Thinking Clearly about the Cloud: Lessons from Real-World Case Studies Steve Else, Ph. D. EA Principals, Inc. stevenelse@eaprincipals. com Federal SOA Conference 12 April, 2011
Use Case/Case Study #1: Migrating a Microsoft Application to the Cloud • Freedom of Information Act Service for the Department of Health and Human Services • Solution Architecture considerations – Access Database • Advantage – Cloud easily supports the provision of a service that can be shared across large agencies/enterprises • Other similar case studies: Share. Point use 2
Use Case/Case Study #2: Writing an Application for the Cloud • Audit Resolution Tracking and Management System for the Department of Health and Human Services • Solution Architecture considerations • Advantages (Enterprise Service) • Other similar case studies 3
Use Case #3: Amazon Web Services -Re-Engineering an Existing App for the Cloud • To build a highly scalable application, some components may need to be re-engineered to run optimally in a cloud environment • Some existing enterprise applications might mandate refactoring so that they can run in an elastic fashion • Majority of reliability issues in the Cloud linkable to problems with the architecting of the app for the Cloud 4
Use Case/Case Study #4: Leveragingan Existing Cloud App • CRM on Salesforce. com by large Federal contractor • Solution Architecture considerations are more focused on analyses of alternatives • Advantages: Quick and inexpensive, an appealing option, if available • Other similar case studies: – Other Salesforce. com offerings – Google Apps – Regonline Instant Access & Common Capability / Data Worldwide 5
Use Case/Case Study #5: Performance and Complexity Limitations of the Cloud • Oil and Gas Exploration in Canada • Solution Architecture Considerations – Complex and extensive infrastructure/apps needed for a competitive advantage • Other similar case study – Epidemiological modeling • But complexity itself doesn’t have to be a show stopper – the Cloud will evolve rapidly in terms of performance…but… 6
Use Case/Case Study #6: Private Cloud • IBM’s focus (among other large vendors) • Business model • • Still sell hardware and software and address maturity issues of Cloud Computing in terms of performance, security, support and overall governance Offer Cloud services • End user value proposition: higher performance, confidence, control and security • Such issues will likely be addressed in the future • Private Clouds provide for the chance to “get use to the idea” without a lot of the uncertainties and risks, esp. associated with complex systems 7
Use Case/Case Study (Private Cloud): #7 Air Traffic Organization • Critical, Complex Infrastructure • Complex System Simulation / Devel. • Communication Among Stakeholders CIO: Private, gov’t-only cloud might support data center consolidation, asset mgt. , virtualization and cybersecurity 8
Use Case/Case Study #8: Cloud Solution not justified for cost and service considerations • Small Association’s CRM systems supporting about 1500 members • Assessment – Cost of migration – Loss of responsiveness – Still need one IT person anyway • Decision to not move the system to the cloud, but instead to upgrade to a single system for higher performance and easier management 9
Use Case/Case Study #9: Large Pharmaceuticals Leverage Cloud for Peak Periods in Processing/Data-Intensive Research • Eli Lilly used Amazon's Elastic Compute Cloud (EC 2) to process bioinformatics sequence data in 20 seconds at a cost of $6. 40 – – – • • • In-house setup for an equivalent 64 machine computing cluster capacity would have required a costly 12 week process Avoided weeks of potential queue delays to critical research Such demand spikes can be absorbed by cloud capability, allowing more level loading and more predictable utilization of internal resources Beyond Eli Lilly, Pfizer, Johnson & Johnson, & Genentech are also among cloud adopters, at least at pilot level Dramatic reductions in cycle times also improve project workflows and outcomes by allowing additional experimental iterations General Challenges: Security for private data, lack of communication /processing standards among providers, “credit card access” encourages rogue use by eager employees through various cloud channels (potentially defeating security policies) 10
Use Case/Case Study #10: Cloud Enables New Process Paradigms Leading t Dramatic Cost and Time Reductions • New York Times used Amazon’s Elastic Compute Cloud (EC 2) to process 4 Terabytes of raw image data (from scans of 11 million archived newspaper articles) and produce and store 11 million finished PDFs into an online newspaper archive storage/access area – – – • Process so fast, automated, and repeatable that an application design error that affected PDF quality was fixed and the entire collection of PDFs were produced again with the same rapid process – • Computation Cost: $240, Execution Time : 24 hours Replaced a much more labor intensive slower PDF assembly process The “Hadoop” Java-based software framework was used to manage the distributed processing for this project (a popular open source, and free, solution for managing data-intensive distributed cloud apps – used by many cloud providers) Again, the luxury of iteration is afforded in a way that would have not been feasible with the conventional time and cost of in-house methods This success on a large scale effort inspired additional successful cloud projects within the company 11
Recommendations • Assess Cloud Readiness for an organization first, if there is a desire or mandate to deploy something to the Cloud • Accomplish an analysis of alternatives in light of desired target state • Prioritize to identify low hanging fruit first • Gather requirements for target Cloud deployment • Accomplish a proof of concept, leveraging the metrics possible in the Cloud for refinements of the Solution Architecture 12
13
Скачать презентацию Thinking Clearly about the Cloud Lessons from Real-World
c6a6d6225c353b2ed21a78899f8ceed4.ppt