Oracle Quality of Service Management - Meeting SLAs in a Grid Environment
- 1. 1
- 2. <Insert Picture Here> Oracle Quality of Service Management Meeting SLAs in a Grid Environment Mark V. Scardina Aris Prassinos Director, Product Management Chief Engineer Quality of Service Management MorphoTrak, SAFRAN Group
- 3. The following is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle. 3
- 4. Improving the Corporate Bottom-line - Proactive SLA Management • Assure predicable performance for consolidated applications • Dynamically allocate resources to meet SLAs • Reduce IT costs by optimizing resource use Quality of Service Management 4
- 5. Online Systems – Outage Causes - Driving IT towards SLAs 5
- 6. Oracle Quality of Service Management - The Objective In real-time you need to be confident that: • When resources across the datacenter are sufficient they are continuously deployed to ensure performance and availability objectives will be met. • When resources are insufficient to meet demand more business critical objectives will be met at the expense of less critical ones. • When load conditions severely exceed capacity, resources remain available. 6
- 7. The Real-Time Performance Equation Code Development Resource Capacity Planning Use Configuration/Deployment Resource Quality of Service Management Wait Application IT Department Performance 7
- 8. Quality of Service Management - The Oracle Approach … • Instrument the entire Oracle stack • Support end-to-end work classification • Generate real-time correlated performance and resource wait data • Accurately identify bottlenecks and their causes • Produce sensible and incremental recommendations and respective impact projections • Respond with Just-in-Time resource allocation • Protect existing workloads from memory starvation 8
- 9. Dynamic Resource QoS Management - The QoS Management Policy 9
- 10. Quality of Service Management - The Oracle Approach in Action Define and Enable Classify and Measure Evaluate and Policy-Driven Analyze and Report Architecture Recommend Implement and Control 10
- 11. Quality of Service Management - The Oracle Approach in Action Sales Clients Back Office Clients Response Time Objectives Sales Pool Back Office Pool NA APAC EMEA Resource (CPU) Most Critical Least Critical Recommendation: NA Performance Objects being CPU Bottleneck. EMEA Performance Class has met. All Performance Class has CPU Bottleneck. Action: Move Server from Server Pool Backfrom Level 2 to Level 0. Promote EMEA Performance Class Office to Sales. No action required. 11
- 12. Memory Guard via Cluster Health Monitor - Protecting Your Server Resources • Receives a continuous stream of OS Memory metrics • Analyses once a minute for over-committed memory conditions • Issues alert should any server be at risk • Protects existing work and applications by automatically closing the server to new connections • Automatically re-opens server to connections once the memory pressure has subsided 12
- 13. Cluster Health Monitor with Memory Guard - Protecting Your Server Resources Sales Clients Back Office Clients Physical Sales Pool Back Office Pool ! Most Critical Least Critical Memory Alert: Memory Pressure detected – Server at Risk. Action: No new sessions accepted. 13
- 14. Integrated, Task-Based Management - New EM Quality of Service Management Tool 14
- 15. <Insert Picture Here> Takenori Otsubu Director of Sales & Marketing IT Infrastructure Solutions Bureau NS Solutions Corporation “NS Solutions highly welcomes the release of Oracle Database 11.2.0.2. Oracle's new QoS Management technology enables more efficient resource utilization and provides greater service stability to our customers. We will utilize this new functionality to enhance our NSGRANDIR+ IT infrastructure optimization methodology.” 15
- 16. MorphoTrak 16
- 17. MorphoTrak: a U.S. Company Gateway to international biometric technologies for U.S. clients Facts . Part of a Global Leader Headquarters: Alexandria, VA U.S. Subsidiary of Morpho, SAFRAN Group Engineering: Anaheim, CA & Tacoma, WA World leader in biometrics & ID management Global market leader in AFIS Operation Centers: Austin, TX & Albany, NY according to Frost & Sullivan More than 450 Employees Over 50% of biometric identity systems worldwide More than 35 Years Experience with U.S. & Global Biometric Solutions http://www.morpho.com http://www.morphotrak.com 17
- 18. MorphoTrak: Our Business Design, Manufacture & Integration of Biometric & Identity Management Solutions Forensic Systems . Federal . Business Identity . Solutions Security .Solutions Integrated Information Federal Forensic Security & Identification Biometric Solutions for: Systems for Agencies Solutions to Governments -FBI Physical Access Control Logical access/sign-on Entry – Exit Systems Time & Attendance -US Visit Building Management Access control terminals Automation -Ports (TWIC), Airports (FAA), MetaMorpho™ AFIS Federal Govt (PIV) Identification Systems MorphoBIS™ AFIS Help Governments Fight -Passenger Security (TSA) Identity Fraud : AFIS Ensures: Driver licenses Accurate Identification of Background checks Individuals Welfare Benefits Readers Suitable for: Automated Border Control ID Cards Faster Resolution of Harsh Industries - Finger-Face-Iris Passports Criminal Cases Corporate Facilities Etc. Luxury Environments 18
- 19. MorphoTrak Customers • Chosen for three largest biometric systems in the world – FBI Next Generation Identification – DHS US-VISIT – India Aadhaar Project (UID) • Over 200 “smaller” ones – Florida FDLE – Netherlands National Police – Swiss National Police – North Carolina driver license 19
- 20. MorphoBIS • Morpho Biometric Identification Solution – Multi-biometric identification system – SOA, Java, Oracle, XML • Biometric templates stored in Oracle as LOBs – OLTP random access – Every biometric search results in thousands of template fetches from the database depending on number of respondents from first stage matching – Unpredictable service times due to varying number of respondents • Multiple “open” workloads of different priorities and SLAs – Border Control, Criminal Investigations, Benefits Verification, Background Checking – Unpredictable request arrival rates and surges – Workload priorities change over time – Need to avoid long queues during demand surges • Border Control during peak crossing hours • Benefits Verification during business hours 20
- 21. Why QoS Management? • Turnkey fixed-cost system with contractually defined SLA – Hosted by customers – Basic onsite administration provided by generalist customer staff – Oracle support provided remotely by MorphoTrak • Performance objective is the response time for fetching individual biometric templates – Can’t throw more CPUs to the problem due to increased Oracle licensing costs – No onsite DBA to constantly monitor and adjust the system • Traditional workload management approaches not adequate (e.g. DBRM directives for CPU allocation, priority queues, varying # of threads) – Complex to determine right settings, imprecise, no clear correlation with SLA – Tend to take more resources than needed from low priority workloads to guarantee SLA of high priority workloads • MorphoTrak beta tested QoS for possible future use 21
- 22. QoS Management Policies 22
- 23. QoS Management - Dashboard 23
- 24. Surge Identified – Recommendation 24
- 25. Recommendation Implemented 25
- 26. SLAs Restored – All Is Well 26
- 27. Metrics – Tracking Performance 27
- 28. Conclusions • Goldilocks “just right” approach to Quality of Service management – Prioritizes workloads to meet all SLAs in a fair manner without overshooting – System resets to default priorities when surges subside • Cost effective – Optimal utilization of system resources without over-configuring the system • Diagnosis and tuning is simple and intuitive – No need for expert personnel on-site – Notifications can be provided to operators when surges occur – Recommendations can be implemented by generalists 28
- 29. Oracle Quality of Service Management - Delivering Datacenter Performance Datacenter Performance Datacenter Performance Today Tomorrow Applications cooperate in sharing Applications fight for resources resources Reactive Resource and Application Proactive Resource and Application Protection Protection Manually allocate resources Allocate resources Just-In-Time to meet SLAs to meet SLAs Grid for Cloud value is not fully realized Grid for Cloud value is fully realized 29
- 30. Quality of Service Management - Live Demonstration Demo Pod W-028 Moscone West 30
- 31. 31