Demystifying Disaster Recovery as a Service Rachel Dines, Senior Analyst, Forrester Research Inc. Alex Foster, Product Manager, Windstream 2011 Windstream Communications, Inc. 008906
Introduction Rachel Dines, Senior Analyst, Forrester Research Rachel serves Infrastructure & Operations Professionals. Her research focus is on IT continuity and disaster recovery services and technologies, next-generation high availability and backup, and data center strategies. Additionally, Rachel conducts research on infrastructure and operations metrics. Alex Foster, Product Marketing Manager, Windstream Alex Foster is the Product Manager for Cloud, Storage & Backup solutions at Windstream Hosted Solutions. Prior to joining Windstream, Alex was the lead product manager for PAETEC's cloud portfolio and played a key role in expanding PAETEC's data center services portfolio to include Infrastructure-as-a-Service (IaaS) offerings.. 2
Agenda The state of disaster recovery today Enter a new model: cloud-based disaster recovery Barriers and benefits to cloud-based disaster recovery Recommendations 3
What s driving the need to run 24/7/365? Competitive advantage Downtime creates an opportunity for competitors to seize market share. Employee and customer expectations Downtime impacts confidence and retention Cost of downtime This includes revenue loss, employee productivity loss, fines, penalties, reputation impact Accountability Including fiduciary responsibility to shareholders and other key stakeholders Supplier expectations Suppliers expect agreements and payments to be nonstop 4
The disaster recovery challenge BC/DR budgets are 5.5% of IT opex/capex Less budget allocated to BC/DR Data explosion Capacity requirements are still growing 20%-40% per year Business owners have less and less tolerance for any data loss 5 No tolerance for data loss More complexity and heterogeneity Increasing recovery demands 25% of servers are non-windows OSes More and more companies operate close to 24x7
Agenda The state of disaster recovery today Enter a new model: cloud-based disaster recovery Barriers and benefits to cloud-based disaster recovery Recommendations 6
Recovery objectives The gap in traditional DR services Seconds Data Loss Synchronous Replication Minutes Hours Asynchronous Replication Gap Hot Sites, Warm Sites Dedicated IT equipment Days $ Recovery from tape Recovery from disk Cold Sites Shared IT equipment $$ This gap can be filled with virtualized and cloud solutions $$$$ 7 DR services cost
Cloud-based DR can be defined in three categories Cloud-based DR Do it yourself cloud-based DR Using the public cloud to architect a custom solution leveraging the agility and speed of the cloud. Cloud-to-cloud DR The ability to failover services from one cloud data center to another DR-as-a-Service Pre-packaged solutions that provide failover to a cloud environment 8
Do it yourself cloud-based DR Production data center A B C D E F Public cloud provider Replication A B C D E F Physical server 1 9 Physical server 2 Primary Storage Physical server 3 Failover is manual, requires skilled staff. Public cloud provider does not usually guarantee any capacity when needed nor will they assist in the failover. If physical servers are being protected, customer must manage the conversion to virtual.
Disaster recovery-as-a-service Production data centers Service provider deploys agents to replicate data and applications to the cloud. Physical machines are converted to VMs to boot in the cloud. Primary Storage DRaaS provider 10 Primary Storage Managed backup appliance Backup service provider uses managed backup to the cloud and converts backups of physical or virtual machines into bootable VMs in the cloud.
Replication can occur at the storage, operating system, application, or hypervisor layer Production data center DRaaS provider Application replication Hypervisor replication Host replication SAN replication Primary Storage Primary Storage 11
Two-thirds of companies show interest in DR-asa-service What are your firm s plans to adopt IT-recovery-as-a-service based on virtual infrastructure at the service provider? Expanding/upgrading implementation Implemented, not expanding Planning to implement in the next 12 months Planning to implement in a year or more 5% 6% 6% 9% Interested but no plans Not interested 35% 34% Don't know 5% Base: 650 server, storage, or data center decision-makers at North American and European enterprises Source: Forrester's Technology Forrsights For Hardware, Q3 2011 12
Cloud-to-cloud disaster recovery Production cloud provider Public cloud provider A B C D E F Replication A B C D E F Failover is automated and can be between different cloud providers or failover within the same provider. 13
Like on-premise DR, recovery sites vary in temperature $$$ Hot cloud site: Recovery cloud is running replica VMs to production site using real-time replication. Recovery time objective (RTO) : 0-2 hours Recovery point objective (RPO): 0-24 hours Warm cloud site: Recovery cloud contains offline copies of virtual machines that can be spun up during disasters or tests. RTO: 2-6 hours RPO: 0-24 hours $ Cold cloud site: Recovery cloud contains backups of production systems that must be first rehydrated and turned into VMs before recovery can occur. RTO: 4-24 hours 14 RPO: 24-48 hours
Agenda The state of disaster recovery today Enter a new model: cloud-based disaster recovery Barriers and benefits to cloud-based disaster recovery Recommendations 15
Benefits of cloud-based DR Better functionality for less cost Easier, more frequent, and less expensive testing More flexible short-term contracts with fast time to market Pay-as-yougo pricing Most of the time, you essentially only pay for storage resources, turning on VMs only in the event of a disaster invocation or a test Little to no upfront investment is required Testing can be automated and non-disruptive. DRaaS contracts usually include testing services and failover assistance Gives you the ability to adapt to changing IT environment and business needs. Deployments are measured in weeks, not months to years On-demand pay-per-use resource that removes the initial investment costs and inflexible contractual agreements, while still delivering recovery capabilities 16
Barriers to adopting DR in the cloud Status quo Firms that provision DR in-house today have massive sunk costs in alternate facilities and infrastructure Finding additional budget to make a change can be a challenge Security and compliance concerns Security concerns regarding compliance and data protection is still the top barrier to any cloud service Companies must ask about practices in data protection, vulnerability management, identity management, etc. Legacy applications Most companies still have legacy applications that can t be run in the cloud, so won t fit the DR in the cloud model More companies are taking multi-tiered approach to provisioning recovery Oversubscription of recovery resources Like traditional providers, DR in the cloud providers also typically oversubscribe resources Companies want visibility into oversubscription and guarantees for certain base requirements 17
Agenda The state of disaster recovery today Enter a new model: cloud-based disaster recovery Barriers and benefits to cloud-based disaster recovery Recommendations 18
Conclusions and recommendations Consider where cloudbased DR might fit into your recovery portfolio It doesn t have to be the strategy for all systems, fewer and fewer companies are using a monolithic approach for recovery today Vet providers in the market today As about SLAs, guaranteed capacity, security procedures, physical location of the data centers, etc. As always, resilient technologies business resiliency You must test recovery, run exercises, train staff, and work with vendors to make sure you can truly meet business requirements for recovery 19
Windstream Hosted Solutions Disaster Recovery as a Service 20
Windstream Overview Advanced Communications & Technology Solutions Fortune 400 Company with $6 Billion in Annual Revenue More than 450,000 Business Customers Nationwide 4 out of 5 FORTUNE 500 Customers Over 150 Offices Across the U.S. Approximately 14,500 Employees More than 115,000 Fiber Miles 6 Network Operations Centers Enterprise-Class Data Centers 21
WHS Division Overview Synopsis Leading Provider of Infrastructure as a Service (IaaS) Delivery Hi- T o u c h TM M a n a g e d S e r v i c e s Cloud Dedicated Colocation Experience 10+ years 1000+ customers Facilities (23) Data Centers SSAE16 SOC1 2N Power Infrastructure 100% Uptime SLA 22
The Hybrid Data Center 23
Hi-Touch Managed Services 24
What is DRaaS? Disaster Recovery as a Service (DRaaS) Combines the best of replication, cloud and virtualization technologies Delivers a fully-managed recovery to a cloud-based disaster recovery infrastructure Ensures your data and applications are safe and secure, and will be there when you need them most 25
How Does DRaaS Work? 26
The Flexibility of DRaaS Multiple replication methods: Host-based replication for heterogeneous physical and virtual environments For EMC-powered data centers, with managed EMC RecoverPoint Appliances and Replication Manager support For NetApp-powered data centers, with managed SnapMirror / SnapVault replication and support for SnapManager Application-layer replication using running VMs (Ex: Exchange Database Availability Groups) 27
The Power of DRaaS Traditional DR Increased hardware and storage Typically limited to 24 hour RPO Long RTO to restore manually from backups Windstream Hosted Solutions DRaaS Leverages cloud economics and economies of scale Variable RPO down to 15 minutes with application consistency Restore entire environment in little more than server boot time Difficult and time consuming to test Off-loads restore burden on provider, customer only has to validate applications Requires significant architectural work Pre-designed and validated for common environments 28
DRaaS Buyer s Guide Cloud Infrastructure Requirements Managed Recovery Requirements Managed Application Availability Solutions Resilient Networking Solutions Production-grade cloud environment Able to run production and DR workloads Able to support application requirements (IO, VLANs) Protects and provides managed recovery of physical and virtual servers Management and monitoring of replication process Self service toosl Support for applicationlayer replication for Oracle/ SQL Server Support for managed n- tier application environments and middleware Managed global load balancing Private network integration Hybrid networking with support for physical servers and network appliances Key Requirements Support for multiple types of replication Ability to run production workloads Support for hybrid/ private networking 29
Thank you Thank you for your interest in Windstream! Need a customized business solution? Contact us at windstreambusiness.com 30