Dell Hybrid Cloud Platform Reference Architecture with VMware

Dell Hybrid Cloud Platform Reference Architecture with VMware A Dell Reference Architecture for VMware Built on Dell PowerEdge FX2, SC4020, and Active System Manager, with VMware vsphere 6 and vrealize Dell Global Solutions Engineering April 2016 A Dell Reference Architecture

Revisions Date August 2015 October 2015 April 2016 Description Initial release Update to Reference Architecture in conjunction with release of Deployment Guide Editorial updates THIS WHITE PAPER IS FOR INFORMATIONAL PURPOSES ONLY, AND MAY CONTAIN TYPOGRAPHICAL ERRORS AND TECHNICAL INACCURACIES. THE CONTENT IS PROVIDED AS IS, WITHOUT EXPRESS OR IMPLIED WARRANTIES OF ANY KIND. Copyright 2016 Dell Inc. All rights reserved. Dell and the Dell logo are trademarks of Dell Inc. in the United States and/or other jurisdictions. All other marks and names mentioned herein may be trademarks of their respective companies. 2 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Table of contents 1 Introduction... 5 1.1 Scope... 5 1.2 Audience... 5 1.3 Customer Challenges... 6 1.4 Benefits of the Reference Architecture Solution... 6 2 Design Principles and Benefits... 11 2.1 Drive Business and IT Agility... 11 2.2 Optimize Service Delivery... 12 2.3 Enhance Control... 12 3 Solution Overview... 13 3.1 Public, Hybrid, and Multi-Cloud Capabilities...14 3.2 Self-Service, Governance, and Workflow Orchestration...14 3.3 Unified Management and End-to-End Automation...16 3.4 Health, Capacity, and Performance Monitoring... 18 3.5 Metering, Showback, and Chargeback... 20 3.6 Data Protection, Replication, and Recovery... 21 3.7 Virtualization Platform... 23 3.8 Hardware Infrastructure... 24 4 User Scenarios and Examples... 26 4.1 Provisioning and Management... 26 4.2 Monitoring... 30 4.3 Consolidation... 36 4.4 Capacity Planning... 38 4.5 Governance and Compliance... 39 4.6 Metering and Chargeback... 44 4.7 Business Continuity... 47 5 Solution Architecture... 53 5.1 Solution Requirements... 54 5.2 Core Infrastructure Pod Architecture and Design... 54 5.3 Management Infrastructure Pod Architecture and Design... 58 5.4 Storage Architecture and Configuration...61 3 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.5 Network Architecture and Configuration... 66 5.6 Data Protection, Replication, and Recovery Architecture... 71 5.7 Sizing and Scaling Recommendations... 73 6 Component Specifications... 77 6.1 Hardware Component Details for Core Infrastructure... 78 6.2 Hardware Component Details for Management Infrastructure... 78 6.3 Firmware Versions of Components... 79 6.4 Software Components... 80 7 Solution Verification... 81 7.1 Storage Verification... 81 8 Solution Summary... 83 A Additional Resources... 84 B Compute System Component Details... 85 B.1 Dell PowerEdge FX2 Architecture... 85 B.2 Dell Networking S4048-ON Switch... 88 B.3 Dell Storage SC4020 Array... 88 B.4 Brocade 6510... 89 B.5 VMware vsphere 6... 89 4 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1 Introduction 1.1 Scope This section describes the objective of the Dell Hybrid Cloud Platform Reference Architecture with VMware vrealize and the scope of this document. It includes the target audience for the document, the challenges that the architecture addresses, and some of the significant benefits that are achieved by using the reference architecture, including metrics on time and effort savings by using a highly automated cloud platform. This paper provides a reference architecture for a general purpose cloud infrastructure solution that is suitable for both a private and a hybrid cloud platform. This reference architecture is: Based on the Dell PowerEdge FX2 converged chassis and modular servers, ultra-low latency LAN and SAN switches, and the Dell SC Series high performance enterprise storage system. Designed for a VMware vsphere 6 infrastructure. Optimized with Active System Manager (ASM), Dell s unified management and open automation framework. Integrated with VMware vrealize Orchestrator (vro) for flexibility and vrealize Automation (vra) for self-service and governance capability. Highly reliable with integrated monitoring across the physical and virtual infrastructure and built-in data protection, replication, and recovery. The VMware vrealize Suite is a cloud management solution purpose-built for the hybrid cloud that enables IT to deliver infrastructure and applications at the speed of business with the control that IT requires. This architecture integrates the best of Dell and VMware products and, with Active System Manager s integration to both vcenter and vrealize, it creates a fully featured and highly optimized private and hybrid cloud platform with end-to-end automation capabilities that span from physical to virtual to cloud. This paper presents a scalable and highly available infrastructure that is designed according to the best practices and recommendations for virtualizing enterprise applications and providing them to business users in a cloud model. The paper provides guidance on the choice of components and configurations, rationale as to why certain design decisions were made, and how they deliver measurable value to customers. Deployment and implementation steps for the proposed solution architecture are outside the scope of this paper but may be covered in associated documents. 1.2 Audience The audience for this paper includes, but is not limited to, sales engineers, technologists and architects, field consultants, partner engineering team members, customers, and anyone else interested in deploying an optimized and validated solution stack with the best of Dell and VMware technology. 5 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1.3 Customer Challenges With ever increasing business demands and growing volumes of data, most customers are under unprecedented pressure to improve efficiency and lower costs. However, current operational models of delivering IT services, which involve procuring technology from best of breed technology providers, can prove not only to be time consuming, but problematic for a number of reasons. In this approach, customers are typically burdened to make design decisions, validate various components, setup and configure components manually, and manage the environment in an ongoing fashion by engaging multiple vendors for support. All of these elements across the end-to-end infrastructure life cycle add up to increased complexity and ongoing costs for customers. The Dell Hybrid Cloud Platform with VMware vrealize reference architecture is designed to provide detailed guidance and measurable benefits to customers and help solve the challenges mentioned above. The architecture addresses the following key requirements that we have heard from customers: High flexibility and less complexity: Flexibility in configuration, scaling, and implementation choices based on specific outcomes desired. Reduction in Capital Expenditure (CAPEX): Reduce CAPEX by starting at the right scale, growing in pre-defined units, and paying as you go. Lower operational costs: How to lower your overall TCO by optimizing the infrastructure resources and reducing management complexity. Optimized environment: How to deploy and configure an optimized environment based on best practices to deliver the best experience to your end users. Meet business SLA s under all conditions: How to design, deploy and manage a solution that handles failures without causing disruption. 1.4 Benefits of the Reference Architecture Solution By offering well-planned design choices and providing the rationale for choosing the right components, this Dell reference architecture takes the guess work out of solution design, which reduces the enormous time it takes to procure, validate, and integrate the components on their own. A predesigned and validated reference architecture not only reduces risk, but also provides an optimized solution that can be deployed and operated quickly and efficiently. Dell Solution Engineering teams have designed and thoroughly validated this reference architecture solution. The reference architecture approach therefore provides: Design principles that are central to each targeted solution Architectural design based on best practices Automated deployment and lifecycle management of the solution including pre-built templates, automated firmware updates, and scaling of physical resources Flexibility to address specific customer needs by further building on or modifying the base reference architecture as required 6 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

The reference architecture, as show in the figure below, is a modular platform that can be deployed precisely as specified and can also be built upon or extended to address a wide range of cloud-based solutions. This architecture recommends specific products at each layer to provide the most comprehensive and integrated solution for implementing a hybrid cloud solution. Each of the recommendations has profound benefits for customers. For example: At the hardware layer, the PowerEdge FX2 architecture provides a highly scalable compute building block; the Dell SC4020 provides a high-performance storage network; and the Dell S4048-ON provides an ultra-low latency LAN network. At the virtualization layer, Dell Active System Manager s tight integration with vcenter Server and vsphere 6, as well as integration with vrealize Orchestrator and vrealize Automation, provides a comprehensive automation solution that spans across the physical and virtual infrastructure. At the software layer, Dell Foglight provides deep unified monitoring of both physical and virtual infrastructure; Dell Active System Manager provides unified infrastructure management and automation; and Dell vranger provides secure data protection with agentless, easy to use, and fast backup and recovery. Figure 1 Dell Hybrid Cloud Platform Reference Architecture with VMware The design and end-to-end validation of the reference architecture reduces risk and helps ensure project success. And the solution can still be tailored to support specific customer requirements and preferences, including different server form factors and storage technologies, or allow customers to leverage existing investments in networking, monitoring, and data protection products. 7 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1.4.1 The Dell Blueprint Approach Among Dell s strengths is its broad product and services portfolio that can be tailored to meet customer s requirements regardless of their size, scale, or business model. But a broad portfolio does not mean that it is more complex or difficult to choose. Dell Blueprints are future-ready, comprehensive, and flexible solutions that have been designed, tested, and validated by Dell engineering teams to remove the complexity, optimize the performance, and match our solutions to the customer s desired outcome. Validated Blueprint solutions, which may range from engineered solutions to reference architectures, incorporate best-in-class building blocks from Dell and Dell partners, including data center infrastructure, IT management software, and industry-leading applications. To address the challenges of deploying and operating a private or hybrid cloud, Dell listened to customers and designed a portfolio of purpose-built reference architectures that are optimized for specific applications and workloads and are flexible enough to scale and adapt as needed. These reference architectures are: Built on our best-of-breed products that are designed for virtualization across the ecosystem. Tested and validated, and fully integrated, yet flexible enough to be tailored for your organization, removing risk, and accelerating your time to value. Optimized and protected with investments Dell has made in software that make our solutions easier to manage and deliver ongoing results. Delivered with Dell s global reach and exceptional execution and delivery to provide consistent deployment, management, and maintenance in every region of the world. 8 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1.4.2 Tangible, Measurable, and Profound Benefits This optimized reference architecture provides significant benefits in speed, efficiency, and performance. As an example, using the wizard-driven automation and pre-built templates of Active System Manager, Dell s unified management and automation solution that anchors the reference architecture, you can: Onboard, configure, and deploy new cloud infrastructure in under three minutes of hands-on administrator time, and 95% faster with 77% fewer steps than with manual processes Deploy a new IT service in just 6 steps and under 30 seconds of admin time using a pre-built automation template Provision bare metal hardware and deploy a new ESXi cluster 80% faster and in 71% fewer steps than one of the leading competitors 1 Update firmware on a 4-node virtualization cluster in 90% fewer steps than manual processes Scale up capacity by adding a data store to an existing cluster with 92% fewer steps and 98% less time than with manual processes These automation capabilities deliver speed, reliability, responsiveness, and consistency to the provisioning, deployment, scaling, and de-provisioning of resources in the cloud environment. 1.4.3 State of the Art Financing Solutions The challenges of deploying a cloud platform are not only technical. They also include paying for capital and operational expenditures, often in the face of uncertain financial environments. Dell Blueprint Reference Architectures are ideally suited for the scale-ready payment systems that are available from Dell Financial Services that allow customers to control how and when they pay: based on their forecasted usage; their deployment schedule; or their actual usage. The Dell scale ready payment solutions for enterprise include the following highly flexible options: Pay as You Grow enables customers to install all of their required technology now and pay for their technology over time based on their forecasted usage Provision and Pay allows customers to grow their technology solutions over time by matching their payments to their deployment schedule Scale on Demand enables customers to install their required technology now and pay for it over time based on their actual measured usage These flexible and innovative financing solutions allow Dell to share the risk, convert CAPEX to OPEX, and allow the customer to align their technology acquisitions to their cash flows. 1 Based on independent testing performed at a third-party testing laboratory. To read more, please visit ASM on Dell Tech Center. 9 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1.4.4 Cloud Services for Every Phase of the Solution Lifecycle The reference architecture method provides a fully designed and validated approach for building a private or hybrid cloud. However, it also offers the opportunity to extend or customize the design to meet unique requirements. Wherever you are in the cloud planning or adoption process, Dell services can help you create a cloud strategy that targets your specific requirements without introducing complexity or risk. With Dell cloud advisory services you can get the guidance you need to go from a virtualized environment to a multi-cloud model, including private, public, and hybrid cloud. You can accelerate your virtualization process, and create a solid foundation for cloud adoption with help from Dell specialists. Dell s cloud advisory services can analyze your current business and IT objectives, and provide guidance that helps you build a pragmatic cloud roadmap. You can also use Dell guidance to create a business case that includes your existing investments. With Dell cloud design and implementation services you can take a cloud project from concept to deployment. We begin by translating your business, technical, and operational requirements into a complete design, based on a reference architecture or on your own unique requirements. Next we define the required infrastructure, including the analysis and adaptation of your applications to take advantage of cloud, if needed, or presenting comparable cloud-based applications. Once your design is complete, our team begins the implementation process of infrastructure hardware and software installation. We then move on to application migration and coding, along with integration of systems residing in multiple clouds, as required. In addition, with Dell managed cloud services you can trust Dell to manage your hosted or on-premise cloud infrastructure, operating systems, and applications. Dell managed cloud services includes: A customer delivery executive who serves as your single point of accountability. ITIL-based service and operations management, a standards-based framework for incident and problem management, monitoring and alerting, and change management. 24/7 intelligent voice-response support, available through a toll-free number. A self-service, portal that delivers IT operations lifecycle management capabilities. Dell services are available to help you across every phase of the cloud lifecycle, from design to implementation to management and operations. 10 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

2 Design Principles and Benefits Dell cloud solutions are designed to provide a faster, simpler path to cloud that provides lower TCO and the flexibility to adapt your cloud to the future. Dell Cloud Solutions help you customize a cloud solution to meet your needs and timelines with an end-to-end portfolio, including cross-platform infrastructure services, private and hybrid clouds, heterogeneous management, and expert guidance. With Dell, you can: Drastically cut your cloud deployment time Slash your operating costs and respond faster to business demands Match the right applications to the right cloud to meet your business challenges The Dell Hybrid Cloud Platform Reference Architecture with VMware is a pre-validated yet fullycustomizable reference architecture that offers automated provisioning across physical and virtual environments with powerful and easily-integrated multi-cloud management capabilities. The reference architecture provides prescriptive guidance for building your cloud with a unique system that connects vrealize directly to your physical infrastructure. You can manage your infrastructure as a programmable pool of resources controlled by the services layer and manage your TCO to avoid over-provisioning by deploying the right application on the right cloud. The Dell Hybrid Cloud Platform Reference Architecture with VMware is designed to help you and your organization: Drive business and IT agility Optimize service delivery Enhance control Each of these benefits is described below. 2.1 Drive Business and IT Agility Agility is defined as the power of moving quickly and easily; nimbleness. It is also defined as the ability to think and draw conclusions quickly; intellectual acuity. These define perfectly the desired outcomes of IT and business agility. To get up and running more quickly; to deploy applications faster; to enter markets more quickly; and to be more responsive are all examples of IT and business agility. The Dell Hybrid Cloud Platform was designed to enable IT agility, which drives and supports business agility. This reference architecture helps you: Quickly deploy your cloud with this pre-validated reference architecture that is customizable to your unique needs Easily stand up new business services across any cloud, private or public, through a unified service delivery framework Accelerate IT responsiveness with automated provisioning that spans your physical and virtual infrastructures, creating a fully-elastic pool of resources, using Dell Active System Manager 11 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

2.2 Optimize Service Delivery Optimization is about maximizing resource utilization and gaining efficiencies wherever possible. In IT, that usually involves scalability and simplification and requires tools and systems to provide the insights needed to manage both. Efficiency, simplicity, and scalability are key design principles of the Dell Hybrid Cloud Platform. This reference architecture helps you: Identify your costs through enhanced inventory visibility across your physical and virtual environments and public and private clouds utilizing ASM and Dell Open Manage Essentials (OME) Scale your physical and virtual infrastructures precisely with Dell PowerEdge FX modularity and tailor payments that meet your needs with flexible financing solutions from Dell Financial Services (DFS) Simplify service deployment, reducing effort and errors, through workload-based automation templates supported in ASM 2.3 Enhance Control Security, compliance, data protection these are some of the most cumbersome, yet critical, responsibilities of an IT organization. To manage these elements and successfully deliver IT as a service there must be proper controls in place. These controls are meant to facilitate and monitor, not obstruct and limit. The Dell Hybrid Cloud Platform provides the controls you need to enhance overall responsiveness and user experience while maintaining governance and proper protections. This reference architecture helps you: Improve governance and better manage shadow IT with robust multi-cloud management Deliver the right user experience through end-to-end service level monitoring and proactive problem detection and resolution utilizing Dell Foglight and OME Protect your data and ensure compliance with capabilities that align investments and data value using Dell vranger and the Dell backup appliance DR6000 These design principles are inherent throughout all aspects of this reference architecture design. In Section 4 (User Scenarios) you can see examples of workflows and usage scenarios for each of the different users of the platform that illustrate very clearly the agility, efficiency, and control that are achieved with this solution. 12 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

3 Solution Overview This paper describes a reference architecture for a general purpose cloud infrastructure solution that is suitable for both a private or hybrid cloud platform based on the Dell PowerEdge FX2 converged system; the latest Dell LAN, SAN, and storage technology; VMware vsphere and vcenter Server 6.0 and components of the vrealize suite; and Dell Active System Manager as an overall unified management and automation framework. The reference architecture provides a scalable and highly-available infrastructure with built-in monitoring and data protection. It is designed according to the best practices and recommendations for virtualizing enterprise applications and providing them to business users in a cloud or Infrastructure-as-a-Service (IaaS) model. The VMware vrealize Suite is a purpose-built cloud management solution for heterogeneous data centers and hybrid cloud environments. With the integrated combination of Active System Manager, VMware vrealize Orchestrator, and vrealize Automation, the end-to-end automation capabilities of this platform span from the bare metal physical infrastructure, inclusive of servers, networking, and storage, to the virtual infrastructure, including hypervisor deployment and cluster configuration, to operating system and VM deployment, and all the way to application-level provisioning. This provides a complete Infrastructure as a Service (IaaS) cloud platform with unparalleled automation capabilities. A high level architecture diagram of the key components of the solution is shown below. Figure 2 Components of the Hybrid Cloud Platform Reference Architecture This reference architecture integrates the best of Dell and VMware products. With Active System Manager s integration to both vcenter and vrealize, integration of physical and virtual monitoring, and built-in data protection, it creates a fully featured, deeply automated, and highly optimized private and 13 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

hybrid cloud platform with end-to-end automation capabilities from the physical to the virtual to the cloud. The core capabilities of the reference architecture are described in the following sections. 3.1 Public, Hybrid, and Multi-Cloud Capabilities Hybrid clouds provide an optimum balance of private and public cloud capabilities, including secure ondemand access to shared but private resources along with the flexibility to deploy or move workloads offsite to public clouds to meet specific needs. This model provides both security and scalability, balances capital and operational expenses, and delivers the benefits of providing for peak demand capacity and disaster protection. This reference architecture describes a hardware and software stack that is designed as a highly optimized private cloud platform and, by providing a flexible management and automation framework from Dell and VMware that allows the incorporation and consumption of public cloud resources, provides the foundation for a hybrid cloud as well. With vrealize Automation, users and administrators can provision virtual machines into private resource pools managed by Active System Manager or provision resources into multiple public clouds, including Amazon Web Services, Microsoft Azure, vcloud Air, and other public cloud providers that can be configured as endpoints of vrealize Automation. (See the vrealize Automation documentation for current information on cloud endpoints.) By extending the private cloud capabilities enabled in this reference architecture into one or more public cloud instances, you achieve not only a hybrid cloud model, but a multi-cloud capability, which enables the maximum degree of business agility and the most flexible economic model. 3.2 Self-Service, Governance, and Workflow Orchestration From the VMware vrealize Suite, vrealize Automation is included to provide a self-service catalog, request and approvals framework, and policy and governance capabilities. vrealize Automation is designed for the Software Defined Data Center (SDDC) and has the control, governance, and multi-tenancy needed for both private and hybrid cloud deployments. vrealize Orchestrator provides a flexible orchestration framework that can be used for automating technical tasks or integrating with business processes such as helpdesk and service ticketing. vro is also the means by which Dell Active System Manager integrates into the vrealize Suite. VMware vra helps IT departments with delivery and ongoing management of infrastructure, applications, and custom services. VMware vra can be used to build both enterprise and service provider cloud solutions. Supporting implementations for private clouds, public clouds, and hybrid clouds, vra can be used in conjunction with existing or new business processes and tools to leverage a very flexible, highly automated, and intuitive self-service experience for the end user. 14 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

VMware vrealize Automation: Delivers infrastructure, application and custom services through a unified IT service catalog Meets specific business needs at the right service level with personalized, policy-based governance Delivers governance and control to enable hybrid cloud deployments Figure 3 vrealize Automation with ASM template integration VMware vra makes it easy to maintain control over provisioned physical and virtual workloads via its centralized provisioning structure, governance, and infrastructure management capabilities. For example, lifecycle management of virtual machine workloads is controlled with vra through built-in processes for request, approval provisioning, management, reclamation, and decommissioning. For orchestration, vro simplifies the automation of complex IT tasks and integrates with other VMware vrealize Suite components to adapt and extend service delivery and operational management, effectively working with existing infrastructure, tools, and processes. VMware vro allows administrators to develop complex automation tasks in a graphical workflow manner, and quickly access and launch workflows from the VMware vsphere client, various components of the VMware vrealize Suite, or other triggering mechanisms. VMware vro is used for many cloud deployments to create custom workflows interacting with governance, provisioning, documentation, monitoring, service ticketing, and IP address management (IPAM) systems. VMware vro can use functionality exposed by the ASM API to create end-user or administrator-specific services. The ASM plugin to vro allows a business analyst or IT architect to embed ASM templates and automation methods into orchestration workflows, and therefore integrate seamlessly into existing business processes, such as requests, approvals, help desk, and self-service. This saves time and manual effort and improves responsiveness and consistency. In addition, the ASM integration with vro allows an 15 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

IT administrator to build templates for common or standardized IT services and infrastructure requests in ASM and publish them to the vra service catalog for end user self-service fulfillment. Figure 4 vrealize Orchestrator with ASM plugin 3.3 Unified Management and End-to-End Automation Active System Manager is Dell s unified management and automation solution that simplifies the deployment, provisioning, and management of heterogeneous resources, including both infrastructure and workloads, and delivers the benefits of automation to service-centric IT management. For the Hybrid Cloud Platform, ASM provides an overall unified management console and end-to-end automation framework. This framework provides broad capabilities that range from hardware discovery, onboarding, initial configuration, and inventory management; to physical resource pooling, bare metal provisioning, and firmware compliance; to virtual infrastructure provisioning, with deployment of hypervisors and configuration of hosts and clusters; to virtual machine, application, and workload provisioning. Together with its tight VMware integration, ASM provides a truly complete end-to-end management and automation experience. ASM delivers: Rapid time to value by accelerating growth and innovation with fast, responsive IT automation Superior ease of use by optimizing IT and simplifying and enterprise systems management Unrivaled flexibility by increasing options with an open and extensible architecture ASM enables system administrators to deploy new services from pre-built templates in 6 steps and 30 seconds of hands-on time and 80% faster than similar products. 16 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 5 Dell Active System Manager ASM takes a top-down, service-centric approach to IT automation that spans across both physical and virtual infrastructure and encompasses everything from servers, switches, and storage, to hypervisors and clusters, to virtual machines, operating systems, and applications. ASM provides a unified and comprehensive user experience for the initial deployment and the ongoing lifecycle management of shared or converged infrastructure. ASM enables IT administrators to accelerate service delivery and improve efficiency in a private cloud platform in a number of ways, including the ability to: Get up and running quickly, with rapid discovery and initial configuration, and wizard-driven automation of server, chassis, and I/O network component onboarding Efficiently manage infrastructure lifecycle, with comprehensive firmware management, compliance monitoring, and updates Define and provision unique or complex physical, virtual, and hybrid services as reusable templates Easily deploy, manage, and scale IT services, with automated provisioning, scaling of physical and virtual resources up and down, and decommissioning of services to recover resources Define and manage user access levels and permissions with role based access control Respond quickly with dynamic, on-demand provisioning of resources from shared pools In short, ASM provides a strong foundation for a private or hybrid cloud environment. ASM supports a plugin to vrealize Orchestrator and, through this vro integration, the ability to publish ASM templates to the vrealize Automation service catalog. The ASM plugin to vrealize Orchestrator: 17 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Allows ASM automation templates to be called from vro orchestration workflows Enables allows automation methods supported by ASM including provisioning, configuration, capacity scaling, and de-provisioning to be embedded into vro workflows Extends the automation and orchestration capabilities of vro to bring ASM automation into business process workflows to improve responsiveness and simplify complex IT processes Enables integration with vrealize Automation so that automation templates for infrastructure and workloads can be published to the vra service catalog for a self-service IT experience 3.4 Health, Capacity, and Performance Monitoring Health, capacity, and performance monitoring of the physical and virtual infrastructure is provided by Dell Foglight for Virtualization, which is integrated with Dell Open Manage Essentials (OME) for hardware monitoring, and plugin cartridges into the storage, LAN, and SAN networking components. The integrated combination of Foglight with the hardware subsystems provides a unified monitoring interface across the physical and virtual layers. Dell Foglight for Virtualization provides insights into the health, risk and efficiency of virtual infrastructure and includes change tracking and performance impact analysis. Using Foglight, administrators can quickly identify bottlenecks and analyze network traffic. Foglight provides a unified dashboard for the overall health of the virtual infrastructure, current resource utilization, and any issues that need an administrator s attention. Using Foglight, IT managers get the visibility into costs across virtualization operations. Administrators can easily charge back cost centers based on the resource utilization at a department or business level. Figure 6 Foglight for Virtualization 18 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Some of the features and benefits of Foglight for Virtualization include: End-to-everything visualization: Experience unmatched visibility into VMware and Hyper-V environments with granular monitoring and management across the virtualization infrastructure stack from the virtual machine down to the physical disk array. vswitch support: Get a complete picture of your infrastructure with comprehensive network topology and mapping with end-to-end network routes between VMs and hosts. Easily monitor network usage and highlight performance alarms from VMs, vswitches and hosts. Optimization: Reduce OPEX with expert advice on optimizing VM density, including vcpu, vmemory, and guest storage. Utilize the Wastefinder tool to locate abandoned, powered off, unused templates, or snapshots and zombie VMs and then select the option to immediately remediate and return wasted resources back to your resource pool. Automation: Enforce proper VM lifecycle management with automation capabilities that help you easily deploy new VMs into an environment, decommission VMs, and remove VM files that are unnecessary. Reduce mean-time-to-resolution (MTTR) with automated root cause analysis and virtual machine performance remediation workflows. Capacity planning and resource utilization management: Control CAPEX with insight into resource utilization across your virtual infrastructure. Easily monitor and access current workload capacity, identify shortfalls, analyze alternative scenarios, and plan for data center growth or additional virtualized workloads. The unified monitoring view of virtual and physical infrastructure is achieved in part through integration between Foglight and Open Manage Essentials, Dell s one-to-many management console for monitoring Dell data center equipment including server, storage and networking devices. OME also provides lifecycle management of Dell PowerEdge Servers. It is the easiest and simplest way for system administrators to maximize Dell system uptime and health. Key features of OME include: Monitoring health status, events and inventory for Dell PowerEdge servers, storage and networking devices Providing hardware level control and management for Dell PowerEdge server, blade chassis, and internal storage arrays Enabling deeper management and control of Dell Blade chassis, Dell storage, and Dell networking devices through context-sensitive link and launch of their respective element management tools 19 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 7 Dell Open Manage Essentials Open Manage Essentials also integrates with Dell SupportAssist, which enables automatic hardware failure notification sent securely to Dell technical support for intelligent analysis and diagnosis to optimize availability and reduce manual intervention. With the Dell OME integration into Foglight for Virtualization, you can: Follow virtual issues right down to the hardware level View historical analysis (not just real-time) Track changes in your Dell server environment from a single view Run predictive analytics that help you understand and address future issues before they happen Easily track changes within your server farm In this reference architecture, Dell Foglight for Virtualization serves as the primary monitoring interface for the overall virtual and physical infrastructure, with hardware health and status being surfaced into it through the OME integration. In addition, the OME client interface is available for detailed hardware monitoring and fault analysis as needed. 3.5 Metering, Showback, and Chargeback The requirements for tracking and metering of resource usage and for reporting, including showback or chargeback capabilities, vary widely from organization to organization. This reference architecture provides several capabilities to address different requirements. Foglight is the primary interface specified in this reference architecture for reporting including showback and chargeback. Foglight allows you to simplify the management of the business of the data center and 20 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

improve the return on investment of your existing IT infrastructure with visibility and accountability across virtualization operations. It provides a clear view of resources consumed and their associated costs so you can chargeback cost centers or external customers with ease.. 3.6 Data Protection, Replication, and Recovery Data protection, replication, and recovery services are built in to the reference architecture with Dell vranger Pro, which has deep VMware integration, and a Dell Data Protection appliance that provides local, remote, and cloud-based backup with inline deduplication. Dell vranger software provides high-speed backup and recovery of virtual infrastructures built using VMware and physical Windows Server environments. It also provides high-speed replication for VMware. It protects entire virtual environments in minutes, detecting and backing up new VMs automatically, and delivers safe, scalable data protection to even the largest VMware and Hyper-V environments. With vranger, you can locate and restore individual files in seconds, even if they are buried in virtual and physical backups, from a single, intuitive interface. Dell vranger, when combined with Dell Storage DR appliances, provides complete backup, replication, deduplication, and recovery capabilities for the virtual infrastructure environment. Figure 8 Dell vranger Pro vranger provides high-speed, resource-efficient backup, replication and recovery of virtual machine images, and supports backup and recovery of Windows physical servers, files and folders. In addition, vranger delivers maximum storage savings when paired with the Dell DR deduplication appliance. 21 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Some of the key features and benefits of vranger Pro include: Change block tracking (CBT): Eliminates the time required to scan for changed blocks in guest images on vsphere hypervisor systems to speed backup and replication jobs. Instant file-level recovery (FLR) for Windows and Linux: Enables you to quickly restore a single file from a backup image in the repository through a one step process. Native, full catalog capability: Provides native, full catalog of every image in the backup repository enabling immediate identification of available recovery positions. Wildcard scanning feature quickly locates backup repository files to be restored. One-step catalog recovery: Provides advanced search (including wildcards) and right-click recovery selection directly from the management console to speed up restore of VMs, save points, and hosts with native catalog. Advanced encryption standard (AES)-256: Secures protected images block-by-block on the VMware host as they are read so they are secure over the network and in the backup repository. Full, incremental, and differential backup: Enables a complete backup cycle for protected images that is optimized for speed and resource efficiency. Full, incremental, and hybrid replication: Provides the full range of options required to efficiently replicate VMware VMs over LANs and WANs. VMware vsphere vmotion support: Automatically protects VMs as they move from one host to another even when backup jobs are running. The Dell DR Series Disk Backup and Deduplication Appliance is a physical appliance and works in close conjunction with vranger Pro to deliver maximum performance and functionality. The DR Series systems are extremely efficient, high performance, disk-based backup and recovery appliances. The DR Series are simple to deploy and manage and offer unsurpassed total cost of ownership (TCO) benefits. Through the use of innovative Dell deduplication and compression technology, the DR Series systems can help achieve data reduction levels up to 15:1. This reduction in data means that more backup data can be retained longer and within the same footprint. Figure 9 DR6000 Backup Appliance 22 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

3.7 Virtualization Platform This reference architecture is based on VMware vsphere 6 with ESXi and vcenter Server as the virtualization management platform. VMware vcenter Server provides a centralized platform for managing all your VMware vsphere environments so that you can automate and deliver a virtual infrastructure with confidence. vcenter Server is used to manage the ESXi hosts and the logical vsphere clusters. vcenter Server provides centralized control and visibility through: vsphere web client: Manage the essential functions of vsphere from any browser. vcenter single sign-on: Allow users to log in once and access all instances of vcenter Server, without the need for further authentication. Custom roles and permissions: Restrict access to the entire inventory of virtual machines, resource pools, and servers by assigning users to custom roles. Inventory search: Explore the entire vcenter inventory, including virtual machines, hosts, datastores and networks, from anywhere within vcenter. Figure 10 vrealize Automation with ASM template integration vcenter Server integrates with vrealize Orchestrator to automate more than eight hundred tasks using out-of-the-box workflows or by assembling workflows with an easy drag-and-drop interface. In addition, Dell Active System Manager integrates with vsphere virtualization platforms and ESXi hosts through a centralized vcenter Server instance to integrate management and automation of the overall physical and virtual environments. 23 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

3.8 Hardware Infrastructure The hardware infrastructure of the Hybrid Cloud Platform consists of a core infrastructure unit or pod and a management infrastructure pod. The core provides the private cloud infrastructure and consists of the Dell PowerEdge FX2 converged chassis with FC630 servers, Dell S4048-ON LAN and Brocade 6510 SAN switches, and Dell SC4020 storage. These components are summarized below and described in more detail in Sections 5 and 6. The management elements are described in detail in Section 5.3. 3.8.1 Compute The PowerEdge FX2 is a 2U hybrid rack-based computing platform that combines the density and efficiencies of blades with the simplicity and cost benefits of rack-based systems. With an innovative modular design that accommodates IT resource building blocks of various sizes compute, storage, networking, and management the FX2 enables data centers infrastructures to be built with greater flexibility. The FX2 enclosure also offers I/O modules to several I/O aggregators that can simplify cabling, improve east/west traffic within the server, and enable LAN/ SAN convergence, reducing cost and complexity. The Dell PowerEdge FC630 is designed to be a workhorse for data centers looking for new levels of efficiency and density in an incredibly small footprint. Powered by up to two 18-core Intel Xeon E5-2600 v3 processors, each FC630 has 24 DIMMs of memory, two 2.5- inch or eight 1.8-inch front access drives, a 10Gb SNA and access to two PCI Express (PCIe) expansion slots in the shared chassis. 3.8.2 Networking The Dell Networking S4048-ON is a 1U highdensity 10/40 GbE ToR switch with 48 dualspeed 1/10GbE (SFP+) ports and six 40 GbE (QSFP+) uplinks. This switch leverages a nonblocking and cut-through switching architecture to provide ultra-low-latency performance for applications. The six 40 GbE ports can be used to create either a Virtual Link Trunk (VLT) between the switches to enable traffic isolation within the solution infrastructure or as a connectivity to the data center core network. 24 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

The Brocade 6510 is a 48-port 1U Gen 5 FC switch that is suitable for high-performance data requirements of server virtualization, cloud, and enterprise applications. This switch can be configured in 24, 36, or 48 ports and supports 2, 4, 8, 10, or 16 Gbps speeds. This enables enterprises to start small and scale the FC infrastructure based on the growth, as needed. By using features such as Virtual Fabrics, Quality of Service (QoS) and zoning, IT departments can build multi-tenant cloud environments. Management and diagnostics features such as Monitoring and Alerting Policy Suite (MAPS), Dashboards, Flow Vision, Fabric Performance Impact (FPI) monitoring, and Credit Loss Recovery help administrators and IT organizations avoid problems before they impact the SAN operations. 3.8.3 Storage The Dell Storage SC4020 storage array offers multi-protocol support and virtualized multi-tier, multi- RAID-level storage policies. Each SC4020 array comes with dual redundant controllers, 24 internal drive slots, eight 8 Gb Fibre Channel (FC) or four 10 Gb iscsi network ports, and one 1 Gb port per controller for out-of-band (OOB) management traffic. The supported drive types range from Write-Intensive (WI) Solid State Drives (SSDs) to Read-Intensive (RI) SSDs to 15K, 10K, and 7.2K SAS drives in small form factor (2.5- inch). This storage array supports multi-tier data placement that improves application performance. This array supports expansion up to 192 drives by adding Dell Storage SC220 or SC200 enclosures. The Storage Center 6.5 Operating System (OS) provides features such as block-level compression, synchronous Live Volumes to restore data in a non-disruptive manner, and Active Directory (AD) Single- Sign-On (SSO). The Storage Center Manager enables easier and out-of-box web-based management of SC4020 arrays while the Enterprise Manager and its components can be leveraged to build a secure, multi-tenant environment with the Dell Storage Center SAN. 25 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

4 User Scenarios and Examples The Dell Hybrid Cloud Platform Reference Architecture with VMware provides a comprehensive solution for your private and hybrid cloud needs. In this reference architecture we segregate the primary platform users into three categories: Cloud admins, who have end-to-end ownership of the platform from infrastructure to virtualization layer. Their responsibilities include setting up infrastructure for LOBs or departments and ensuring ongoing systems operations and readiness. Tenant admins, who are responsible for managing specific line of business or departmental needs in the shared infrastructure platform. They are responsible for creating blueprints, defining policies, and exposing the blueprints for consumption through the self-service portal to end users. They have restricted access to provision infrastructure and can monitor virtual and physical infrastructure relevant to their LOBs or departments from Foglight. End users, who request workloads from the self-service portal in vra. The following user scenarios provide examples of how the architecture can help drive business and IT agility, optimize service delivery, and enhance control for your cloud environments. These scenarios are organized into seven areas: Provisioning and Management Monitoring Consolidation Capacity Planning Governance and Compliance Metering and Chargeback Business Continuity Most of the use cases below are relevant to both cloud and tenant admins across multiple functional areas. 4.1 Provisioning and Management 4.1.1 Set Up and Provision End-To-End Infrastructure from the Self-Service Portal vrealize Automation provides end user access to a self-service portal for provisioning VMs on demand under guidelines set by IT. This allows application teams to provision application workloads without being stranded for resources. However, this does not eliminate the need for the cloud administrator to work with multiple infrastructure teams to onboard, configure, and deploy new cloud infrastructure. Without ASM, manual approaches involve submitting a request for provisioning server, storage, networking, and virtualization layers with various teams and then manually configuring each layer 26 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

separately. This slows down the process of end-to-end infrastructure provisioning and can impact the experience of the end user and their organization. Dell s Hybrid Cloud Platform provides a unified infrastructure management layer. ASM supports a plugin to vrealize Orchestrator and, through this vro integration, provides the ability to publish ASM templates to the vrealize Automation service catalog. This integration lets cloud admins set up both virtual and physical infrastructure right from the vra self-service portal and drives IT and business agility. Consider a scenario where the cloud administrators have to setup a new ESXi cluster for a specific LOB on the cloud platform. The cloud admin logs in to the vrealize portal and within the Service Catalog selects the service blueprints to provision a new ESXi cluster, as shown below. Figure 11 vrealize Portal provisioning of ESXi cluster They then select the right ASM template for provisioning the end-to-end infrastructure, shown below. 27 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 12 Dell Active System Manager templates ASM automates the end-to-end provisioning across storage, networking, compute, and virtualization and creates an ESXi cluster in a matter of minutes. In validated testing, ASM has on-boarded, configured, and deployed new cloud infrastructure in less than three minutes of hands-on administrator time and was 95% faster with 77% fewer steps than with manual processes. ASM s end-to-end automation does more than save time for infrastructure provisioning. By providing a single configuration routine (templates) that is based on best practices it reduces possible errors and improves overall service quality. ASM also provides a single abstraction layer for compute and storage, reducing complexity and the learning curve for administrators to set up cloud infrastructure. 4.1.2 Scale On-Demand through ASM and vrealize Integration While the unified infrastructure management layer accelerates the initial setup of the virtual infrastructure, the real value of infrastructure automation exposed into vra is enabling rapid scale-up and scale-down of infrastructure with a single click. Dell Hybrid Cloud Platform s unified infrastructure management layer, ASM, supports the ability to scale-up, scale-down, and de-provision compute and storage resources from the vra service catalog. Imagine a scenario where a business unit you support needs additional compute infrastructure for a brief period of time web servers to support peak demand over a holiday season, for example. This usually requires requesting multiple teams for server provisioning, building the OS, and adding the new hosts into existing ESXi infrastructure. These manual steps and the involvement of multiple teams for repetitive tasks either slows down IT from reacting quickly for new business demands or requires significant upfront planning and, perhaps, expense. By integrating ASM and vrealize, Dell helps you provision much more quickly and simply, and without the need for a lot of upfront planning and cost. 28 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

As shown in the figures below, when a new request for infrastructure provisioning comes from the application team, the cloud admins login to vra s service portal and kick off ASM flows that provision additional hosts. They then select the service running within ASM and trigger an ASM flow to add additional hosts. ASM then configures the new hosts, deploys the OS as needed, and adds them to the existing cluster, scaling up the environment within a matter of minutes. Figure 13 Services in vra to scale-up compute and storage for an ESXi cluster Figure 14 ASM triggered to add hosts 29 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 15 ASM configures hosts, deploys OS, and adds to the cluster With ASM, cloud admins can use the service catalog in vra to provision on-demand physical resources (servers and storage) to meet spikes or troughs in business demand. This integration makes reacting and adapting much simpler and contributes directly the business agility and responsiveness. 4.2 Monitoring 4.2.1 End-to-end View from Physical to Virtual to Cloud Monitoring is a critical component when multiple organizations are using resources from a common pool and running a diverse set of applications. Most monitoring platforms stop at the virtualization layer with little or no visibility into the underlying physical infrastructure. Dell Hybrid Cloud Platform solves this problem by providing a single end-to-end view for monitoring physical, virtual, and cloud infrastructure. As shown below, cloud admins can monitor current utilizations as well as the health state of related objects within Foglight. When a user clicks on VMware Explorer in Foglight and selects an individual virtual machine, Foglight captures not only the standard performance metrics of CPU and memory utilization, but also network, storage throughput, IOPS, response time, and capacity utilization. 30 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 16 Foglight captures performance metrics Consider a scenario where your application team complains about poor latency issues. With a traditional approach, cloud admins have to work with multiple teams, multiple consoles, and little shared information to identify the root causes and take corrective action. This can take a significant amount of time and severely impact the business unit experience and, in some cases, result in unplanned downtime of applications. Dell s end-to-end monitoring alleviates these issues. As seen below, when admins are notified about poorly performing applications they log into Foglight to first understand the SAN topology with an end-to-end view from a VM or workload to the network fabrics to the storage subsystem. Figure 17 Foglight end-to-end topology view helps identify trouble spots 31 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Dell Foglight collects performance information from the storage infrastructure, arrays, virtual machines, and hypervisor hosts and correlates them end-to-end to give a complete view of performance over time. As seen below, Foglight gathers near-real-time VM, Fibre Channel, iscsi, and storage performance data from the environment and maps it to VMs, hosts, and service paths, helping administrators to: Optimize storage allocation to accelerate application performance. Improve virtual machine, server, array, switch, and fabric utilization without compromising performance. Identify the best VM, host, array, and switch resources to allocate to a new application Figure 18 Foglight gathers performance data and maps to VMs, hosts, and service paths This complete visibility from physical to virtual to cloud helps prevent issues before they happen. It also provides faster troubleshooting. Admins can proactively monitor the environment for issues and fix them before they cause any impact on the application workloads. All of this helps ensure optimal service delivery for business units and end users. 4.2.2 Monitor and Resolve Issues by Rapid Scaling Dell s Hybrid Cloud Platform, via the integration of compute and storage monitoring through plugins into Foglight, not only offers a single pane of glass for monitoring physical and virtual infrastructure, but also allows administrators to drive insights and take action. In the examples shown below, administrators monitor data store usage for existing free space as well as growth rate and set up alarms to be notified in advance about capacity issues before active workloads running on the platform are disrupted. When these alarms are triggered, cloud admins can react quickly and allocate more storage to the environment using ASM services in the vra service portal. Scaling up physical resources through ASM workflows that are integrated into vra not only saves time for admins when reacting to potential issues, it also lets them allocate storage only as needed, optimizing resources. 32 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 19 Foglight shows data store for cluster running out of capacity Figure 20 Action can be taken immediately add data store to the ESXi cluster to resolve issues Contrast this capability to a platform without the unified infrastructure monitoring and automation that Dell Hybrid Cloud Platform provides. Administrators would have to work with individual teams that manage compute, storage, and networking to isolate the issue to a specific layer and take corrective action in a manual fashion. This can significantly impact the experience for the business unit and end users. The ability to proactively resolve issues with a few clicks helps you deliver a higher quality of service. 33 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

4.2.3 Monitor and Resolve Issues through Easier Troubleshooting and Remediation Foglight for Storage Management makes troubleshooting very easy for cloud admins. It helps administrators determine the cause of virtual machine performance issues and assists in remediation. Even transient, unpredictable issues can be readily addressed. Admins can scroll back in time to precisely when a performance issue was reported and examine exactly what was affecting virtual machine performance at that moment. Consider a scenario where your application team complains about poor latency issues. In such a scenario, admins can drill down from the reported VM to the volume and the LUN on the storage system to identify the root cause of the issue. As shown, the admin can select the VM and then drill down to analyze the SAN storage. Foglight presents a view of performance from the perspective of the virtual machine, showing latency and workload as compared to both configured thresholds and typical activity. Figure 21 Foglight shows virtual machine performance at the LUN level In this example, the storage array is reporting that the LUN is responding with latency that exceeds the pre-defined thresholds and is higher than what is typically seen at this time of day. This could mean that other LUNs within the storage pool are consuming excessive IOPS causing high latency for this VM s LUN. The admin can dig deeper using Pool Change Analysis to identify if a specific LUN in the pool has experienced significant changes in IOPS compared with an admin-configurable time period in the past. This analysis helps identify whether another LUN in the pool started driving more traffic than normal. 34 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

In the example below, a LUN is experiencing a lot more I/O than is typical. If a volume or LUN had a disk failure or was rebuilding that would be reported. With no failed disks or other storage array issues, this LUN simply can t keep up with the workload being demanded of it. A storage vmotion of the virtual machine to a more lightly loaded or higher performing volume should address the problem. Figure 22 Performance pool change analysis identifies failures or workload issues so corrective action can be taken With such deep integration into the underlining SAN layer, admins can quickly troubleshoot and resolve issues and improve IT service delivery experience. 35 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

4.3 Consolidation 4.3.1 Infrastructure Consolidation Implementing a private or hybrid cloud has the benefit of consolidating servers and storage in the process. By integrating data center silos into a cloud infrastructure, the benefits of server, storage, and data center consolidation can be attained, including lower costs, less complexity, and simplified deployment and management. Figure 23 Consolidating silos into a cloud infrastructure leads to lower costs and less complexity 4.3.2 Common Architecture for Any Scale When it comes to infrastructure sizing, customers rarely know the final amount of compute and storage they will need. These requirements are tied to business needs and pace of growth, which can change over time. They are faced with the tradeoff of buying a smaller solution that will need an upgrade later or buying a larger solution and paying for capacity that they do not need yet. What customers need is the flexibility to operate at any scale with no tradeoffs of compute power between the convenience and flexibility of rack deployments and the density and performance of blade servers, or the different levels of performance offered by various storage solutions. This reference architecture uniquely adopts a single infrastructure architecture regardless of scale. This means no change in core technology or added complexity as the infrastructure grows. The reference architecture offers customers: Flexibility to start small and grow in smaller chunks: The reference architecture supports deployments that can start small with a single FX2 chassis with four servers and a single storage 36 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

array SC4020 and then scale to add more compute nodes (FX2 chassis), storage enclosures (SC220), or both to support constantly changing workload demands. Reduced OPEX and operational complexity: By adopting a single, standardized design across the board and transforming the resource consumption model from an inefficient over-provisioning model to a pay as you grow model, OPEX and operational complexity are reduced. The reference architecture has been designed around the concept of a scale-unit. Each scale-unit is a self-contained, balanced pod where the computer, networking, and storage are built to be sized appropriately for each other based on the initial requirement. Each pod can either be scaled individually or together, based on your ongoing requirements. Figure 24 Scale units allow growth within the rack and beyond To help manage scale with costs, Dell Financial Services allows customers to control how and when they pay based on their forecasted usage, their deployment schedule, or their actual usage. The Dell scale ready payment solutions for enterprise include the following highly flexible options: 37 Pay as You Grow: allows customers to install all of their required technology now and pay for their technology over time based on their forecasted usage Provision and Pay: allows customers to grow their technology solutions over time by matching their payments to their deployment schedule Scale on Demand: allows customers to install their required technology now and pay for it over time based on their actual measured usage Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

These flexible and innovative financing solutions allow Dell to share the risk, convert CAPEX to OPEX, and allow you to align your technology acquisitions to your cash flows. 4.4 Capacity Planning 4.4.1 Plan In Advance In a self-service environment, with many users across different LOBs, monitoring usage trends and growth rates is extremely important. Lack of visibility for proper planning can lead to cases where the platform won t have enough resources to manage business requirements. Dell Hybrid Cloud Platform s single monitoring pane, Foglight, solves this problem. It ensures that usage trends and growth rates are tracked to alert the admin of current and future capacity problems. Cloud admins can track growth rates at the cluster level. In the example below, the cluster is under allocated for memory and has fifteen months until it is full for storage; whereas CPUs are sufficient for greater than two years at the current growth rate. These insights, based on past and current infrastructure usage, ensure better capacity planning and help optimize infrastructure costs, alerting you where to spend now and where you can wait. Figure 25 Foglight aides capacity and investment planning 38 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

4.4.2 Drive insights to prevent failures Foglight can help administrators derive insights that can prevent possible disruption to workloads. In the example below, the cluster is measured against an N+1 model. The admin wants to track and monitor if the cluster can handle the current load when one of the servers in the cluster fails. However, the cluster is over-allocated by two VMs and does not have sufficient resources if one of the nodes in the cluster fails. It also shows other bottlenecks in the environment. In our example it shows storage as next possible bottleneck. Figure 26 Foglight helps admins proactively spot trouble before it happens This level of visibility ensures that cloud administrators have insights to help maintain high availability and helps address issues in a proactive fashion. 4.5 Governance and Compliance 4.5.1 Define Alarms and Be Notified of Possible Issues You can define at what points your admins receive alarms and what actions are to be taken when trigger points are reached, including automating proactive or recovery operations. The ability to stay in front of issues helps keep IT and business operations running smoothly. 39 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 27 Define alarms to stay in front of potential issues 4.5.2 Multi-Tenancy at the Resource Level Using Constructs in ASM Given that different organizations are serviced from a shared pool of resources, ensuring that each of them has access to only those resources that they have security and have paid for is critical to providing optimized service delivery. The Dell Hybrid Cloud Platform Reference Architecture with VMware supports secure multi-tenancy at two different levels. vrealize Automation s role-based access control (RBAC) restricts user access to the self-service portal based on their business role Resource isolation at the physical layer can be ensured through the resource pooling construct in ASM As shown below, cloud admins can map compute resources to specific resource pools to ensure physical and logical separation between functional areas such as Production, Testing and Development, and different LOBs, such as Finance and HR. While provisioning a new ESXi cluster or while adding hosts to an existing cluster this logical separation restricts functional areas to using only those resources they have paid for or have access to. 40 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 28 Logical separation of resources ensures guaranteed service delivery to various organizations 4.5.3 Optimize the Environment with Resource Usage for Active Workloads With different organizations serviced from a shared pool of resources, having capabilities to tune active VMs running in the platform is critical to ensure optimized service delivery. During blueprint creation within vrealize tenant admins can specify the maximum machine resources (CPU, Memory) that end users can use to provision VMs. This gives a degree of control to ensure optimal usage of resources, however it still leaves it to the discretion of the tenant admin to optimize resource usage while provisioning. Foglight, through its optimizer feature: Provides visibility into the average and peak resource consumption per VM Provides a recommendation on how to optimize Allows admins to reclaim poorly used resources and bring them back to the common pool In the example below, we see a list of VMs that are underutilizing their CPU usage and Foglight is providing a prescriptive recommendation of Decreasing CPU allocation. This allows the admin to reclaim resources that can be used elsewhere. 41 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 29 Foglight prescribes actions to optimize resource utilization 4.5.4 Optimize the Environment to Remove Zombie VMs Another issue that admins often run into with cloud environments is to track and reclaim resources of those VMs that have been inactive and unused for a long period of time. A good example of this would be a test/dev environment where a developer set up a LAMP stack a few months ago and hasn t used it again. Foglight lets admins track such inactive and unused VMs and reclaim resources for use elsewhere. In the following example, Foglight lets you to identify abandoned and potential zombie VMs and reclaim resources that can be used elsewhere. This ensures that resources are optimally utilized and rogue processes don t affect the performance of other active VMs. 42 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 30 Foglight identifies unused resources for reclamation and utilization 4.5.5 Keep Firmware Up To Date Hidden infrastructure configuration and workload provisioning errors can impact service delivery. Even if errors are minimized during initial deployment, ongoing updates and modifications can create unexpected hidden errors. To maintain system performance and optimization, as well as compliance, it is important to keep firmware updated. With ASM, cloud admins can ensure that the firmware for the underlying hardware platform including servers, network switches, and storage controllers is compliant and update it easily when needed. This ensures the infrastructure is in a balanced state and can help prevent any disruption to service delivery. ASM provides a streamlined experience in managing resources and associated firmware with centralized management and control. As shown below, cloud administrators can log into ASM and select the firmware update. They also can schedule the update to cause minimal disruption to the VMs running in the environment. 43 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 31 ASM makes firmware management easy and managed 4.6 Metering and Chargeback In a cloud platform where multiple LOBs are using a consolidated pool of resources, having the ability to account for the utilization of all resources within the environment is critical. Foglight provides a simple costing mechanism for a cloud environment and can be configured to support a multi-tenant environment where accounting needs to be done separately for each individual LOB on the platform. Foglight s chargeback models are quite flexible. Users of the platform can be charged per host with a flat rate or by the actual resource utilization, or at a VM level at a flat rate or based on actual resource utilization per virtual machine. In the figure below we can see the charge back for the Human Resources LOB is based on a flat rate usage per host. 44 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 32 Foglight chargeback at a per host flat rate In order to charge based on resources consumed, cloud admins can create a Measure Resource Usage template and assign hosts to the template for each LOB, as shown below. 45 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 33 Measured Resource Usage template creation for consumption charges A report can then be scheduled based on the Measure Resource Usage template. VM cost is calculated periodically based on actual utilization of resources as weighted in the template definition, as shown below. Figure 34 Resource weighting for chargeback purposes 46 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Below is an example report that gets generated which breaks down the cost per VM based on the usage for a given period of time for a specific LOB. Figure 35 Cost per VM usage report 4.7 Business Continuity 4.7.1 Fast and Easy Backup and Restore Data protection is critical to ensure that workloads are protected from failures that occur in the virtual infrstructure or primary storage. Traditional backup applications require the installation of agents in each VM, which is cumbersome for adminstrators to manage. Dell vranger provides fast application, consistent whole VM backup protection for vsphere infrastructures by integrating with the vstorage API for Data Protection (VADP) and Volume Shadow Services (VSS). As shown below, vranger automatically discovers new VMs and other vsphere inventory elements when they are added to the virtual infrastructure. vranger is seemlessly integrated with the vsphere API and has the ability to create backup jobs from any vsphere element (Cluster, Resource Pool, Folder, ESXi Host, individual VM etc.) in the VMware integrated vranger inventory view. 47 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 36 vranger integrates with vsphere API and automatically discovers new VMs With vranger you can protect your entire core infrastructure pod compute density (from 100 to 1000 VMs) from a single pane of glass. To make sure all your important workloads are protected, vranger and the DR appliance can support multiple simultaneous backup write streams to meet demanding backup workloads that require high data ingest performance. The examples below show how vranger allow you to back up from any level of virtual inventory You have complete control over what is backed up and how often. Figure 37 vranger can back up a complete cluster 48 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 38 vranger can back up a specific VM vranger also integrates with vsphere Change Block Tracking (CBT) which can be combined with the proprietary Active Block Mapping (ABM) feature to minimize the amount of data transferred during backup and reduce overall backup time. Only active blocks (i.e. not idle free space or white space blocks) that have changed since the last backup are copied to the backup repository when CBT and ABM are enabled. Further backup performance improvement can be achieved by the reduction of data movement through vranger integration with the DR Appliance Rapid Data Access (RDA) library which performs source side deduplication. ABM, CBT and RDA collectively ensure that only active blocks which have changed and are unique (i.e. the block is not stored in the DR appliance yet) get copied during backup. Following screen shot shows how ABM can be enabled for a backup job. 49 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 39 Enabling ABM for a backup job 4.7.2 Disaster Protection with Replication For any cloud environment, it is critical to provide data protection in case of disaster. Your most critical data should not only be backed up and securely stored on a regular and automated basis, but should also be replicated offsite in real time to reduce chances for downtime and data loss. DR Appliance Replication provides offsite backup storage for disaster protection of physical or virtual replication targets. Admins can specify which containers should be replicated through the vranger user interface to allow for coordinated backup and replication scheduling and the replications are WAN optimized deduplicated, compressed, and securely encrypted. Figure 40 Simple, specific, and secure offsite replication with vranger and DR Appliance 50 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

4.7.3 Granular Search and Recovery To maintain compliance and to aid in recovery of lost files due to disasters or accidents, you need to be able to search for and recover at a file level. vranger has the ability to catalog the file and folder contents inside virtual disks with its Virtual Disk Cataloging and Search & File Level Restore capabilities. As shown below, an admin can use the Catalog Search and Browse page to search for files with multiple selection criteria. Figure 41 vranger Search and Browse functionality The search result items can then be selected and a file level restore (FLR) launched to automatically restore any file or folder. Figure 42 vranger file level selection and restore 51 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 43 Shows how a specific file can then be restored in vranger 52 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5 Solution Architecture The Dell Hybrid Cloud Platform architecture includes both management infrastructure and core compute infrastructure. This section describes the overall solution architecture and includes details of architectural recommendations across the management and core infrastructures that include compute, storage, and networking subsystems. The figure below illustrates the high-level scalable architecture of the basic building block (pod). The architecture has been designed with a building bock approach with clear segregation of the infrastructure: the management layer that hosts cloud management software components; data protection infrastructure that includes the backup appliance; and core infrastructure where the actual workloads/vms are hosted. Figure 44 Scalable architecture of the solution 53 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.1 Solution Requirements This reference architecture depends on certain infrastructure elements within the existing data center implementation. The following list summarizes the infrastructure elements that are required. 1. Active Directory (AD) Active Directory Domain Services are required for Dell Hybrid Cloud Platform. Users such as cloud administrator and tenant administrator are created in AD. Management components such as VMware vra, OME, and Foglight integrate into AD. 2. Domain Name Server (DNS) must be available on the management network. 3. Network Time Protocol (NTP) Server must be available on the management network. 4. SMTP is optionally utilized by other components of the management stack, including Foglight, OME and vcenter for notifications. 5. An existing Ethernet infrastructure to integrate with is required. 10Gb or 40Gb Ethernet infrastructure is recommended. Additional components, such as Dell Networking cables and transceivers, are needed to uplink the solution to the customer network. The necessary components depend upon customer networking and uplink requirements. 6. Sufficient power and cooling to support the solution must be present. 7. For integrating with public cloud and publishing the self-service portals to external users in the company and allowing access to internal resources within the Dell Hybrid Cloud Platform infrastructure, it is recommended that the data center implementation has reverse proxy servers, edge gateways, and firewalls deployed. 5.2 Core Infrastructure Pod Architecture and Design The core infrastructure pod consists of the following components: 6 x PowerEdge FX2 chassis Dual-controller Dell Storage SC4020 4 x Dell Storage SC220 enclosures 2 x Brocade 6510 switches for SAN fabric 2 x Dell Networking S4048-for LAN fabric Dell PowerEdge FX2 is a 2U hybrid rack-based computing platform that combines the density and efficiencies of blades with the simplicity and cost benefits of rack-based systems. It includes in-chassis switching that reduces ToR port usage and cabling complexity by switching the traffic among the four 2- socket blade servers and among other stacked chassis. The Dell Storage SC4020 is a highly virtualized modern storage area network that features multiple tiers of storage and automated data tiering. This solution architecture implements non-converged LAN and SAN fabrics. Brocade 6510 FC switches provide connectivity between the Dell Storage SC4020 FC storage arrays and the compute cluster. The 54 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

storage architecture and configuration are described in detail in Section 6.4. Dell Networking S4048-ON switches provide 10 GbE network connectivity between the compute cluster nodes and the rest of the data center. Section 6.5 describes the network architecture in detail. Figure 45 Scalable architecture for the core infrastructure pod While this reference architecture presents a six PowerEdge FX2 based scale-unit as the initial deployment, the overall solution has enough room to grow beyond this capacity. The Dell Networking S4048-ON and Brocade 6510 switches provide the port density to add more compute chassis. The following table describes the compute server configuration in this architecture: 55 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Table 1 Compute configuration in the solution architecture Resource Compute Nodes Processors Memory Network FC Host Bus Adapter OS Volume Description 4 x PowerEdge FC630 in each PowerEdge FX2 2 x Intel E5-2660v3 processors in each FC630 server 128 GB in each FC630 1 x Qlogic 57840 Quad-Port Blade Network Daughter Card (bndc) 1 x Qlogic QLE2562 Dual-Port 8Gbps FC adapter mapped using PowerEdge FX2 PCIe expansion slots 2 x 16 GB SD cards in the internal SD card module The figure below illustrates the end-to-end connectivity from PowerEdge FC630 in slot-1 of the PowerEdge FX2 chassis. This includes connectivity from the LAN and SAN fabric to the VM switch in the hypervisor. While this represents only the core compute infrastructure connectivity to LAN and SAN fabrics, a similar architecture is implemented for the management infrastructure components as well. The end to end I/O connectivity represented also consists of multiple port-channel and VLT configurations for LAN connectivity. 56 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 46 End-to-end network and Fibre Channel connectivity in the solution 57 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.3 Management Infrastructure Pod Architecture and Design Implementing the infrastructure that hosts the management software stack (vra, vcenter, Foglight, ASM, OME, Enterprise Manager) requires careful planning and deployment. The architecture recommended is designed as a foundation for a large scale consolidation of virtual infrastructure in the future. The architecture ensures that choices made today don t limit the size of the cloud environment that a customer can grow into over a period of time. The components described in the above subsections represent the necessary core building blocks for managing the Dell Hybrid Cloud Platform for VMware. The Solution Overview section describes the recommended software components for the management infrastructure. Most of this management software, other than Dell Cloud Manager, Dell vranger Pro, and Dell DR6000 appliance, can be implemented as virtual machines and share the storage capacity available on the Dell Storage SC4020 array used for the virtual infrastructure. Each of these management components requires connectivity to the core compute infrastructure for monitoring and management purposes. Therefore, it is recommended to connect the management infrastructure to the same LAN and SAN fabrics as the core compute infrastructure. This means you can leverage the same VLAN configuration for traffic classes between the core compute and management infrastructure. For example, the Host Management and vmotion VLAN configuration and IP address scope configuration deployed for the core compute infrastructure can be used the for the vsphere cluster nodes in the management infrastructure. The Dell Network S4048-ON and Brocade 6510 switches used in the core compute infrastructure provide the necessary port density to connect the management infrastructure. The following table provides an overview of the virtual infrastructure requirements for deploying the management components. Table 2 Management infrastructure VM sizing 58 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

In summary, 51 vcpus and 142 GB of vmemory are required to implement the virtualized management infrastructure. Dell vranger Pro and Dell DR6000 appliance are implemented as physical hosts. Dell vranger Pro uses VDDK SAN transport to improve the backup performance for the virtual infrastructure. This functionality can only be implemented on a physical host. The figure below illustrates the management architecture connecting to the existing LAN and SAN infrastructure used for the compute. Figure 47 Management architecture connectivity When virtualizing management components, administrators must ensure that the infrastructure is highly available. For storage and network the underlying core compute infrastructure provides high availability (HA). Deploying at least two physical hosts outside the core compute infrastructure is recommended for enabling complete HA of the management components. The choice of physical servers and their configuration for the management infrastructure depends on the number of management components 59 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

being deployed. PowerEdge R630 servers provide the right balance of cost and performance for virtualizing management components. The following figure illustrates how the management VMs are connected to different networks in the management vsphere cluster. Figure 48 Management VMs connected to management cluster The table below provides the specifications for the servers in the management infrastructure. Table 3 Server specifications for management infrastructure Component Virtualization Infrastructure for management software Details 2 x Dell PowerEdge R630 Processor 2 x Intel Xeon E5-2683v3 Memory OS Volume 128 GB; 8 x 16GB DDR4 DIMMs 2 x 16 GB SD cards in internal SD module 60 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Component Details Network Qlogic 57840S quad-port NDC SAN Qlogic 2562 OS VMware ESXi 6 U1 Dell vranger Pro 1 x Dell PowerEdge R630 Processor Memory 2 x Intel Xeon E5-2620 v3 Family 64 GB; 8 x 8GB DDR4 DIMMs OS Volume 2 x 600 GB 15K SAS drives in RAID 1 Network Qlogic 57840S quad-port NDC SAN Qlogic 2562 OS Windows Server 2012 R2 For deploying the management components described in this table, two PowerEdge R630 servers are used in a VMware vsphere cluster. This virtualized implementation considers 2:1 vcpu oversubscription and no memory oversubscription. By using a vsphere cluster along with HA and DRS features, the virtualized management infrastructure can be made highly available and resilient in case of a host failure. Dell vranger Pro is deployed on a PowerEdge R630 server and provides the backup and recovery functionality for the virtual infrastructure. The DR6000 appliance is a physical host that gets connected to the existing LAN infrastructure used for compute. For DR6000, Qlogic 57800 (2 x 10GbE SFP+ and 2 x 1Gb) NDC card is used to connect directly to the system hosting vranger Pro. 5.4 Storage Architecture and Configuration The following sections provide details of the storage architecture and configuration deployed in this solution. In the Dell Hybrid Cloud Platform for VMware, SAN fabric and storage arrays are shared between the core compute and management infrastructures. 5.4.1 Drive and RAID Configuration This reference architecture employs the Dell Storage SC4020 array. The array is configured with a combination of Write-Intensive (WI) SSD drives, Read-Intensive (RI) SSD drives and 10K SAS drives to provide approximately 26 TB of raw disk capacity. This hybrid deployment offers both the performance and capacity needed for the solution infrastructure. The table below provides the configuration details of the SC4020 storage array. 61 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

The choice of a hybrid drive configuration enables scaling of compute infrastructure beyond 24 nodes by providing deep I/O performance. This performance capability of the hybrid configuration can be complemented by adding more capacity in terms of Dell Storage SC220. Section 6.3 provides guidance on how the compute and storage scale out can be achieved and the considerations for the scale-out. Table 4 Component Disk Drives Storage Profile Dell Storage SC4020 storage array configuration with enclosures Description 6 x 400 GB WI SSD 6 x 1.6 TB RI SSD 12 x 1.2 TB 10K SAS drives RAID 5 with stripe width 9 RAID 6 with stripe width 10 Volumes Replay Profile Enclosures 2 x 6 TB volumes for VM Store Standard Daily Replay Profile with one week expiration 4 x SC220 storage enclosures each with 24 x 1.2 TB 10k SAS drives 5.4.2 Storage Fabric Configuration As shown in the next figure, ports 1 and 2 from each controller connect to FC switch 1, while ports 3 and 4 connect to FC switch 2. Each FC switch represents a fault domain for the FC storage and enables port failover. Brocade 6510 switches support both 8 Gbps and 16 Gbps SFP+ transceivers. In this solution architecture, 8 Gbps transceivers are used as the SC4020 storage array supports only 8 Gbps FC. 62 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 49 Dell Storage SC4020 connectivity to FC SAN fabric 5.4.3 Enabling Redundancy As shown in the figure above, fault domains are implemented to maintain continuous connectivity to stored data and the server. Fault domains are established to create redundant I/O paths. These fault domains provide for continuous connectivity with no single point of failure and without loss of bandwidth, and provide fault tolerance at the controller level. With Storage Center version 5 and above the virtual ports feature is used to implement fault domain. In addition to controller level redundancy, virtual ports can also offer port level redundancy. At a minimum, Storage Center requires two ports in the same fault domain on the same controller to support port level failover. For the compute and management VM cluster to connect to the FC fabric, Qlogic QLE2562 FC HBAs are used. For the compute infrastructure, PowerEdge FX2 architecture provides the PCIe slots that are mapped through a PCIe switch to the PowerEdge FC630 servers in the chassis. For the compute cluster connectivity, one QLE2562 adapter is used per PowerEdge FC630 server. For the management infrastructure, one QLE2562 adapter per PowerEdge R630 connects to the FC switch fabric. The two figures below illustrate how the compute and management components are connected to the SAN fabric for FC connectivity. 63 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 50 Compute Server to SAN fabric connectivity using PowerEdge FX2 PCIe slots Figure 51 Management Server to SAN connectivity For high availability of Fibre Channel connections, ports from each FC HBA are connected to two different switches. These ports along with the FC connections from the SC4020 array are configured to be in the same zone to enable storage volume access to the compute cluster. Ports from the each of the management server FC HBA are configured similarly. The figure below shows the FC volumes mapped as datastores on the ESXi hosts within the vsphere 6 cluster. 64 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 52 FC volumes mapped as data stores in vsphere cluster for compute and management Multiple data stores within the vsphere cluster enable Data store heartbeating, which ensures that partitioning or isolated host networks do not trigger VM movement within the cluster. The figure below shows the default configuration for the Data store heartbeating in the vsphere cluster. By default, the vsphere cluster selects up to five data stores for the Data store heartbeating method. Figure 53 Data store heartbeat configuration Redundant FC switches and data store configuration in the vsphere cluster ensure that the storage in the infrastructure is highly available. The redundant controllers and power supplies in the storage array along with the fault domains within the Storage Center configuration ensure that the storage can sustain component level failures and provide HA of the data stored. 65 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.5 Network Architecture and Configuration The following sections provide details of the network architecture and configuration in this solution. 5.5.1 Network Design with mvlt The network architecture in this solution employs Virtual Link Trunking (VLT) between the two ToR switches and between the IOAs in the PowerEdge FX2 chassis. In a non-vlt environment, redundancy requires idle equipment which increases switch costs. In addition, the idle equipment adds value only in the event of an equipment failure. On the other hand, in a VLT environment all paths are actively utilizing bandwidth and switches to their fullest potential. This doubles the throughput, thus increasing performance and adding immediate value. VLT technology allows a server or bridge to uplink a single trunk into more than one Dell Networking S4048-ON switch and to remain unaware of the fact that the single trunk is connected to two different switches. The switches, a VLT-pair, appear as a single switch for a connecting bridge or server. Both links from the bridge network can actively forward and receive traffic. VLT provides a replacement for Spanning Tree Protocol (STP) based networks by providing both redundancy and full bandwidth utilization using multiple active paths. Major benefits of VLT technology are: Dual control plane on the access side that lends resiliency Full utilization of the active LAG interfaces Rack-level maintenance is hitless and one switch can be kept active at all times. The mvlt architecture employed in this solution provides multiple active paths between the compute chassis and the ToR switches. mvlt is essentially a port-channel VLT between two VLT domains. Two VLT domains one between the IOAs in the PowerEdge FX2 architecture and another between the Dell Networking S4048-ON switches - exist in the network architecture. Implementing mvlt within the network architecture enables a completely loop-free layer 2 (L2) network topology while ensuring that the east-west traffic within the chassis stays in the chassis. By restricting the east-west traffic to the chassis, the available bandwidth to the ToR switches can be efficiently used. In this architecture, the total available uplink bandwidth from a PowerEdge FX2 chassis to the ToR switches is 60Gbps. For a general-purpose virtual infrastructure, the 60Gbps uplink bandwidth, along with the 10GbE non-blocking switching architecture at the ToR switches, provides sufficient bandwidth for the virtualized applications running in this solution infrastructure. 5.5.2 Compute Infrastructure LAN Configuration The two Dell PowerEdge FN410S IOAs in the PowerEdge FX2 architecture provide the top-of-rack connectivity for the PowerEdge FC630 servers. Each IOA provides eight internal server facing ports and four external ports. FN IOA provides the low-touch configuration mode called VLT mode in which Port 9 from each IOA in the chassis form a Virtual Link Trunk Interconnect (VLTi). Ports 10, 11, and 12 form portchannel 128, which in turn connects to the ToR switches. In this architecture, Ports 10 and 11 from FN IOA 66 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

1 connect to TOR 1 while Port 12 is connected to TOR 2. On FN IOA 2, Ports 10 and 11 connect to TOR 2 while Port 12 is connected to TOR 1. The following figure illustrates this connectivity. Figure 54 Server LAN configuration Since Ports 10, 11 and 12 from each IOA are connected to two different Dell Networking S4048-ON switches, a VLTi between the S4048-ON switches is mandatory. This provides complete redundancy for the network traffic from the chassis to the ToR switches. These ports are combined using port channel configuration and enable up to 60 Gbps of bandwidth between the FX2 architecture and the ToR switches. As shown in the figure above, two 40 GbE ports from each Dell Networking S4048-ON switch are connected together to create a VLT. This provides up to 80 GbE bandwidth between the ToR switches. The remaining four 40 GbE ports on each ToR switch can be used to connect to the data center core network. This ToR configuration uses the management ports on both the switches for the VLT backup link. 5.5.3 Management Infrastructure LAN Configuration For the management servers, the PowerEdge R630s are connected directly to a Dell Networking S4048- ON switch through the Qlogic 57840S quad-port NDC, as shown in the figure below. Figure 55 Management server connectivity 67 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

As shown in the above figure, the servers hosting vsphere cluster for management VMs and the DCM deployment connect to the ToR switches using 4 x 10GbE ports. Refer to the data protection section for an overview of how the vranger Pro and DR6000 are connected to the LAN fabric. 5.5.4 NPAR and Hypervisor Network Configuration for LAN Traffic Each Dell PowerEdge FC630 server in the PowerEdge FX2 architecture and the PowerEdge R630 management servers in the management vsphere cluster provide four 10 GbE ports through a Qlogic NDC. To ensure the bandwidth prioritization for different traffic classes such as host management, vmotion and VM network, switch independent network partitioning (NPAR) is configured. By using NPAR, each port on the NDC is partitioned into two logical partitions. The Qlogic NDC adapters allow setting a maximum bandwidth limitation to each partition. Setting maximum bandwidth at 100 will prevent the artificial capping of any individual traffic type during periods of non-contention. For customers with specific requirements, NPAR maximum bandwidth settings may be modified to limit the bandwidth available to a specific traffic type, regardless of contention. These adapters also allow relative bandwidth assignments for each partition. The table below shows the NPAR schema and the relative bandwidth assignments on each partition. Table 5 NPAR schema and relative bandwidth weights bndc Port Port 0 Port 1 Port 2 Port 3 Network Partition Label Traffic Class NIC 0 Host Management NIC 4 vmotion 70 NIC 1 Host 30 Management NIC 5 vmotion 70 NIC 2 VM Network 40 NIC 6 NA 60 NIC 3 VM Network 40 NIC 7 NA 60 Relative Bandwidt h 30 Maximum Bandwidt h 100 The figure below illustrates how NPAR is deployed on each physical host. As shown below, network partitions NIC6 and NIC7 are not deployed to carry any host or VM network traffic. The relative bandwidth constraints are used only when there is a need for a traffic class to be given priority. 68 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 56 Management and core compute host LAN configuration using NPAR The figure below shows the host-level implementation of virtual switches as illustrated in the Host LAN diagram. 69 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 57 vswitch implementation on every physical host (management and core compute) The load balancing algorithm for the virtual port groups is left to default setting which is Route based on originating virtual port. Customers may consider changing this configuration based on the workload that is deployed on the virtualized infrastructure. 5.5.5 VLAN Configuration for LAN Traffic and Traffic Isolation for Different Tenants For isolating different traffic classes from a host, Virtual LAN (VLAN) is deployed. VLANs are configured in the port groups of the virtual switches for each traffic type. The virtual LAN configuration is needed on the FN IOA and ToR switches to ensure end-to-end connectivity. Cloud administrators can also use VLANs to segregate traffic from different tenants. Port groups are created in the virtual switch for each tenant and each port group is configured with a unique VLAN. Through vra, virtual machines for each tenant are provisioned and configured to their corresponding port group. The Dell Networking S3048 1 Gb switch in this architecture is an optional component and can be replaced by an existing out-of-band (OOB) management switch in the customer s existing data center environment. The network architecture design described in this section implements key design principles and considerations to provide a highly available LAN fabric that is optimized for virtual and enterprise application traffic. 70 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.6 Data Protection, Replication, and Recovery Architecture The Dell Data Protection solution for the Dell PowerEdge FX2 converged system uses the DR appliance with Rapid Data Access (RDA) source side deduplication technology and the vsphere VDDK SAN Transport. The combination of RDA inline deduplication and VDDK SAN transport provides a premium data movement service that is capable of satisfying demanding backup workloads at scale. Additionally, leveraging SAN for backup transport avoids taxing virtualization resources for backup purposes, thereby conserving hypervisor capacity for production workloads. For backup, the solution leverages the deep VMware integration of Dell vranger to acquire virtual disk read access and SAN transport via vsphere. After attaining virtual disk access, vranger reads the virtual disk blocks and negotiates with RDA inline deduplication to determine if the block is unique. Only unique blocks are copied to backup storage. Additional space savings techniques such as vranger Active Block Mapping and VMware Changed Block Tracking can be enabled to further reduce the data transport volume. Recovery is performed using RDA to read the Recovery Point stored in the DR appliance and vsphere to establish a virtual disk write access SAN transport. For Disaster Recovery, the DR Appliance provides the capability to replicate from the DR appliance across the WAN to a remote physical or virtual DR appliance residing in either an enterprise network or in a public cloud. The figure below shows the data protection architecture. Figure 58 Data protection architecture 71 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

A Dell PowerEdge R630 server is used to deploy Dell vranger Pro backup software running on Windows Server 2012 R2. This server is connected directly to a Dell DR6000 appliance using 2 x 10GbE SFP+ ports. This provides a private network between the vranger Pro system and the DR6000 appliance for backup traffic. This architecture ensures that the backup duration for the compute infrastructure can be reduced while keeping the core LAN infrastructure free from all backup traffic. The figure below illustrates this architecture. Figure 59 Backup architectures Within the data protection infrastructure, only the PowerEdge R630 hosting Dell vranger Pro is connected to the FC fabric. As described earlier, this connectivity is used for the VDDK SAN transport when protecting the compute infrastructure built on PowerEdge FX2 architecture. 72 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

5.7 Sizing and Scaling Recommendations 5.7.1 Sizing Recommendations Initial Deployment of Core Infrastructure Pod The six PowerEdge FX2 based scale-unit for the core compute infrastructure can be used to deploy approximately 1,000 virtual machines each with 2 vcpus, 4 GB memory, and 40 GB VMDK. Considering an I/O profile in which each VM requires 30 IOPS, the storage backend must be capable of sustaining 30,000 IOPS during peak load. The drive configuration of Dell Storage SC4020 used in this reference architecture can support more than 50,000 sustained IOPS. Therefore, with a 4:1 vcpu oversubscription and 1.5:1 memory oversubscription for the virtual machines, the six PowerEdge FX2 based scale-unit will be capable of running approximately 1,000 virtual machines. To scale out the core compute infrastructure, additional scale-units containing six PowerEdge FX2 chassis and LAN and SAN fabric switches can be added. As described in the Management Infrastructure Pod Design and Architecture section, some of the management components in this solution infrastructure can be deployed as virtual machines. To host these virtual machines, the management infrastructure pod includes two PowerEdge R630 servers in a VMware vsphere cluster. The solution specification for these virtual cluster hosts and the specification for each management virtual machine is capable of supporting up to 3,000 virtual machines. For example, the configuration used for the vcenter Server Appliance in the management infrastructure pod is capable of managing up to 100 physical hosts and up to 3,000 virtual machines. The vrealize Automation appliance with 4 x vcpu and 16 GB of memory is capable of supporting up to 10,000 virtual machines. These configuration maximums enable the initial compute infrastructure pod to scale out to an infrastructure of up to 3,000 VMs without changing the management infrastructure. 5.7.2 Sizing Recommendations- Adding More Core Infrastructure Pods For the initial core compute infrastructure pod, the architecture for the management virtual machines leverages HA provided by the vsphere cluster. However, as more core compute infrastructure pods are added, it may become necessary to add more capacity to the management vsphere cluster. While the core compute LAN and SAN infrastructure HA and management cluster HA provide some level of failover for the services, it is recommended to scale-out the management components where supported. For example, vcsa supports dividing the services in Platform Services Controller and vcenter Server services and hosting them on different virtual machines running behind a load balancer. These virtual machines can then be hosted across multiple physical servers to ensure that one host failure does not degrade or impact the services offered. Also, for the initial core compute infrastructure, the vrealize automation server is configured to use the embedded vrealize Orchestrator. This supports up to 100 concurrent workflows. When scaling out to support more virtual hosts and virtual machines, it is necessary to deploy an external vrealize Orchestrator instance and configure vrealize automation services to use the external orchestrator. For Single-Sign on (SSO), the management infrastructure leverages the SSO services offered by Platform Services Controller (PSC) in vcsa. As mentioned earlier, to support more virtual machines or higher 73 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

authentication requests than in a smaller pod, it is necessary to deploy multiple load-balanced instances of PSC across different physical hosts. Starting with a non-ha (at the service level) configuration for these management virtual machines and scaling-out these VMs to enable server level may require complex reconfiguration. To eliminate this, it is recommended to build these infrastructure components with future scale-out in mind. For all service endpoints, using a generic endpoint DNS name other than the host name is recommended. For example, consider that the vrealize Automation host name is vra1-vcloud.financecloud.local. This hostname can be mapped to a DNS alias pointing to vra.financecloud.local and the alias (CNAME) can be used in all component configurations. The DNS alias name can be assigned to a load balancer that is used to load balance vra services. This eliminates any need for service reconfiguration and makes scaling out of the management infrastructure to support more compute infrastructure pods seamless. When adding capacity for running more management virtual machines, additional physical hosts running VMware vsphere 6.0 can be added to the existing vsphere management cluster. To scale out beyond three core compute infrastructure pods, adding two PowerEdge R630 servers with similar specification as the existing management vsphere clusters is recommended. The figure below shows how infrastructure can be scaled for rapid growth in the environment. Figure 60 Add pods for rapid growth 5.7.3 Scaling Out The Networking of the Solution The solution can be scaled by adding multiple Dell Hybrid Cloud Platform (HCP) pods in the customer data center. The Dell Networking S6000 switch can be used to create a simple yet scalable network. The S6000 switches server as the spine switches, in the leaf-spine architecture. Each S6000 can support 32 40GbE ports and could aggregate multiple racks with little or no over-subscription. When connecting multiple HCP racks, using the 40GbE uplinks from the rack, you can build a large fabric that supports multi-terabit clusters. The density of the S6000 allows flattening the network tiers and creating an equalcost fabric from any point to any other point in the network. 74 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 61 Multiple infrastructure pods can be scaled out using the leaf spine architecture For large domain layer-2 requirements the extended VLT (evlt) can be used on the S6000, as shown in the figure below. The VLT pair formed can scale in terms of hundreds of servers inside multiple hybrid cloud racks. Each hybrid cloud rack has 4 40GbE links to the core network providing enough bandwidth for all the traffic destined from and to each rack. Figure 62 Extended Virtual Link Trunking 75 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

The S6000 can be connected to a Dell Networking Z9000 as the core switch. Large scale networks routinely deploy the Z9000 in HPC and cloud environments. Compared to a chassis (modular) switch, the density available on Z9000 gives the best ROI in terms of switching capacity, form-factor (2U), and power measured in fractions compared to modular chassis. Supported by the FTOS operating system that has features suitable for high end switches and routers, the Z9000 becomes a great next-generation choice for a core switch. 76 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

6 Component Specifications This section provides a list of the hardware and software components used in the solution, which include VMware vsphere 6.0 on the PowerEdge FX2 architecture with Dell Networking S4048-ON switches and Dell SC4020 Storage. The table below provides a list of components needed to build a virtualized infrastructure for the cloud reference architecture as described in Section 5. Table 6 Hardware specification for 6 x PowerEdge FX2 architecture deployment Component Virtualization Infrastructure Virtualization Hosts Details 6 x Dell PowerEdge FX2 12 x FN410S; 10GbE SFP+ IOA (2 per PowerEdge FX2) 24 x Qlogic QLE 2562 DP PCIe Add-on FC HBA (4 per PowerEdge FX2) 24 x Dell PowerEdge FC630 Servers ( 4 per PowerEdge FX) Processor Memory OS Volume Network OS 2 x Intel Xeon E5-2660v3 Family 128 GB; 8 x 16GB DDR4 DIMMs 2 x 16 GB SD cards in internal SD module Qlogic 57840 quad-port NDC VMware ESXi 6 U1 Storage Arrays Networking Cable Requirements Dell Storage SC4020 with 6x 400 GB (WI) SSD, 6 x 1.6 TB (RI) SSD and 12 x 10K SAS drives 4 x Dell Storage SC220 expansion attached with 24 x 1.2 TB 10K SAS drives 2 x Dell Networking S4048-ON for Local Area Networking (LAN) 2 x Brocade 6510 FC switches 1 x Dell Networking S55 for OOB network (optional) 42 x 10GbE SFP+ DAC cables 48 x LC-LC FC cables 2 x 40GbE QSFP+ cables 77 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

6.1 Hardware Component Details for Core Infrastructure The table below provides specification of hardware components of the infrastructure pod. Table 7 Specification of components used in this reference architecture Component Virtualization Infrastructure Virtualization Hosts Details Dell PowerEdge FX2 PowerEdge FN410S; 10GbE SFP+ IOA Qlogic QLE 2652 DP PCIe Add-on FC HBA for storage connectivity Dell PowerEdge FC630 Servers Processor Memory OS Volume Network 2 x Intel Xeon E5-2660v3 Family 128 GB; 8 x 16GB DDR4 DIMMs 2 x 16 GB SD cards in internal SD module Qlogic 57840 quad-port NDC OS VMware ESXi 6 Storage Arrays Dell Storage SC4020 with 6x 400 GB (WI) SSD, 6 x 1.6 TB (RI) SSD and 12 x 10K SAS drives Infrastructure Deployment Dell Active System Manager (ASM) 8.1 6.2 Hardware Component Details for Management Infrastructure The table below provides the specifications for the servers in the management infrastructure. Table 8 Server specifications for management infrastructure Component Virtualization Infrastructure for management software Details 2 x Dell PowerEdge R630 Processor 2 x Intel Xeon E5-2683v3 Memory OS Volume Network 128 GB; 8 x 16GB DDR4 DIMMs 2 x 16 GB SD cards in internal SD module Qlogic 57840S quad-port NDC SAN Qlogic 2562 OS VMware ESXi 6 U1 78 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Component Dell vranger Pro Details 1 x Dell PowerEdge R630 Processor Memory 2 x Intel Xeon E5-2620 v3 Family 64 GB; 8 x 8GB DDR4 DIMMs OS Volume 2 x 600 GB 15K SAS drives in RAID 1 Network Qlogic 57840S quad-port NDC SAN Qlogic 2562 OS Windows Server 2012 R2 6.3 Firmware Versions of Components The tables below show the recommended firmware versions for the components of this reference architecture. Table 9 Device PowerEdge FC630 Servers Version BIOS 1.1.9 CPLD 1.0.0 idrac8 Enterprise 2.10.10.10 (49) Life Cycle Controller (LC) 2 2.10.10.10 PERC H730 Mini 25.2.2-0004 Network Controller QLogic BCM57840S 7.12.15 QLogic BCM57840S Driver 7.12.2.0 Qlogic QLE2562 8Gb FC 03.22.00 Table 10 Device Dell CMC PowerEdge FX2s Chassis, Compellent SC4020 Storage, and Switches Version 1.20.A00.201502058 Dell FN410S IOA 9.6 (0.0) 79 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Device Version Dell S4048-ON 9.8.0.0 Brocade 6510 v7.2.0a PERC H730 Mini 25.2.2-0004 Dell Storage SC4020 6.5.20 6.4 Software Components The tables below list the primary software components for the reference architecture. Embedded and required element manager software elements for the Dell hardware components are not shown. Table 11 Dell software components Dell Software Version Notes Dell Active System Manager 8.1.1 Unified management and automation software Dell ASM Plugin for VMware ASM Plugin to vrealize Orchestrator. Enables vrealize 1.0 vrealize Orchestrator Automation integration. Dell Foglight for Virtualization, Monitoring for virtual infrastructure. This version 8.3 Enterprise Edition includes integration with OME. Dell Open Manage Essentials 2.0.1 Monitoring for physical infrastructure Dell Enterprise Manager 2015 1.10 Storage management console Dell vranger Pro 7.3 Backup server for virtual infrastructure. Optional Dell Software: Dell OpenManage integration for VMware vcenter Dell Storage vsphere Web Client Plugin 3.0 3.0 Plugin to vcenter Server for server monitoring and configuration from vsphere client Plugin to vsphere for storage monitoring and configuration from vsphere client Table 12 VMware software components VMware Software Version Notes VMware vcenter Server 6.0 vsphere management server VMware vsphere ESXi 6.0 Server hypervisor VMware vrealize Automation 6.2.1 VMware cloud management and infrastructure, service catalog VMware vrealize Orchestrator 6.0.1 Orchestration engine 80 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

7 Solution Verification Dell Solution Engineering has tested and validated each of the configurations referred in this reference architecture. The validation entails: Interoperability between each of the components in the architecture Guidance on the firmware version used for the testing Example test scenario that provides guidance on sizing The solution design in this paper employs best practices and recommendations at all infrastructure components. Any solution infrastructure must be verified for component level functionality before it can be used in production. This should involve verification of infrastructure component failover to ensure that the services offered by the solution infrastructure are highly available and resilient against component failures. The solution infrastructure described in the paper was verified to ensure that the component level highavailability and resiliency are indeed available. There was no user-perceived downtime of any services running inside the virtual machines in the infrastructure during the component failover. 7.1 Storage Verification For a general purpose virtualization infrastructure it is important to ensure that each component within the infrastructure is appropriately sized to run the virtualized applications. One of the unique aspects of the reference architecture is that it adopts a single storage architecture regardless of scale. Hence, performance of SC4020 plays a major role in the overall solution performance and influences the number of VMs that can be deployed on this infrastructure. While it is not practical to size every type of application that may run on a general purpose infrastructure as described in this paper, it is still important to determine the overall performance that the storage subsystem can offer. To this extent, the storage verification of the solution infrastructure was carried out to understand how Dell Storage SC4020 performs. An Online Transaction Processing (OLTP) I/O load profile was simulated using IOMeter. The table below provides the I/O profile details. Table 13 I/O profile for storage verification I/O Profile Type Read Write Access Size IO Pattern Database OLTP 70% 30% 8K 100% Random For IOMeter verification, two VMs per cluster node were created and the storage verification tests were performed within the VM. The table below provides the VM configuration used for the IOMeter verification. Table 14 VM configuration for IOMeter verification VMs per Host vcpu / VM Memory / VM OS VMDK Data VMDK 2 20 60 GB 40 GB 40 GB 81 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

As shown in the table above, a 40 GB thick provisioned VMDK was used for the IOMeter verification. These VMDK files were stored across eight 750 GB volumes created on Dell Storage SC4020. IOMeter verification was performed with varying queue depth values. This verification was performed on a six PowerEdge FX2 architecture containing a total of 24 PowerEdge FC630 servers. The table below provides the data generated from the IOMeter runs on the six PowerEdge FX2 architecture. Table 15 IOMeter performance data from six PowerEdge FX2 architecture (48 workers) Queue Depth 1 4 16 IOPS 23714 43790 47992 MBps 185 342 374 Read MBps 129 239 262 Write MBps 55 102 112 Transactions per Second 23714 43790 47992 Average Response Time 1.93 4.2 15.33 Average Read Response Time 2.42 5.56 11.59 Average Write Response Time 0.8 1.01 24.06 Average Transaction Time 1.93 4.2 15.33 The storage performance data shown above clearly shows that the storage subsystem in this architecture has enough IOPS and disk capacity to scale beyond an eight-node deployment. Let us look at an example: The solution design assumes one VM per logical processor core equivalent on each workload server. Further, other Dell lab tests show that the SC4020 can deliver 129,661 Oracle 100% read OLTP IOPS with less than 1ms latency 2, further showing that this test didn t fully saturate the SC4020 array. However, it is always recommended to size the storage infrastructure by using a real-world workload scenario that is planned for deployment. The six PowerEdge FX2 based scale-unit can be used to deploy approximately 1,000 virtual machines each with 2 vcpus, 4 GB memory and 40 GB VMDK. Considering an I/O profile in which each VM requires 30 IOPS, the storage backend must be capable of sustaining 30,000 IOPS during peak load. The drive configuration of Dell Storage SC4020 used in this reference architecture can support more than 50,000 sustained IOPS. Therefore, with a 4:1 vcpu oversubscription and 1.5:1 memory oversubscription for the virtual machines, the six PowerEdge FX2 based scale-unit will be capable of running approximately 1,000 virtual machines 2 Based on Dell internal testing at Dell labs using Oracle ORION toolkit in March 2014. Performance shown indicative of 12 WI drives. Oracle workload performance will vary depending upon workload I/O profile and SC4020 drive configuration, usage and manufacturing variability. 82 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

8 Solution Summary Dell Blueprint solutions for cloud provide a faster, simpler path to cloud deployment and operation that provide lower TCO and the flexibility to adapt your cloud to future requirements. With Dell, you can cut your cloud deployment time and operating costs and respond faster to business demands by matching the right cloud design to the right application. The Dell Hybrid Cloud Platform Reference Architecture with VMware is a fully-customizable, pre-validated reference architecture that provides prescriptive guidance for building your cloud. It is designed to help you and your organizations improve and align business and IT agility, gain efficiencies and optimize service delivery, and enhance control. Dell engineering teams developed this reference architecture based on direct customer input. They have done the sizing, testing, and validation to help remove the complexity so you can deploy more quickly and with confidence, while maintaining the flexibility to tailor the solution to your needs. To match your costs with actual deployment and utilization, Dell provides scale-ready payment systems that allow you to control how and when you pay based on forecasted usage, deployment schedule, or actual usage. This architecture is highly scalable, allowing you to meet your needs today and easily grow over time. You can grow within a pod or add additional pods. It utilizes a common, modular infrastructure architecture regardless of scale. This means no change in core technology or added complexity as your infrastructure grows. The architecture is also highly integrated, which enables significant automation, simplified deployment and provisioning, advanced monitoring and proactive resolution capabilities, and high availability and data protection features. The Dell Hybrid Cloud Platform Reference Architecture with VMware is ideal for private and hybrid cloud implementations where scale without compromise, control without complexity, and efficiency without constriction is desired. In summary, this reference architecture integrates the best of Dell and VMware products. With Active System Manager s integration to vcenter and vrealize, integrated physical and virtual monitoring, and built-in data protection, the reference architecture provides a fully featured, deeply automated, and highly optimized private and hybrid cloud platform with end-to-end automation capabilities from the physical to the virtual to the cloud. 83 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

A Additional Resources A selection of additional resource is shown below. 1. Support.dell.com is focused on meeting customer requirements with proven services. 2. Build Your VMware Cloud contains all the VMware private and hybrid cloud resources on Dell.com. 3. Dell TechCenter is an IT Community where you can connect with Dell customers and Dell employees for the purpose of sharing knowledge, best practices and information about Dell products and installations. 4. Dell Cloud Blueprint on Dell TechCenter is where the current version of this reference architecture can be found. 5. Dell Active System Manager on Dell TechCenter provides information on ASM, including links to resources, technical documents, and videos 6. Referenced or recommended Dell publications: a. Dell PowerEdge FX2 b. Dell Storage SC4020 c. Dell Storage SC220 7. Referenced or recommended VMware publications or websites: a. New features in VMware vsphere 6 b. VMware vrealize Automation c. VMware vrealize Orchestrator 84 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

B Compute System Component Details The following sections provide the technical and product overview of the compute configuration options available with PowerEdge FX2 architecture. B.1 Dell PowerEdge FX2 Architecture PowerEdge FX2 is a 2U rack mount chassis and accommodates PowerEdge FC630, PowerEdge FC430, PowerEdge FC830, PowerEdge FM120x4 server sleds and PowerEdge FD332 storage sled. A sled is a pluggable component in a modular architecture similar to that of PowerEdge FX2. The PowerEdge FX2 architecture offers flexibility in terms of infrastructure compute, storage and networking. This table presents an overview of all component configurations supported in PowerEdge FX2 architecture. Table 16 Dell PowerEdge FX2 supported components overview Feature Server Compatibility Form Factor Number of Server Sleds Number of Storage Sleds I/O Power Supplies Management Network Description PowerEdge FC630/FC430/FC830/FM120x4 servers PowerEdge FD332 storage 2U rack enclosure Up to four FC630 Up to eight FC430 Up to two FC830 Up to four FM120x4 Up to four FD332 8 PCIe slots (supporting Ethernet and FC) Up to 2 x 1600w PSU (FC630 and FC430) Up to 2x 1100w PSU (FM120x4) 1 Chassis Management Controller 2 x pass-through I/O Modules (IOM; 1 GbE or 10 GbE) 2 x 10 GbE SFP+ IO Aggregator (IOA; FN410s) 2 x 10 GbE 10 Base T IO Aggregator (IOA; FN410T) 2 x FC and 10 GbE combo IO Aggregator (IOA; FN2210S) The server connectivity in a PowerEdge FX2 architecture can be extended by adding supported PCIe peripheral cards. There are eight PCIe slots available for this purpose and they are internally mapped to servers. The following figure shows the internal mapping of the PCIe slots to servers in the PowerEdge FX2 architecture: 85 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Figure 63 PCIe slot mapping in PowerEdge FX2 (with half-width servers) The PowerEdge FX2 architecture is managed through the Chassis Management Controller (CMC). The CMC Web console provides features to manage servers and IOMs and monitor PCIe slots in a PowerEdge FX2 architecture. The following figure shows the group management of multiple PowerEdge FX2 architecture in the CMC web console: Figure 64 PowerEdge FX2 CMC web console - Group Management Using the group management feature of CMC it is possible to manage up to 20 PowerEdge FX2 architectures from the same CMC console. This can be beneficial when there are multiple chassis participating in an infrastructure deployment such as the virtualization infrastructure implemented in this reference architecture. 86 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

B.1.1 Dell PowerEdge FC630 Server PowerEdge FC630 servers support Intel Haswell processors up to 18 processing cores and up to 768 GB physical memory. This enables PowerEdge FC630 to deliver the performance required for computeintensive tasks. A Qlogic 57810 or Qlogic 57840 10 GbE blade Network Daughter Card (bndc) used within the server provides the network connectivity through the 10 GbE pass-through IOM or 10 GbE IOA in the PowerEdge FX2 architecture. PowerEdge FC630 servers use blade mezzanine cards to connect to the PCIe add-on peripherals in the PowerEdge FX2 architecture. This connectivity can be leveraged to build optional redundancy for network connections. For operating system (OS) drives, PowerEdge FC630 supports a wide variety of PowerEdge RAID controller (PERC) devices. PowerEdge FC630 servers in a PowerEdge FX2 architecture provide the necessary compute required for the virtual infrastructure. This server strikes a balance between the density provided by PowerEdge FC430 and the capacity and performance provided by PowerEdge FC830. PowerEdge FC430 servers scale memory up to 768 GB (unlike FC430 where the maximum supported memory is 256 GB) when needed. By using four FC630 servers in PowerEdge FX2 architecture, the solution architecture enables the same computing capacity offered by two FC830 servers in PowerEdge FX2 while ensuring that there is greater physical isolation for virtual applications. Therefore, PowerEdge FC630 is an ideal choice for the virtualized architecture presented in this paper. For an overview of PowerEdge FC430, PowerEdge FC830 and PowerEdge FM120x4, refer to Appendix C. The following table describes the components supported in a PowerEdge FC630 server: Table 17 CPU Supported components in PowerEdge FC630 Supported in PowerEdge FC630 Up to 2 Intel Xeon E5-2600 family processors Memory 24 DIMMs; up to 768 GB 3 Networking Storage Form Factor PCIe slots Qlogic 57810 Dual Port 10 GbE bndc Qlogic 57840 Quad Port 10 GbE bndc Emulex 10GbE adapters PowerEdge Express Flash NVMe PCIe SSD, SATA HDD/SSD or SAS HDD/SSD Up to 8 x 1.8 SSD or 2 x 2.5 ½ width Sled 2x PCIe slots (x8) provided by PowerEdge FX2 3 Some processor heat sink configurations may not support 768 GB. Refer to the product manual for more information. 87 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Systems Management idrac8 Express or idrac8 Enterprise B.1.2 Dell PowerEdge FN410S As mentioned in Table 11, PowerEdge FX2 supports multiple networking options for server connectivity to the top-of-rack (ToR) switches and/or to the data center core. These networking options provide flexible choices when designing a solution infrastructure. This reference architecture employs PowerEdge FN410S for the ToR connectivity from the PowerEdge FX2 chassis. Using the IOA configuration simplifies connectivity by as much as 8-to-1, greatly reducing cabling complexity. The FN IOA is a plug-n-play networking device providing a low-touch experience for server administrators. Most IOA initial setup and networking functions are automated making for minimal touch for basic to advanced features. B.2 Dell Networking S4048-ON Switch Dell Networking S4048-ON is a 1U high-density 10/40 GbE ToR switch with 48 dual-speed 1/10GbE (SFP+) ports and six 40 GbE (QSFP+) uplinks. This switch leverages a non-blocking and cut-through switching architecture to provide ultra-low-latency performance for applications. The six 40 GbE ports can be used to create either a Virtual Link Trunk (VLT) between the switches to enable traffic isolation within the solution infrastructure or as a connectivity to the data center core network. The Dell S4048-ON supports the open source Open Network Install Environment (ONIE) for zero-touch installation of alternate network operating system including feature rich Dell Networking OS and also supports Dell Networking s Embedded Open Automation Framework, which provides advanced network automation and virtualization capabilities for virtual data center environments. B.3 Dell Storage SC4020 Array Dell Storage SC4020 belongs to the Storage Center (SC) 4000 series arrays based on the SC8000 platform. This storage array offers multi-protocol support and virtualized multi-tier, multi-raid-level storage policies. Each SC4020 array comes with dual redundant controllers, 24 internal drive slots, eight 8 Gb Fibre Channel (FC) or four 10 Gb iscsi network ports and one 1 Gb port per controller for out-of-band (OOB) management traffic. The supported drive types range from Write-Intensive (WI) Solid State Drives (SSDs) to Read-Intensive (RI) SSDs to 15K, 10K, and 7.2K SAS drives in small form factor (2.5-inch). This storage array supports multi-tier data placement that improves application performance. This array supports expansion up to 192 drives by adding Dell Storage SC220 or SC200 enclosures. Using the largest capacity disk drives supported a Dell Storage SC4020 array can support up to 1 Peta-Byte (PB) of raw disk capacity. The 32-GB controller cache (16 GB per controller) provides the necessary storage performance for most enterprise application needs. The Storage Center 6.5 Operating System provides features such as block-level compression, synchronous Live Volumes to restore data in a non-disruptive manner and Active Directory (AD) Single- 88 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware

Sign-On (SSO). The Storage Center Manager enables easier and out-of-box web-based management of SC4020 arrays while the Enterprise Manager and its components can be leveraged to build a secure, multi-tenant environment with the Dell Storage Center SAN. B.4 Brocade 6510 Brocade 6510 is a 48-port 1U Gen 5 FC switch that is suitable for high-performance data requirements of server virtualization, cloud and enterprise applications. This switch can be configured in 24, 36, or 48 ports and supports 2, 4, 8, 10, or 16 Gbps speeds. This enables enterprises to start small and scale the FC infrastructure based on the growth, as needed. By using features such as Virtual Fabrics, Quality of Service (QoS) and zoning, IT departments can build multi-tenant cloud environments. Management and diagnostics features such as Monitoring and Alerting Policy Suite (MAPS), Dashboards, Flow Vision, Fabric Performance Impact (FPI) monitoring and Credit Loss Recovery help administrators and IT organizations avoid problems before they impact the SAN operations. B.5 VMware vsphere 6 VMware vsphere 6 is an industry leading server virtualization platform that enables creation of a highly available, resilient, and on-demand infrastructure for virtualizing enterprise applications or as a general purpose virtual infrastructure. vsphere 6 provides improved features that are available to enterprises looking at data center consolidation and enhancing resource utilization. Apart from the scalability improvements, vsphere 6 provides enhanced vsphere Fault Tolerance support for multi-vcpu VMs, compression support for vsphere Replication, and added support for multisite content library. The table below provides an overview of the capabilities in vsphere 6. Table 18 VMware vsphere 6 cluster capabilities VMware vsphere 6 VM Hardware Version Virtual Hardware 11 Maximum Number of vcpu per VM Maximum Memory per VM 128 4 TB Cluster Nodes 64 Maximum CPU per Host 480 Maximum Memory per Host 12 TB Maximum VMs per Cluster 8000 The virtualization capabilities offered by vsphere 6 along with the management capabilities offered by vcenter enable building an efficient platform for virtualizing enterprise applications. 89 Reference Architecture Dell Hybrid Cloud Platform Reference Architecture with VMware