Planning A Cloud Implementation

Editor's Notes

• #7: IDC Cloud Services Survey from the last two years group of these motivations into two area - Reduce costs / increase business agility.Interestingly last survey (Oct2009), pay per use passed easy to deploy as the top benefit.most enterprises want to do both And in most enterprises have multiple cloud projects or initiatives In each case (whether small project or major transformation) first decide what is the business goal
• #8: There is a lot knowledge about existing public cloudsBut internal enterprise private clouds may be different.Consider the 5 NIST characteristics - On-demand self-service, Resource pooling, Rapid elasticity, Measured service, Broad network access.For each of these characteristics an enterprise private cloud may be quite different from what comes to mind for public cloudsConsider ‘on-demand self service’ - at AWS it may take minutesto deploy an instance of Oracle database 11g - there is no IT department controlling that deployment But an enterprise has enterprise standards, authority and budget requirements,governance issues, etc. So enterprise private cloud user self service may mead hours instead of minutesBut that’s better than weeks of the old wayEach characteristic may have it’s own ‘enterprise cloud’definitionSecurity another one. – e.g., customer data subject to regulatory issues, private clouds different from public clouds
• #9: And these differences are why organizations choose private cloudsSecurity (and related compliance issues) is at the top of that list of concerns, along with QoS issues / reliability / availability. For these reasons it may be easier to turn existing services into an internal cloud, Creating shared resource pools is faster path to cloud than moving applications and data outsideAnd ultimately most enterprises want to able to deploy hybrid clouds
• #12: What are common use cases we seeAugmentation – handling peak load or anticipated spikes in demand for services in the cloud. Attractive , but may be difficult.dependent on the statefulness of the application and the interdependence Shared Services –the logical extension of SOA movement, shared databases, middleware, security services, Development and Test –perhaps the easiest cloud use-case Shared Infrastructure – Cloud Computing modern evolution of IT consolidation projects, with the Cloud’s capabilities for self service.
• #13: cloud computing for shared services drill downmany different levels from a shared portal and front end infrastructure to shared data services. you’ve heard about Oracle’s Middleware as a Service and Database as a Service capabilities ExaLogic and Exadata. You’ll hear later about Shared Security and Identity Infrastructure. In each area there are specific trade-offs to achieve the potential benefits.
• #14: What kind of opportunities are we talking about for cost savings from moving to consolidated shared services.Credit Suisse a Private Cloud using Oracle technology for Middleware and Database.220 applications on pool of 400 servers,just 15% of the previous of 2800 systems.in line with industry surveys that show 12-15% IT utilization.this architecture precededengineered systems but the end result and the architecture are comparable.
• #15: Let’s drill down a bit further on Database as a Service as a consolidation target. Here we have illustrated 3 different architectures You have seen in the keynote three different architectures mentioned Shared Server, Shared database and Shared Schema. There are potential benefits in both cost saving, the ability to rapidly scale and the self service nature of the database cloud. Each architecture has specific advantages. For an enterprise with common database standards, consolidation to a shared single database instance, may be relatively easy With large database heterogeneity a shared server design provides freedom of choice but there is more administrative overhead and scaling is limited to the server size.A shared operating system design on a database cluster model provides increased scaling and availability, coupled with the ability to run multiple database instances for performance or versioning requirements.Choose the right one or more for your needs
• #16: The key considerations Savings in HW and licenses reduced system management – though you still have db mgmtShared Database is appropriate where there already exists a common hardware/operating system and db version mixed workloads may be better served by deployment of individual VMs in a shared server consolidation.The Shared operating system design is appropirate for Mission critical applications they benefit from the availability architecture of a RAC cluster, while still having the advantage of independent database instances.there is no single best DB consolidation architecture, it depends on the workload and related application architecture, and most enterprises are likely to want to run more than one of these models.
• #17: Having decided which cloud architecture for which applications – what is the migration planGiven that there are usually some migration costs – apps may be selected by assessing suritability, difficulty and ROIa common general strategy is to migrate apps that are peaky, have potentially frequent changes and are likely to see future investment.those applications are probably the early targets for migration as it is easier to demonstrate a positive ROI for the resulting shared infrastructure. Of course all the new development can be moved to the cloudAnd this can happen even it there are migration delays for existing apps
• #20: Achieving cost reduction implies some model for measuring and tracking costs, both direct and indirect costs. power, cooling, and real estate costs, may be allocated in facilities budgets outside the IT organization.And achieving business flexibility implies existing measures of agility, such as the overall time-to-deploy new functions.where adequate models don’t exist or aren’t tracking current costs … - consideration should be given to implementing cost measurement and tracking before begin the cloud implementation. - to achieve the maximum benefits of cloud computing model may mean changing the way the enterprise calculates and allocates costs of services. Perhaps first consider a new more comprehensive cost model and how cloud computing charge backs could be distributed across the IT user base. Oracle’s Enterprise Management cloud infrastructure products are evolving to support not just the management of the cloud, but also the reporting to achieve cloud efficiency and cost allocation.
• #21: direct savings of reduced hardware investment are obvious, but the biggest benefits may be in operations and licensingSimilarly it can be difficult to measure indirect service level benefitsMore importantly - These benefits may be beyond the enterprises current measurement or cost allocation schemas. Moving to higher cost and performance visibility can be a parallel project to the actual physical implementation of the proposed cloud infrastructure. Where these service levels aren’t being currently tracked, then a pilot sample can be done by tracking a current implementation workflow.
• #22: this Oracle customer saved almost 400,000$ over 5 years in hardware and environmental costs by implementing a database as a service over their existing database operations, but as previously noted this is dwarfed by the savings on software and operations.
• #23: Having developed an ROI case what is a desirable pricing model for an enterprise cloud?The simplest model is simply to take cost and amortize it out to the participating business units or applications.On the other hand many enterprises may choose to go farther in implementing a more involved ‘pricing’ or cost allocation model.This could include some form of market rate pricing. Esp if there is a plan to move to some form of hybrid cloud Then the pricing/cost model could then accommodate the rationalization of internal vs external resources.This may go farther to provide some form of value pricing that provides a tax or base to accommodate future investment. This of course would imply a price advantage to an external cloud, unless the enterprise cloud pricing model provided a similar tax on external hosting contracts
• #25: As noted cloud computing may require transforming existing operational procedures, Is the organization readyFor example cloud self-service implies automating operational procedures. Are the existing procedures well documented? Are there existing planning models and tools? Similar consideration should be given to other areas such as IT governance and architecture. Does the current security model support adequate user permission for the proposed cloud model?The enterprise should consider a framework for evaluating the current organizational readiness
• #27: Given a business case and planned architecture – how will the cloud be rolled out – including organizational changesCloud computing implies ‘sharing’. Sharing across applications, department and business units.For example - Will you start with a project and then extend it across the enterprise
• #29: Whichever cloud strategy is selected it is apparent that cloud computing represents some key changes in IT culture. Architecture -Up front vs late bindingIT operations focussesvs self serviceManagement of systems vs cloud control plane (ie., hybrid ready
• #31: The process of creating a roadmap to move to cloud typically involves phases that are parallel to traditional IT consolidationFirst standardizeIdentify all the current implementations of asolution and then Architect a single solution / versionSecond Consoidate – Analyse and review all instances running on the same platform version – plan and then carry out the migrationUse Oracle migration planning services and integration products – GoldenGate, ETL tools etc.May deployment and migration to Exalogic and Exadata or the use of OVM and associated servicesThird AutomateIdentify the deployable units and meta-data, develop the required scripts, and then implement the supporting provisioning tools and workflowe.g., development of custom automated deployments for OVM images and MW platforms, and perhaps the use of SOA and BPM tooling for workflow Finally OptimizeDeploy tools and dashboards and present usage, Capture the metrics and relate to usage, Adapt pricing to demand and evolve new business modelsThis phase may include the deployment of CRM and billing apps to manage charging for usageThis is non-linear / continual
• #32: Here is a customer example implementing a Platform as a ServiceThe company is Norther Trust and their example illustrates a multi-year journey to build this cloud serviceThe result is called“JavaArch” and currently supports 200 applications in 90 WebLogic groups across three separate WebLogic clusters duplicated in 2 live/live data centersThis architecture is designed to provide Improved Time to Market – with a hosting pre-build shared services that every applicationBetter Service Levels – with one model and one set of tools across the enterpriseReduced Costs – the 12 year graph shows % investment in App Dev vs. Infrastructure.The inflection point mid graph is where the JavaArch platform picks up, and infrastructure spend takes a nose diveGetting to this Java based shared services multi-year journey rather than an overnight phenomenon.The four phases depicted are:JavaArch 1.x - The first component JavaArch1.x shared service Web SSO.Web2000 - introduced co-hosting and additional services beyond security.JavaArch8 - Based on WebLogic introduced Automation,templating and scriptingJavaArch11 - the newest environment – adds virtualization for true elastic resource demand withfully automated environment creation. Architecture, repeatable design patterns and platform metrics were essential to this multi-phase roadmap
• #34: Finally let’s summarize the points we have made today :Decide what kind of cloud is under consideration – Infrastructure, Platform… , Private, Hybrid….Identify few select, measurable benefits…. And the potential risks and challengesUse what models may be appropriate to evaluate ROI Evaluate the required areas of organizational readinessDevelop a clear roadmap for organizational deploymentIdentify key transformations
• #35: Speaker Notes:Message #1:Oracle offers you an examination of your most vital data center processes. The Oracle Insightfor Data Center Technology is a program that links leading industry advisors with yourexecutive team to develop a comprehensive process-oriented approach to achieve Data Centerexcellence. Oracle Insight for Data Center Technology provides a consistent, consultativeprocess to assess current state of IT database foundation and the optimum path forward. Thiscomplimentary service to Oracle’s customers brings to bear Oracle’s extensive technology andindustry experience in a highly focused activity that provides a catalyst for data centertransformations.Oracle Insight delivers a strategic assessment of process capability, prioritized roadmap basedon multiple scenarios, and quantifiable economics of strategic decisions. This powerfulcombination delivers a common vision to your executive team with a pragmatic plan ofexecution. With Oracle Insight for Data Center Technology, you have a partner in acceleratingvalue identification and benefits capture from your Data Centers.Message #2:Oracle Insight for Data Center Technology utilizes benchmark practices and an analysisframework based on the industry’s leading standards and practices. We will work with yourexecutive team to customize the program to achieve your objectives. The typical OracleInsight for Data Center Technology has the following steps.1. Data center industry perspective. Oracle facilitates an interactive discussion withyour executives about Oracle’s data center experiences and key industry trends. Inthis phase a scope of analysis and key hypothesis of your data center is developed.2. Discovery. Collaboratively analyze your current business processes and review yoursupporting applications technology through a series of one-on-one and workshopsessions. Opportunities are mapped to technology and process enablers to illustratethe effort required to attain higher-level capabilities. These findings are summarizedin the solution priority map. Oracle identifies the highest-value opportunities anddevelops a prioritized business case with both qualitative and quantitative benefits.3. Solution design. Oracle recommends best-practice business processes andsupporting technology with a road map that includes time-to-benefit analysis and ahigh-level implementation plan for the areas we jointly agree upon.4. Solution presentation. In partnership with you, the Insight team prepares anddelivers a presentation that includes industry trends and key supporting technology,business benefits, and value drivers, as well as a phased implementation plan androad map for the highest- priority opportunities.Message #3:The Oracle Insight program is a global program and leverages the same methodology globally for the benefit of key strategic customers. Message #4:It is also no-cost program for the customer.