WA1929 Cloud Computing for Architects using OpenStack EVALUATION ONLY

Transcription

1 WA1929 Cloud Computing for Architects using OpenStack Web Age Solutions Inc. USA: Canada: Web:

2 The following terms are trademarks of other companies: Java and all Java-based trademarks and logos are trademarks or registered trademarks of Sun Microsystems, Inc. in the United States, other countries, or both. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. IBM, WebSphere, DB2 and Tivoli are trademarks of the International Business Machines Corporation in the United States, other countries, or both. Other company, product, and service names may be trademarks or service marks of others. For customizations of this book or other sales inquiries, please contact us at: USA: , getinfousa@webagesolutions.com Canada: toll free, getinfo@webagesolutions.com Copyright 2013 Web Age Solutions Inc. This publication is protected by the copyright laws of Canada, United States and any other country where this book is sold. Unauthorized use of this material, including but not limited to, reproduction of the whole or part of the content, re-sale or transmission through fax, photocopy or is prohibited. To obtain authorization for any such activities, please write to: Web Age Solutions Inc. 439 University Ave Suite 820 Toronto Ontario, M5G 1Y8

3 Table of Contents Chapter 1 - The Rise of the Cloud Where did Cloud Computing originate? Cloud Computing Wikipedia Entry Gartner on Cloud The NIST Perspective Five Characteristics Five Characteristics The Cloud Computing Spectrum / Service Models Cloud Deployment Models...18 Understanding Cloud Computing Understanding by Phone Service Analogy Understanding by Electrical Power Grid Service Analogy What is so special about Cloud? Synergy is Powerful Moving to the cloud Capacity Planning Challenge Measuring Capacity Capacity Planning Concepts and Challenges Capacity Planning Utilization Risk Utilization Risk Mitigation Capacity Planning Different Workloads Multi-Tenancy Model Common Characteristics of Multi-tenant Applications Data Management in the Cloud Data Management in the Cloud Data Physics By the Numbers Summary...28 Chapter 2 - Cloud Computing Value Proposition Why does Cloud matter? Cloud Value Proposition Cloud Value Business Case # Cloud Value Business Case # Cloud Value Business Case # Cloud Value Business Case # Cloud Business Cases Cloud economics Cloud economics Cloud economics Do Clouds Compute? Select Expected Benefits Identify applicable cost scenario Calculate initial, simple return...39

4 Calculate returns for on-going usage Summary...39 Chapter 3 - Cloud Computing Myths Myth #1: Cloud == Virtualization Myth #2: Cloud == Grid Myth #3: Cloud == SAAS Myth #4: Cloud == SOA Myth #5: Cloud == Security Risk Summary 1/ Summary 2/ Chapter 4 - Cloud Computing Components The Cloud Computing Stack Cloud Computing Components Tightly Coupled Enterprise Breaking the Silos Understanding SOA Applying SOA to the Cloud Cloud Computing without SOA Cloud Component - Virtualization Hypervisors Hypervisor Types Type 1 hypervisors Type 2 hypervisors Applying Virtualization to the Cloud Cloud Component - SaaS Applying SaaS to the Cloud Web 2.0 Should I upgrade? Web 1.0 vs Web Applying Web 2.0 to the Cloud Summary...54 Chapter 5 - Categorizing Clouds Consider the kind of cloud Cloud Scope Public clouds Cloud Scope Private clouds Cloud Scope Hybrid clouds Discussing Cloud Scope Cloud Type - User Interface Cloud Type Application Service Cloud Type Data Services Discussing Cloud Types Intersection of Scope & Type Cloud Role Discussing Cloud Categories Cloud Integration Cloud Integration Summary...61

5 Chapter 6 - Real World Case Study Analysis Case Study Amazon Web Services (AWS) Amazon EC2 Value Discussing Amazon Case Study TuneCore TuneCore s Value Discussing TuneCore Case Study Salesforce.com Salesforce.com Value Discussing SalesForce Case Study Google Apps Google Apps Value Discussing Google Case Study Pitney Bowes Pitney Bowes Value Discussing Pitney Bowes Case Study X.Commerce/OpenStack OpenStack Value Discussing X.Commerce/OpenStack Summary...70 Chapter 7 - Cloud Risks and Risk Mitigation Failure-As-A-Service in Failure-As-A-Service in Notable Breaches in Notable Breach # Notable Breach # Notable Breach # 4 & # Notable Breach # The Cost of Failed Clouds Risks When Consuming Clouds: Service Reliability Service Quality Problem Resolution Data Back-up Total Cost of Ownership (TCO) Risks When Supporting Clouds: Provisioning The Scale of Scale Financial Management How to Practically Estimate Your Cloud Bill? Managing Service Levels Redundancy / Failover Vendor Lock-In Liability Security Cloud Security Access Control Application Security...85

6 7.26 Application Multi-Layer Security Design Information and Data Security Network Security Operational Security Mitigating Cloud Computing Risks: Identifying Cloud-ready Solutions Governing Cloud Services Business alignment Asset Ownership Contract-driven Services Financial Management and Tracking Governance and Risk Mitigation Some Best Practices Summary 1/ Summary 2/ Chapter 8 - Cloud Standards What Exactly Are We Standardizing? What Exactly Are We Standardizing? Standardizing on a Definition Why Standardize? Simple Concept, Difficult Implementation Perspective #1 Turf Wars Turf Wars Turf Wars Other Groups Defining Standards Recent Progress Recent Progress (cont.) Perspective #2 Let the Market Decide Standardization is Restrictive not Creative Fostering Creativity Survival of the Fittest OpenStack Foundation Model Perspective #3 The Simple View Standards to Date OCCI Details OCCI Client Handshake OCCI Server Response OCCI Server Response (cont.) OCCI Cloud Infrastructure Categories Best Practices for Working with Cloud Standards: Building Cloud Solutions Today Advice #1 Build on Proven Standards Advice #2 Focus on Solid Design Advice #3 Good Standards Take Time Review 1 / Review 1 / Chapter 9 - Cloud Computing Infrastructure...111

7 9.1 Does It Really Matter? Cloud Infrastructure Categories Understanding Virtualization Cloud Management Controllers and Agents More Cloud Management Strategies Abstracted Networking Abstracted Computing Abstracted Storage Pulling it all together Eucalyptus Eucalyptus (cont.) Eucalyptus (cont.) Eucalyptus (cont.) Eucalyptus (cont.) OpenStack What is OpenStack? What is OpenStack? OpenStack main components OpenStack main components OpenStack main components OpenStack Conceptual Architecture Summary Chapter 10 - Looking Under the Cloud Hood Amazon Cloud Overview Working with Amazon Web Services (AWS) Building an Amazon Cloud Service Google Cloud Overview Google Cloud Storage Working with the Google App Engine Building a Google App Microsoft Azure Overview Working with Windows Azure Building an Azure Cloud Service OpenStack Overview Working with OpenStack Building OpenStack Environment Building OpenStack environment by hand Using automated configurators Creating OpenStack VM Instances Managing OpenStack Clouds Summary Chapter 11 - Cloud Services Defining Cloud Services The Typical Cloud Services Application Services...136

8 11.4 Messaging Application Service Application Service Cache Application Service Specialized Application Service Storage Services Object Storage Archive Storage Relational Storage NoSQL Storage Monitoring Services Review Chapter 12 - Cloud Computing Sanity Check The Cloud Shift Adapting to a Broader Market Before You Leap to the Cloud Cloud Sanity Check # Cloud Sanity Check # Cloud Sanity Check # Cloud Sanity Check # Summary Chapter 13 - Adopting Your Very Own Cloud How can my organization explore Cloud? Cloud Adoption Best Practices Cloud Adoption Phase Identify your business drivers Get Educated Articulate a Value Proposition Define one or more scenarios Cloud Adoption Phase Cloud Adoption Phase Produce a Road Map Gain Stakeholder Buy-in Cloud Adoption Phase Cloud Adoption Phase Establish Governance Invest in Infrastructure Cloud Pilot Scoping the Pilot Project Pilot Project Scope (cont d) Enterprise Roll-out Start small and grow incrementally Summary Chapter 14 - Cloud Reference Model Parsing the Cloud Service Model Cloud Reference Model Cloud Infrastructure...156

9 14.4 Cloud Infrastructure Vendor Comparison Cloud Infrastructure - Cloud Storage Cloud Platform Cloud Software SaaS - Cloud Services SaaS - Cloud Applications OpenStack Solution Stack OpenStack main components/services Compute (Nova) Main Compute (Nova) modules/services Image (Glance) Object Store (Swift) Components of Swift Block Storage (Cinder) Identity (Keystone) Network (Quantum) Dashboard (Horizon) Pulling It All Together Summary Chapter 15 - Cloud Layering Cloud Layering Cloud Layering Overview Content Services Logic Services Orchestration in the Cloud Utility - Security Services Security Service Example Utility - Data Services Layering Example 1/ Layering Example 2/ Layering Example 3/ Layering Example 4/ Layering Example 5/ Summary Chapter 16 - SDLC in the Cloud Software Development Lifecycle Phases SDLC Models Waterfall RAD SDLC Practices The Criticisms of RAD Enterprise Technology Delivery Frameworks ETDF Phases Project Initiation Project Classification Requirements Discovery Analysis and Design...188

10 16.12 Development Testing Production Implementation Post-implementation Monitoring of Cloud Solutions Retirement Summary Chapter 17 - Requirements Discovery Discovering Cloud Requirements Discovery Workshops Running a Discovery Workshop Cloud Requirements Scoping Cloud Requirements Documenting Expected, Average and Peak Usage Defining Cloud Service Levels Discovery Best Practices Discovery Best Practices (Cont'd) What is Six Sigma? Discovery Gotchas Summary Chapter 18 - Analysis and Design Analysis and Design in the Cloud Analyzing Cloud Requirements Requirements Management Analysis Workflow Mapping Cloud Requirements to Usage Scenarios "Good/Not so Good" Use Cases for the Cloud Introduction to Cloud Design Designing Cloud Service Solutions Design the Cloud Service Interface Designing for Cloud Non-Functional Requirements Analysis and Design Best Practices A&D Best Practices - Prototyping A&D Best Practices System Partitioning A&D Best Practices -Leveraging Cloud Platform Services A&D Best Practices - Using Asynchronous Communication Patterns A&D Best Practices - Design for Failure A&D Best Practices - Caching A&D Best Practices - Staying Hands-On Analysis and Design Gotchas More Design Gotchas Summary Chapter 19 - Cloud Design Strategies Cloud Design Strategies Designing for Cloud Availability Designing for Cloud Security Designing for Cloud Security - OWASP

11 19.5 Designing for Cloud Security - OWASP 10 (Cont'd) Designing for Cloud Security Multi-Factor Security Designing for Cloud Storage Stepping Across Site Silos Stepping Across Site Silos SAML and OpenID Stepping Across Site Silos OAuth Selecting the Right Storage Cloud Storage Model Designing for Cloud Management Designing for Cloud Maintainability Designing for Cloud Service Reuse Designing for Cloud Agility Designing for Cloud Usability Additional Usability Considerations Summary Chapter 20 - Cloud Development Implementing Cloud Services Common Pitfalls for Cloud Developers Building Composite Solutions Cloud Development Stacks Creating Services for Amazon WS AWS Toolkit for Eclipse AWS Explorer AWS Toolkit for Visual Studio Testing in the Amazon Cloud Deploying Amazon Web Services Consuming Amazon Web Services Creating Services for OpenStack Creating Applications for OpenStack Testing OpenStack Solutions Consuming OpenStack Solutions Creating Services for Google Testing Google Cloud Services Deploying Google Services Consuming Google Services Summary Chapter 21 - Cloud Governance IT Governance Agile IT in the Cloud SOA Governance Overview SOA Governance in Practice Cloud Governance Top Cloud Computing Consumer Risks Top Cloud Computing Provider Risks Risk Mitigation Defining Cloud Governance...235

12 21.10 Cloud Governance Model Key Artifacts Governance Life Cycle Policies and Procedures Roles and Responsibilities Governance Best Practices Governance Gotchas Summary Chapter 22 - Cloud SLAs The Importance of Cloud SLAs What Belongs in a Cloud SLA? Minimal Cloud SLA Robust Cloud SLA More SLA Items Governing Cloud Service Quality Supporting Clouds Summary...241

13 Chapter 1 - The Rise of the Cloud Objectives Key objectives of this chapter: Describe the major aspects of Cloud Computing Identify the different trends that converge into Cloud Computing Explain the five attributes of Cloud Computing Describe the different elements of the enterprise that can be moved to the Cloud 1.1 Where did Cloud Computing originate? Cloud computing represents a convergence of threads Virtualization, 1960s, 1990s, 2000s abstracting computer resources to support efficiency and availability Grid computing, early 1990s, late 1990s harvesting of computer resources as a collective Software as a Service (SaaS), late 1990s hosting of software in a centralized fashion with access and licensing provided on-demand Web Services (WS), late 1990s standards-based messaging integration technology Service Oriented Architecture (SOA), early 2000s connecting service providers and consumers in a distributed fashion across ownership domains Web 2.0 / Web Oriented Architecture (WOA), early 2000s collaboration, rich multimedia, data mash-ups 1.2 Cloud Computing Everyone has their own definition, perspective, or angle regarding what 'Cloud' is and what impact it will have upon the industry Cloud computing is

14 Chapter 1 - The Rise of the Cloud One scoop of SOA A dash of SAAS Stir in Virtualization A pinch of Grid Computing Layer with Web 2.0 / WOA as desired 1.3 Wikipedia Entry Cloud computing refers to the delivery of computational resources from a location other than your current one. In its most used context it is Internet-based ("cloud") development and use of computer technology ("computing"). The cloud is a metaphor for the Internet, based on how it is depicted in computer network diagrams, and is an abstraction for the complex infrastructure it conceals. It is a style of computing in which IT-related capabilities are provided as a service, allowing users to access technology-enabled services from the Internet ("in the cloud") without knowledge of, expertise with, or control 14

15 Chapter 1 - The Rise of the Cloud over the technology infrastructure that supports them. 1.4 Gartner on Cloud Gartner defines Cloud computing as A style of computing in which scalable and elastic IT-enabled capabilities are delivered as a service to external customers using Internet technologies The What, Why, and When of Cloud Computing Gartner, The NIST Perspective The National Institute of Standards and Technology (NIST) s definition of cloud computing "Cloud computing is a model for enabling ubiquitous, convenient, ondemand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction." Furthermore, the cloud model promotes availability and is composed of five essential characteristics, three service models, and four deployment models." 4 Deployment Models 3 Service Models 5 Characteristics Notes The NIST Perspective NIST is a U.S. agency and laboratory that focuses upon innovation and standardization in science and technology. More information is available at NIST.org 15

16 Chapter 1 - The Rise of the Cloud "The NIST Definition of Cloud Computing" (NIST Special Publication ) Five Characteristics 4 Deployment Models 3 Service Models 5 Characteristics On-demand self-service - A consumer can unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with each service s provider. Broad network access - Capabilities are available over the network and accessed through standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, laptops, and PDAs). Resource pooling - The provider s computing resources are pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand. There is a sense of location independence in that the customer generally has no control or knowledge over the exact location of the provided resources but may be able to specify location at a higher level of abstraction (e.g., country, state, or datacenter). Examples of resources include storage, processing, memory, network bandwidth, and virtual machines. 16

17 Chapter 1 - The Rise of the Cloud 1.7 Five Characteristics 4 Deployment Models 3 Service Models 5 Characteristics Rapid elasticity Capabilities can be rapidly and elastically provisioned, in some cases automatically, to quickly scale out and rapidly released to quickly scale in. To the consumer, the capabilities available for provisioning often appear to be unlimited and can be purchased in any quantity at any time. Measured service Cloud systems automatically control and optimize resource use by leveraging a metering capability at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported providing transparency for both the provider and consumer of the utilized service. 1.8 The Cloud Computing Spectrum / Service Models 4 Deployment Models 3 Service Models 5 Characteristics 17

18 Chapter 1 - The Rise of the Cloud Notes The Cloud Computing Spectrum / Service Models The Cloud Computing Spectrum: Infrastructure-As-A-Service (IAAS) The vendor provides infrastructure for you to deploy your own software solution, OS images, database, etc. Amazon EC2 and GoGrid are hosted in the provider s cloud, RightScale is a management platform to use with an IAAS for scaling and resource management. Platform-As-A-Service (PAAS) The vendor provides a framework that your solution must fit within. You code against their APIs, management resources, and other elements of the framework to ensure that your software stack works within their environment. Software-As-A-Service (SAAS) The vendor provides a complete solution that you simply purchase access to in the form of some sort of licensing subscription model. 1.9 Cloud Deployment Models 4 Deployment Models 3 Service Models 5 Characteristics Private cloud - The cloud infrastructure is operated solely for an organization. It may be managed by the organization or a third party and 18

19 Chapter 1 - The Rise of the Cloud may exist on premise or off premise. Community cloud - The cloud infrastructure is shared by several organizations and supports a specific community that has shared concerns (e.g., mission, security requirements, policy, and compliance considerations). It may be managed by the organizations or a third party and may exist on premise or off premise. Public cloud - The cloud infrastructure is made available to the general public or a large industry group and is owned by an organization selling cloud services. Hybrid cloud - A composition of two or more clouds (private, community, or public) that remain unique entities but are bound together by standardized or proprietary technology that enables data and application portability (e.g., cloud bursting for load-balancing between clouds). Understanding Cloud Computing 1.10 Understanding by Phone Service Analogy Landline phone service (non-cloud) Select a single provider Always use that provider for your service (unless you switch) If that provider goes down, you are hosed Cellular phone service (cloud) Select a standard (CDMA, TDMA, GSM) Select a provider and negotiate rates Provider offers seamless support wherever you travel (sometimes renting bandwidth from others, but abstracting this via the cloud) If the provider is unavailable, the contract may allow you to use another service at a higher rate 19

20 Chapter 1 - The Rise of the Cloud 1.11 Understanding by Electrical Power Grid Service Analogy Another analogy is the Electrical Power Grid that generates and delivers electricity from suppliers to consumers, where Power stations are data centers Transmission lines are the network Consumer electrical appliances are client applications Electrical power is available on demand Power distribution is a utility service Clients pay as per as you go model Much more efficient and reliable then managing your own power generator 1.12 What is so special about Cloud? There are some that will argue that Cloud Computing is merely the next fad in a string of hype and buzzwords Does Cloud Computing actually offer anything novel or unique? Not exactly. You would be hard-pressed to identify a 'Cloud Standard' or 'Cloud Platform'. On the other hand, it provides a unifying theme to several evolving threads. Synthesis is valuable Synergy is Powerful Virtualization offers technology cost savings and productivity increases. SOA provides a foundation for the creation and governance of services and business processes. SaaS offers a paradigm for software delivery and shifts cost models and capital expenditures to a lean, on-demand model. Web 2.0 delivers compelling user interface experiences. Cloud brings all of these together and synthesizes a solution to a business 20

21 Chapter 1 - The Rise of the Cloud problem Moving to the cloud A big part of what makes the concept of Cloud Computing so interesting is the ability to move select IT concerns "to the cloud" Establish a contracted agreement Pay for service and actual usage Disregard up-front fixed costs and avoid on-going maintenance costs (just fee-for-service) What can you move to the cloud? Data Software Business logic / processes / rules User interface 1.15 Capacity Planning Is the key part of assessing the practicality of moving your solutions to the cloud In many cases, you need to decide upfront: how many instances you will need (create a pool of stand-by instances or let the cloud scale out elastically). You need to define what resources are cloud, what resources aren't e.g. Database might be outside the cloud Determine critical parameters Disk Capacity 21

22 Chapter 1 - The Rise of the Cloud I/O Database CPU Memory You need to make sure you don't overrun your cloud bill To help you estimate your cloud bill (per month or year), cloud vendors offer resource utilization calculators Capacity Planning Capacity planning is a key aspect of cloud deployment. One of the key value propositions of cloud is that capacity can be shared and optimized across many application deployments. What resources are cloud resources, and which aren't For example, database might be outside the cloud Determine which parameters are likely to be critical Disk capacity May be distinguished by instance or attached storage and remote or SAN storage Database capacity Common to have a traditional database cluster outside the cloud serving the cloud. If so, then apply traditional capacity planning for the database. I/O Capacity (disk read/write operations) CPU capacity how to measure difficult since cpu types vary. Amazon, for instance defines a compute unit as equivalent to 1.0 or 1.2 GHZ 2007 Opteron or 2007 Xeon processor Memory Capacity 1.16 Challenge Measuring Capacity Different nodes have different capacities Capacity is virtualized 22

23 Chapter 1 - The Rise of the Cloud e.g. CPU difficult to compare May require arbitrary measurements Amazon, for instance defines a Compute Unit GHz 2007 Xeon processor equivalent 1.17 Capacity Planning Concepts and Challenges As number of applications, instances or deployment units grows, capacity becomes a linear commodity Smaller deployments need to be planned on a what-if basis Also applies to applications that take a large slice of resources Traditional capacity planing focused on new servers for applications that had funding On-demand provisioning changes the equations Fewer signals of capacity need Shorter time horizon when needs are known with any degree of certainty Need to speed up ability to add capacity Capacity Planning Concepts and Challenges As number of applications/instances/deployment units grows, capacity can be viewed as a linear commodity. For a small number of deployments, capacity needs to be planned on a what-if basis Also applies to deployments that take a large proportion of the cloud's resources. Traditional capacity planning focused on new servers for applications able to achieve capital funding. This model is undermined by allowing application teams to assume that servers can be provisioned on-demand, which of course is the whole idea of the cloud. Fewer capacity signals are available, and the time frame for capacity planning becomes much shorter If applications are thought-up, written, and deployed in a matter of weeks or months, with 23

24 Chapter 1 - The Rise of the Cloud on-demand assumptions, capacity planning has a much shorter time horizon with any degree of certainty. May need to drastically reduce your reaction time in adding cloud resources Capacity Planning Utilization Risk Utilization Risk Risk that utilization of a resource is different from expected Utilization Risk Overuse leads to excess wait times and failure to meet requirements Underuse wasted money, low return on investment In traditional server per application model, the app owner also owns the server, thus bears the risk In the cloud, cloud provider bears the risk Utilization risk Risk that the utilization of a resource is higher than expected Overuse, typically leading to excess wait times and failure to meet requirements or lower than expected Wasted money, low return on investment In traditional Server-per-application, application's sponsor owns the server, thus bears the utilization risk. In cloud, the cloud provider bears the utilization risk Utilization Risk Mitigation App owners can provision and release resources on-demand Could lead to seasonal or other periodic surges Solutions: Controls on how much, how often 24

25 Chapter 1 - The Rise of the Cloud Requires policy setting and negotiation Charge-back model Price is an efficient rationing mechanism Allow use of public cloud Good overflow flexibility, but what about costs and data physics? Utilization Risk - Mitigation Application owners can provision and release resources on-demand (perhaps subject to some limitations). Could lead to seasonal or other-periodic surges in demand One solution charge-back costing model to pass on the costs to users. Price is an efficient rationing mechanism. Application owners might also have the option to use a public cloud when necessary. Good overflow flexibility, but what about costs? 1.20 Capacity Planning Different Workloads Applications have different needs for CPU, memory, I/O How do you plan without knowing the mixture? Offer different types of VM instances to users Users can decide (and pay for) the type of workload they need to provision for Again, cost tends to lead users toward informed choices Capacity Planning Different Workloads Some applications will be more CPU, memory, or I/O-focused than others how to plan for them? Offer different types of VM instances to users. Then users can decide (based on insight or testing) which instance type they need. 25

26 Chapter 1 - The Rise of the Cloud 1.21 Multi-Tenancy Model Cloud providers pool their IT capabilities to serve multiple cloud service consumers by employing multi-tenancy models Multi-tenancy often relies on the use of virtualization technologies Cloud IT resources are partitioned so that computational resources, data, network, etc. can be shared between multiple organizations Each organization sees only its own tenanted (rented) resources 1.22 Common Characteristics of Multi-tenant Applications Tenant isolation Run-time behavior of one tenant does not affect others Data isolation Tenants receive individual databases and / or schemas Network I/O operations are totally isolated Scalability It is a tenant-scoped operation; it does not affect application availability and / or performance of other tenants Security Enforced on the per tenant basis Vendors may employ common cloud public-key infrastructure (PKI) shared among tenants Usage Metering and Billing Resources (CPU, Network, Data, etc.) consumed by tenants are metered and tenants are billed at current rates only for what hey consumed Data back-up and recovery Can be executed on demand by a tenant without affecting other tenants 26

27 Chapter 1 - The Rise of the Cloud 1.23 Data Management in the Cloud All the usual good practices apply Canonical Data Model Master Data Management Horizontal Scaling affects data Database Sharding partition data across multiple nodes to parallelize bandwidth and access. NoSQL Databases are a good fit Eventual Consistency - non-transactional programming style Stateless servers help with scalability BUT! - data physics applies 1.24 Data Management in the Cloud Multi-Tenancy cloud model affects data Three approaches Separate database Separate schema Shared Tables with discriminator column - tenant id 1.25 Data Physics Data Physics considers the relationship of data and the processing elements that use the data There is a cost to moving data So data should be located near the processing If data is stored in the cloud rather than on local disk, it will take time and bandwidth to move it. In a networked environment, it's often faster to calculate a value than 27

28 Chapter 1 - The Rise of the Cloud retrieve it Architects need to be able to specify the locality of virtual components and storage Maintain well-defined relationship between processing units and the data they process By the Numbers Cloud Computing certainly seems to be gaining some traction Over half of respondents to a survey by IBM identified that developing new applications for the cloud would be the top activity for the next 24 months. (2011 Tech Trends Report) 37% of businesses globally are deploying cloud to either remotely host applications or host data, or both. (AMD 2011 Global Cloud Computing Adoption, Attitudes, and Approaches Study) 39% of SMBs with employees expect to be paying for cloud services within three years [an increase of 34%] (Microsoft Global SMB Cloud Adoption Study, March 2011) 1.27 Summary Cloud Computing represents a convergence of several different trends, including Virtualization, Grid Computing SOA, Web Services, Web 2.0 / WOA Software As A Service (SAAS) Cloud brings all of these elements together and synthesizes a solution to a business problem Many aspects of the enterprise can be moved to the cloud Data Software 28

29 Chapter 1 - The Rise of the Cloud Business logic / processes / rules User interface 29

30

31 Chapter 2 - Cloud Computing Value Proposition Objectives Key objectives of this chapter Introduce the value and potential associated with Cloud Computing Explore several Cloud business cases Examine a framework for quantifying the value of Cloud Computing solutions 2.1 Why does Cloud matter? Why should enterprises examine Cloud Computing? Three reasons (well, maybe just two): Be more efficient Gain a competitive edge It s too depressing to spend all day watching the market waffle and talking to potential customers that you know don t have any money to spend Efficiency put idle resources to work Frugality 2.2 Cloud Value Proposition minimize capital expenditures, shifting fixed costs to variable costs Reliability offer strong redundancy solutions and seamlessly roll over to alternative providers as needed Scalability more easily adapt to changing user needs and usage scenarios

32 Chapter 2 - Cloud Computing Value Proposition Centralization move capabilities and their support to a centralized provider that can be commonly shared by many Maintainability off-load updates, upgrades and burden of hardware obsolescence 2.3 Cloud Value Business Case #1 Cloud desktop Virtualize the desktop environment, serving up operating systems (along with file and network access) to users via a wide range of thin client interfaces Examples Hosted OS (e.g. OS is accessible on-demand through various devices, served from a centralized location) Operating System User Profile Preferences File Access Network Access Notes Cloud Value Business Case #1 A classic example of this type of solution would be CITRIX solutions like XenDesktop where you host an OS along with configured software through a thin client interface. UI commands are captured at the client-side and then a message is sent over the wire to the server-side where the instruction is interpreted and the command is carried out. The screen update is then sent back to the client-side to display the result of that action. VMWare offers a product called View with similar capabilities. 32

33 Chapter 2 - Cloud Computing Value Proposition 2.4 Cloud Value Business Case #2 Cloud software Deliver software on-demand from a centralized location, rather than supporting a more distributed and diverse environment Examples Google Apps (e.g. replace MS Office, Outlook/Exchange, or similar) Salesforce.com (e.g. move sales team to Web-based CRM) Web-based Software anytime access (24-7) global availability automatic patches/updates per-user incremental cost Notes Cloud Value Business Case #2 Another great example of this is GoToMeeting in which you can access a hosted conference and connect in from anywhere in the world through the use of a subscription plan. 2.5 Cloud Value Business Case #3 Cloud service Provide information or offer a capability to meet requester needs ondemand Examples Data retrieval (e.g. data aggregation, filtering, caching, etc.) Authentication / authorization (e.g. security system abstraction) Resource pooling (e.g. locksmith or tow truck driver) 33

34 Chapter 2 - Cloud Computing Value Proposition Service Common interface Seamless availability Location independence Cloud processing 2.6 Cloud Value Business Case #4 Support peak demand for data and transaction processing on an asneeded basis without paying for those resources during off-hours Examples Batch processing Business process execution Data warehousing / BI 34

35 Chapter 2 - Cloud Computing Value Proposition Cloud Provider (EC2, RightScale, Flexiscale, etc.) Cloud desktop 2.7 Cloud Business Cases Corporate Infrastructure Virtualize the desktop environment, serving up operating systems (along with file and network access) to users via a wide range of thin client interfaces Cloud software Deliver software on-demand from a centralized location, rather than supporting a more distributed and diverse environment Cloud service Provide information or offer a capability to meet requester needs ondemand Cloud processing Support peak demand for data and transaction processing on an asneeded basis without paying for those resources during off-hours 35

36 Chapter 2 - Cloud Computing Value Proposition 2.8 Cloud economics Graphic credit Luca Simonetti, ENGINE Networks Blog, 36

37 Chapter 2 - Cloud Computing Value Proposition 2.9 Cloud economics Graphic credit Luca Simonetti, ENGINE Networks Blog, Cloud economics Dave Powers, Eli Lilly Company ($20B global pharmaceuticals company): We were recently able to launch a 64-machine cluster computer working on bioinformatics sequence information, complete the work, and shut it down in 20 minutes," he says, describing a project the firm executed using Amazon's Elastic Compute Cloud (EC2) service. "It cost $6.40. To do that internally to go from nothing to getting a 64-machine cluster installed and qualified is a 12-week process." 2.11 Do Clouds Compute? Determining a true value for a cloud solution involves several steps Select expected benefits from the cloud solution Identify an applicable cost scenario Calculate the initial, simple return (if any) 37

38 Chapter 2 - Cloud Computing Value Proposition Keeping the benefits constant, calculate the returns for the second and subsequent implementations Select Expected Benefits Possible benefits Cost savings Lower fixed costs Lower variable costs Lower operating costs during peak demand Better service quality Availability improvements Concurrency / performance improvements Better team composition Workforce reduction Staffing re-alignment Service consumer Identify applicable cost scenario Costs associated with developing front-end service consumption interfaces Costs associated with governing the consumption of cloud services Service provider Costs associated with Cloud infrastructure Costs associated with service interface(s) Costs associated underlying business systems Service consumer and provider Add applicable costs from both categories 38

39 Chapter 2 - Cloud Computing Value Proposition Calculate initial, simple return A very simple ROI formula can be applied Benefits you ve assigned to Cloud divided by the cost scenario The initial, simple return is simply calculated based upon whether or not the initial deployment of these services yield an immediate return If moving to a cloud model eliminates certain fixed costs, then an immediate ROI can be realized from your initial Cloud deployment Other returns may only be realized over time Calculate returns for on-going usage Keeping the benefits constant, returns are calculated for the on-going utilization of this new solution Simple calculations may assume a linear profitability driver (via savings and/or revenue) More advanced calculations may follow a non-linear, even exponential profitability curve The key here is to project savings and or new revenue based upon likely usage scenarios You may even find it useful to develop several models, factoring in different usage figures 2.16 Summary Cloud Computing holds considerable value for enterprises Efficiency Frugality Reliability Scalability Business cases abound 39

40 Chapter 2 - Cloud Computing Value Proposition Cloud desktop Cloud software Cloud service Cloud processing Effective tracking and monitoring is essential in order to protect and preserve a return on your cloud investment 40

41 Chapter 3 - Cloud Computing Myths Objectives Key objectives of this chapter Understand common myths surrounding Cloud Computing. Identify the subtle distinctions between Cloud Computing and the common understanding regarding what it is and what it is not. Explain how Cloud Computing differs from these common misconceptions. 3.1 Myth #1: Cloud == Virtualization Myth #1 Cloud Computing is really just virtualization with a web-based interface put in front. Reality Cloud Computing designs often rely heavily upon virtualization technology for server consolidation, better resource utilization, management and driving operational costs down, but encompass a much broader set of supporting technologies. Cloud Computing solutions are rarely implemented without virtualization, but do not explicitly require resource virtualization. Cloud Computing supports a broad set of business cases (cloud dekstop, cloud service, cloud software, cloud storage). Cloud Computing delivers solutions via an elastic, utility computing model (pay per use), shifting fixed costs to variable costs. Cloud Computing supports loosely-coupled integration techniques through the use of standardized Web service interface technologies. 3.2 Myth #2: Cloud == Grid Myth #2 Cloud Computing is just a new term for Grid Computing. Reality They are similar in their efforts to optimize resources and support smooth scaling of capacity. Grid computing helps with resource provisioning on service demand, but it is nearly opposite in its objectives to

42 Chapter 3 - Cloud Computing Myths Cloud computing. Grids are designed for one user to have all resources. Clouds are designed for one user to have a fraction of available resources. Grids are designed to carry the notion of federation to the nth degree. Even ethernet connections are federated. If this design strategy were applied to a Cloud, it would result in a colossal policy challenge to disentangle as multiple users compete for resources. Clouds are designed to virtualize and abstract a portion of resources in order to support elastic demand and scalable solutions. 3.3 Myth #3: Cloud == SAAS Myth #3 Cloud Computing is just the next generation of Application Service Providers (ASPs) in the form of Software As a Service (SAAS) like Salesforce.com, Web-based , Twitter, and others. Reality Cloud Computing supports various design strategies and business cases, with SAAS being a common but not exclusive model. Multiple Cloud service models exist, including Infrastructure-As-A- Service (IAAS), Platform-As-A-Service (PAAS), and Software-As-A- Service (SAAS) For examples of IAAS vs PAAS vs SAAS, see notes section Notes Myth #3: Cloud == SAAS Examples Please! IAAS Amazon Web Services, Google Compute Engine, RackSpace, OpenStack, GoGrid, and RightScale PAAS Force.com, Google App Engine, Microsoft s Azure Services Platform, and Rackspace Cloud SAAS Salesforce.com, Web-based (various), Netsuite, SAP Business ByDesign, Rackspace & Apps 3.4 Myth #4: Cloud == SOA Myth #4 Cloud Computing is basically Service Oriented Architecture (SOA) sitting on top of Virtualization. 42

43 Chapter 3 - Cloud Computing Myths Reality Clouds tend to incorporate service-oriented architectural principles, but incorporate additional considerations beyond the scope of most SOA environments. Cloud services have added requirements to support configurable / declarative data storage, configurable / declarative security, as well as interfacing with cloud controllers and/or agents Clouds tend to leverage SOA design patterns, interface and messaging standards, and service policy enforcement mechanisms SOA carries with it proven governance and service portfolio management strategies and supporting toolsets that Cloud solutions can utilize 3.5 Myth #5: Cloud == Security Risk Myth #5 Cloud Computing represents a huge security risk, exposing critical enterprise resources and valuable intellectual property in the public domain. Reality Not all clouds are created equally. Private clouds can be constructed behind a company firewall, in which an enterprise-level organization serves as the cloud provider for various divisions and business units. Public clouds can be used selectively, moving sensitive data to the cloud and/or moving processing blocks to the cloud out-of-context so as to off-load their processing without revealing their usage. Many organizations house data in 3rd party data centers today and are reasonably confident of the providers security measures, storing data with a Cloud provider could be handled in much the same way. 3.6 Summary 1/2 Myth #1: Cloud == Virtualization Cloud Computing designs often rely heavily upon virtualization technology, but encompass a much broader set of supporting technologies. 43

44 Chapter 3 - Cloud Computing Myths Myth #2: Cloud == Grid They are similar in their efforts to optimize resources and support smooth scaling of capacity, but they are nearly opposite in their objectives. Myth #3: Cloud == SAAS Cloud Computing supports various design strategies and business cases, with SAAS being a common but not exclusive model. 3.7 Summary 2/2 Myth #4: Cloud == SOA Clouds tend to incorporate service-oriented architectural principles, but incorporate additional considerations beyond the scope of most SOA environments. Myth #5: Cloud == Security Risk Not all clouds are created equally. Private clouds can internalize risks behind a corporate firewall. Public clouds can be used selectively and securely. 44