1 / 27

Architecture-Based Virtual Management

Architecture-Based Virtual Management. ZHOU Minghui School of Electronics Engineering and Computer Science Peking University, Beijing, 100871, China Oct.29, 2007. Agenda. Why virtual management Architecture-based virtual management management view Management process

damian
Télécharger la présentation

Architecture-Based Virtual Management

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Architecture-Based Virtual Management ZHOU Minghui School of Electronics Engineering and Computer Science Peking University, Beijing, 100871, China Oct.29, 2007

  2. Agenda • Why virtual management • Architecture-based virtual management • management view • Management process • Architecture-based approach • Case study • PKUAS • JASMINe • Summary Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  3. Why virtual management? • Computing environments are becoming increasingly sophisticated • Numerous complex software • Heterogeneous programming models • Configuration facilities are generally proprietary • Resources virtualization is necessary • Shielding users from their complexity and allowing them to focus on what they wish to do, rather than how the technology can get it done • Automation management is important • Ease burden • Achieve efficiency Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  4. Mod_jk CMI CJDBC Why virtual management? –A scenario • A scenario-J2EE cluster management • Complexity: Different layer, Different technology, Different mechanism JGroup Http session replication EJB session replication CMI CJDBC Database replication Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  5. User – access User – access Challenge • So, we have to build an infrastructure to support this • “Virtual, adaptable, open, on-demand” • How? (What are the problems?) • How to virtualize? • How to automate? Virtual Machine Site A Second EchoGrid Workshop Beijing – 29 & 30 October 2007 Site D

  6. Architecture-Based Virtual Management Approach

  7. Management view-three dimensions Resources … … Statistic Diagnose Modify View Contents Runtime Configuration Cluster Web Service Component Application Middleware Service CPU, Memory, Thread Operations Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  8. Management Process • Monitoring • Measuring parameters of the controlled system that are relevant for the function under control • Decision • Deciding the corrective action to be taken. • Analyzing and planning • Execution • Actually performing the actions on the controlled system. • Knowledge base • Represents information on the controlled system.  Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  9. Process-How to make it? • Mechanism • Control process • Virtual operating system • Points: how to monitor? How to analyse? How to execute? How to make them well-structured? • Policy • Predict system’s behavior and action • Capability of learning adjustment policy dynamically by history data • Point: What to be drawn as rules? • UI • Easy way to know the managed resources • Easy way to manage policy • Evaluation • Management overhead • Management efficiency • Point: tradeoff… Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  10. Runtime Software Architecture Reflective Programs Reflective GUI Reflective APIs Correctness and Security of Reflection Formal Description of SA AppSA Specific Meta Entities Application SA PlaSA Specific Meta Entities Platform SA Refine Base Entities SA in Development Deploy Causal Connection Causal Connection Reflective Middleware Based System Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  11. SASA Design of Adaptive SA deliver refactory Deployment of Adaptive SA deploy Re-deploy Execution of Adaptive SA Highlight: Architecture-Based Approach • Philosophy • Designers can predict some runtime changes and plan correct adaptations • Self-Adaptive Software Architecture (SASA) • When to change • Classical SA analysis methods, e.g., ATAM and SAAM • What to change • Dynamic SA, e.g., Taylor@UCI and Garlan@CMU • How to change • Runtime SA by reflective middleware or reflective component Fractal: dynamic SA Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  12. Case Studies 1 Self-Adaptive tuning in PKUAS 2 Autonomous management in JASMINe

  13. Management Module JTA JMS JDBC JAAS IIOP Web Container EJB Container State Model StatisticsModel Event Model ManagementApplication ManagedObject Model Case study -Self-Adaptive tuning in PKUAS Tuning – Make AS running more high-performace by Administrator • Self-Adaptive Tuning – Tuning AS by itself via rules • PKUAS(JO2nAS): Peking University Application Server • Why self-tuning • make AS running high-performance without depending on Administrator • Selects the system requirements and then outputs a diagram of the system configuration and estimates the construction fees Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  14. Case study 1--management process • Management process: 3 step • Get realtime Performance Statistics • Analyse Statistics and give a result • Tuning AS according to the result and runtime information Management Module ManagedObject Model Statistics Model AS Runtime Information Statistics Information Tune Actions AnalyseResult AnalyseRuleLib Statistics Analyser Tuning Controller TuningRule Lib Config File Environment Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  15. Architecture based Deployment Automatic calculation of deployment factors Visualization of architecture models in the development Drag-and-drop deployment of components Visualization of servers and their capabilities Fractal: software deployment and configuration management Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  16. Case study 1-test case • <StatisticAnalyser Name="JDBCConnectionPool"> • <KeyStatisticName>WaitingThreadCount</KeyStatisticName> • <Rule> • <Conditions> • <Condition> • <Trend>1</Trend> • </Condition> • </Conditions> • <ResultCode>JDBCConnectionPool_1</ResultCode> • </Rule> • </StatisticAnalyser> • <TuningRule> • <AnalyseResult>JDBCConnectionPool_1</AnalyseResult> • <TuningAction Object="Analysee"> • <SetAttribute Name="MaxCapacity" Value="30" Type="java.lang.Integer" SetType="Increase"/> • </TuningAction> • </TuningRule> Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  17. Case study 1—test case • Test tool: LoadRunner • Comparison of WaitTime • Comparison of WaitThreadCount Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  18. Case study 1- Self-Healing of Correlated Faults DB failures cause TX failures Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  19. Case study 1--Experiences • Learned • Mechanism is easily built, especially for application server developers • The point is how to build an elegant infrastructure (extensible, modules independent with each other ) • People like architecture • Challenge • How to get Rule for Analysing and Tuning • How to describe the Analyse Rule Lib and Tuning Rule Lib exactly and completely • How to make architecture to explain system exactly • Tradeoff between abstraction and serviceability Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  20. Conf Prop. XML Case study 2- Autonomous management in JASMINe • Build a configuration is quite difficult • Heterogeneous files (properties, XML, …) • Numerous files and parameters • JOnAS = 46 files, Apache configuration file =160 lines • SOA platform => several products, cluster mode => multiple nodes, large scale deployment • Same problem for the monitoring task • Huge number of events to monitor • Distributed architecture • Multiple levels !!! Not manageable by human beings !!! Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  21. Case study 2- Autonomous management in JASMINe • JASMINe: New OW2 project • Bull, SERLI, INRIA, PKU • Tool for administration of SOA platforms • For the designer • Configuration (clusters) • Deployment • For the administrator • monitoring, error detection, performance tracking • Autonomic administration configuration • Configuration and Deployment • Based on logical configuration through ADL and JADE • Monitoring and autonomous management • Probes, filters, rules engine

  22. Case study 2- JASMINe architecuture Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  23. Case study 2- JASMINe: Design GUI • Eclipse EMF/GMF for architecture description

  24. ADL JADE JASMINe ajp-port=8009 driver=mysqlDriver http-port=80 ajp-port=8009 Case study 2- Configuration and Deployment • Logical configuration TO physical configuration Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  25. LeWYS • JOnAS • JVM • System sonde sonde probe sonde sonde sonde sonde probe probe sonde sonde probe Case study 2- Monitoring and Autonomous Management Data JASMINe Data Data Rules Log Mail SMS … Data Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  26. Summary • All roads lead to Rome: virtualization and automation on management • Is computer determinedly on the road of virtualization and automation? • Management process is general, Essence is the same: rule, Architecture makes this efficient • Can machines make decisions and control complex situations without inflexible dependence on predetermined programs? • GRID infrastructure can benefit from current technical solution whose goal (or current goal) isn’t GRID Second EchoGrid Workshop Beijing – 29 & 30 October 2007

  27. Thanks

More Related