Grid Enabled Services Infrastructure (GESI) ‏

1. Grid Enabled Services Infrastructure (GESI)‏

2. Transformation from the Global Information Grid (GIG) to Net-Centric Environment (NCE)‏ DoD NCE Grid Computing The Net-Centric Environment (NCE) “grid” is envisioned as a federation of distributed computing resources available over local and wide area networks that appear to an end user or application as one large virtual computing system. NCE Grid Computing will be built on pervasive services oriented Internet standards which will allow the DoD to share computing and information resources across departmental and organizational boundaries in a secure, highly efficient manner. Background Office of the Secretary of Defense for Networks & Information Integration (OSD NII) presented a roadmap for transformation of the Global Information Grid (GIG) to the Net-Centric Environment (NCE). source: John Daly Department of Defense Computing Infrastructure Brief NCOIC Stakeholder Outreach Working Group, 6 December 2007 http://colab.cim3.net/file/work/SICoP/Virtualization/2007-12-06/JDaly12062007.pdf

3. Technology Enablers of the NCE Clustering • Collection of computer, storage and application resources in a fixed configuration designed to be operated and managed in a unified, high-performance manner. Server Virtualization • Provides the ability to deploy a discrete number of “virtual machines” on a single hardware platform. Network Virtualization • Allows for the establishment of Virtual LANs (VLANs) between servers. Storage Virtualization • Pooling of physical storage from multiple network storage devices into what appears to be a single storage device that is managed from a central console Agile Architectural Framework • Providing a very agile grid framework upon which to build a Service Oriented Architecture (SOA)‏ Technology NCE will be based on business approaches and technology solutions.

4. Booz Allen Hamilton Has Proven Expertise in Document Management within Federal Government spaces An enterprise system: • 7-node Oracle Real Application Clusters (RAC)‏ • 70TB total storage available, 30TB used for database and content storage • High-Availability Clustered JBoss Application Servers Recognized for its success and quality: • Recognized for Technical Achievement, noting superior operational support • OCIO confirmed SOA implementation compliance, supporting scalable, cost-effective architecture goals Deployed Architecture continues to successfully demonstrate: • Scalability – Increased workload from 18k documents per day in 3/06 to over 100k in 3/07. • User Buy-In - A 600% increase in unique users since going operational • Technology enhancement Background In April 2004, our Document Management pilot capability became an accredited and operational capability

5. “ In 2006, humans created 161 exabytes of data, 3 million times the information in all books ever written. - IDC US Green IT Survey ” The demand for computing, network and storage resources shows no signs of decreasing • Approaching power/AC/space limitations • Handling increasing data and growing user base • Inevitable system load spikes • Peak load can be orders of magnitude greater than the average load • Satisfying stability, availability, and agility requirements • Updating antiquated technology solutions • Paying unreasonable maintenance costs, which tie up budgets • IT professionals / expertise unavailable What are your current technical challenges?

6. Enterprise systems are struggling with the accidental complexity Facts • Individual components are not expensive enough…yet: • Compute Power, Storage, Energy • Hardware used to compensate for poor software design • Demand for Information Technology is still growing • Effectiveness ≠ Efficiency source: “Improving Data Center Energy Efficiency: A Holistic Approach”, Andrew Kurz, Burton Group (September 2007)‏ How did we get to this point? Fallacies • The network is reliable • Latency is zero • Bandwidth is infinite • The network is secure • Topology doesn’t change • There is one administrator • Transport cost is zero • The network is homogenous Peter Deutsch’s 8 Fallacies of Distributed Computing “ The significant problems we face cannot be solved by the same level of thinking that created them. - Albert Einstein ”

7. Current Data Center Capacity Planning Current Data Center development paradigms lead to large amounts of under-utilized siloed resources

8. Capacity Planning and GESI Architecture Modular Architecture GESI allows servers and storage to be more efficiently utilized, reducing use of power and cooling resources

9. Business Process Layer Service Interface Layer Applications Layer QOS Monitoring/Security Virtual Services Layer Foundation Layer “ Physical Layer First and foremost, a loosely coupled architecture allows you to replace components, or change components, without having to make reflective changes to other components in the architecture/systems. - David Linthicum, SOA Magazine, October 2007 ” Complex heterogeneous entities can be decomposed Into simpler, individual components and responsibilities How do you decouple the system components in your Enterprise Architecture? Enterprise Services Task Centric Business Services Entity Centric Business Services Application Wrapper Services Virtual Processing Virtualization of Database Virtual File System/Network Storage, Processing, Throughput, Memory SANs, Servers, Power, Cooling

10. An open architecture allows for the ability to add, upgrade and swap components • Commodity based approach • Low Start-Up and Maintenance Costs • Lower Technology Refresh Expenses • Open Source Promotes Faster Deployment • High Availability • Energy Efficient System Scaling – “Green” • Encourages Standardization • Open Standards = Business Process Agility • Reuse of Business Components • Agile Infrastructure • Virtualized Pools of Resources • Transparent Sharing of Distributed Resources by Multiple Clients • Realign current infrastructure • Reduce current footprint by powering down excess equipment • Reclaim previously wasted resources • Distributed Processing / Storage “

11. An Integrated Architecture Solution Services Integration Architecture to Host: Business Processes, Tools, Data Workflow UI Components Web Mash-Up (Ability to access multiple systems through single UI) Multi-Tenants Legacy Capabilities Legacy Databases Database Storage GESI Infrastructure GESI is customizable to users’ initial requirements and scalable to users’ future requirements • High Availability • Extensible • Load Balancing • Ability to Host Multiple Databases and Legacy Databases • Remote Mgmt. Capability • Online Growth • Store File as a Globally Unique Identifier (Store once and only once)‏ • SOA • Open Source • Open Standards • Interoperability • Business Agility • Virtualization – Maximizing Resource Use • More SW across LessHardware “By 2010, Enterprise Web Mash-Ups will be dominant (80%) for composite enterprise applications” -Gartner Symposium, Orlando, October 2007

12. “ Virtualization success depends on tight cooperation between server, storage, network and security teams. • Chris Wolf, Burton Group CIO Magazine, “Virtualization in the Enterprise Survey” ” The combination of service, virtualization and grid techniques can reduce cost while also improving flexibility What new challenges will server consolidation introduce? • Cluster multiple physical servers to gain failover capabilities • Share resources across physical servers source: “Automated failover and recovery of virtualized guests in Advanced Platform”, Rob Kenna, Redhat Magazine

13. GESI is part of a Service Oriented Architecture Foundation • As the Open Group Architecture Framework (TOGAF) states, good enterprise architecture brings important business benefits: • A more efficient IT operation: • Lower software development, support, and maintenance costs • Increased portability of applications • Improved interoperability and easier system and network management • Improved ability to address critical enterprise-wide issues like security • Easier upgrade and exchange of system components • Better return on existing investment, reduced risk for future investment: • Reduced complexity in IT infrastructure • Maximum return on investment in existing IT infrastructure • The flexibility to make, buy, or out-source IT solutions • Reduced risk overall in new investment, and the costs of IT ownership • Faster, simpler, and cheaper procurement: • Buying decisions are simpler, because the information governing procurement is readily available in a coherent plan. • The procurement process is faster - maximizing procurement speed and flexibility without sacrificing architectural coherence. Infrastructure Architecture is part of the SOA Alliance's SOA Foundation

14. GESI provides the essential elements of a Service Oriented Infrastructure Service Oriented Infrastructure Moving from a dedicated infrastructure to a dynamic infrastructure • Addresses all the aspects of the infrastructure • Networks, servers, data centers, and firewalls • Application infrastructure, security, monitoring, middleware, etc. • Key elements • Life cycle support to manage the deployment of SOI components • Virtualization of infrastructure resources to SOI users • Service management to assure the SOI solution provides the required service characteristics • Combines architecture building blocks and role-based portals • High reuse of common services • Reuse of infrastructure and foundational components • Reduction in time needed to deliver capabilities “ SOI delivers bottom-line benefits to the enterprise. It provides the basis for greater IT automation which results in higher IT productivity and lower operational costs. SOA Practitioner's Guide, part 2 ”

15. Benefits of the GESI Solution Improve performance, availability, and cost-effectiveness of compute- and data-intensive applications Run growing volume of complex, resource-intensive, high-performance computing jobs within existing distributed infrastructure Provision additional capacity dynamically as it becomes available, and failover gracefully around unavailable capacity, without interrupting jobs in progress Reduces total cost of ownership of ongoing information technology (IT) operations by making more efficient use of available computing , storage, and network capacity Postpone the need for deployment of additional capacity to support growing transaction loads Provision, scale, and reconfigure virtualized computing resources within a service-oriented environment What are the benefits of a GESI solution? “ Grids enable more flexible, dynamic resource sharing among diverse physical or logical computing environments. - Burton Group Grid Services, “Pooling Distributed Resources for Virtual Supercomputing” ”

16. Isaac Christoffersen Associate christoffersen_isaac@bah.com Christopher Dale Associate dale_christopher@bah.com GESI Physical Layer Doug Johnson Associate johnson_doug@bah.com David Schillero Associate schillero_david@bah.com