1 / 25

Instrumentation Strategies For The Cloud

01 April 2011. What you need to tell your execs David Halbig. Instrumentation Strategies For The Cloud. Agenda. What is it and what’s the big deal? Where are we today? Why does performance need special attention? What you gonna do about it? Q&A. What Is It and What’s the Big Deal?.

huey
Télécharger la présentation

Instrumentation Strategies For The Cloud

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. 01 April 2011 What you need to tell your execs David Halbig Instrumentation Strategies For The Cloud

  2. Agenda • What is it and what’s the big deal? • Where are we today? • Why does performance need special attention? • What you gonna do about it? • Q&A

  3. What Is It and What’s the Big Deal? Adams, Scott; “Cloud Consultant”, Licensed from www.thedilbertstore.com on 30Mar2011

  4. What Is it? • On-demand self-service • Broad network access • Resource pooling • Rapid elasticity • Measured Service Source: Nat’l Institute of Science and Technology, http://csrc.nist.gov/publications/drafts/800-145/Draft-SP-800-145_cloud-definition.pdf, accessed on 30Mar2011

  5. What’s the Big Deal? Figure 1: Before the Cloud: Outsider’s view of IT Negotiations KAL, “Blame Game”, accessed at: http://www.economist.com/printedition/displayCover.cfm?url=/images/images-magazine/2011/01/15/CN/20110115_CNA400.jpg on 15Jan2011

  6. What’s the Big Deal? Figure 3: “Power of the Cloud”,,BusinessWeek downloaded @ www.businessweek.com on 07Mar2011

  7. What’s the Big Deal? • Potentially much lower TCO • Potentially easy access to the ‘ilities’ (scalability, manageability, share ability) • Absolutely much lower hurdle to prototyping • Potentially much faster time to market

  8. Where Are We Today? Figure 3: The Gartner Hype Cycle The Gartner Group, The Gartner Hype Cycle, accessed from http://en.wikipedia.org/wiki/File:Gartner_Hype_Cycle.svg on 30Mar2011

  9. Why Does Performance Need Special Attention? • On-demand self-service • Interference effects • Geographic dispersion • Broad network access • Resource pooling • Interference effects • Geographic dispersion • Heterogeneous machine architectures • Rapid elasticity • Measured Service

  10. Geographic Dispersion 186,000 Miles Per Second It’s not just a good idea It’s the law

  11. Geographic Dispersion App Server Same Data center – 2 ms latency per SQL call = 10 * 2ms = 20ms DB Server Time

  12. Geographic Dispersion App Server Remote Data center – 40 ms latency per SQL call = 10 * 40ms = 400ms DB Server Time

  13. Interference Effects – I/O GUEST GUEST GUEST GUEST GUEST GUEST GUEST I’m OK I’m OK I’m OK I’m OK I’m OK I’m OK I’m OK Virtualization Layer Server Hardware HELP!!! SAN

  14. Interference Effects - CPU Source: VMware ESX 3: Ready Time Observations - Feb 2004

  15. The Cloud Does Not Fix STUPID • BUILDs promoted with test databases still imbedded • Undocumented off-campus transactions/special cases • Un-benchmarked / un-measured BUILDs, particularly in high-volume environments

  16. What Are You Gonna DO About It? DBaseServer Tier AppServer Tier WebServer Tier Desktop Tier Business Transaction Management Platform “Deep Dive” Diagnostic Monitoring WASWeblogic Tomcat JBoss OracleDB2 SQL Server ApacheIHS System, Host & Process Level Monitoring AIX /LPAR Windows /VMware Linux /VMware Solaris /Zones

  17. What You Gonna DO About It?Performance Monitoring Products DBaseServer Tier AppServer Tier WebServer Tier Desktop Tier Business Transaction Management BTM BTM BTM AWR Spotlight Grid Perfmon Platform “Deep Dive” Diagnostic Monitoring CA-Introscope UpTime VIC ESXPlot UpTime VIC ESXPlot System, Host & Process Level Monitoring UpTime perfPMR UpTime DTRACE

  18. System, Host, and Process Level Monitoring • Continuous Monitors • Windows • Unix (and dialects) – SAR & I/O Stat • Specialized Use Monitors • Network • Windows – WireShark • Solaris – Snoop • Linux – WireShark • AIX – Iptrace • OS • Windows – Perfmon/SysInternals • Solaris – DTRACE • AIX – PerfPMR • VMware – ESXtop/ESXplot

  19. Continuous Monitoring Wikipedia: PerpetuumMobile of Villard de Honnecourt (about 1230)

  20. Intermittent Monitoring In Case of Fire Break Glass

  21. Middleware Monitoring • Continuous Monitors • Database • SQL Server - Perfmon (maybe) • Oracle – Automated Workload Repository (AWR) • JVMs (Tomcat/Jboss) • CA-Wily • Compuware – Agent Vantange • DynaTrace

  22. There Are No Standards for BTM Capabilities Figure 4: BTM-capable Ham Sandwich

  23. Agent Agent Agent Agent DTE DTE DTE DTE BTM High-Level Architecture(Typical multi-tier distributed environment) Management Server & Web GUI Transaction topology map Transaction profiles collected User request DB Server Observed Tier App Server (.Net/J2EE) MSG BUS App Server Web Server DTE = Dynamic Tier Extension

  24. Dave’s Sure-Fire List of BTM Capabilities • Horizontal view of aggregate and single transactions across all tiers of interest • Resource consumption information from each monitored tier • Auto-discovery of transaction path • Capture path contribution of non-monitored tiers • Continuous operation at volume • Low transaction path overhead

  25. Progress!! Figure 1: With the Cloud: Outsider’s Reality of IT Negotiations KAL, “Blame Game”, accessed at: http://www.economist.com/printedition/displayCover.cfm?url=/images/images-magazine/2011/01/15/CN/20110115_CNA400.jpg on 15Jan2011

More Related