Text Expo 2009 Site Confidence & Seriti Consulting Load Test Case Study

Loading...

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

0 comments

Post a comment

    Post a comment
    Embed Video
    Edit your comment Cancel

    Favorites, Groups & Events

    Text Expo 2009 Site Confidence & Seriti Consulting Load Test Case Study - Presentation Transcript

    1. SERITI CONSULTING Load Testing Case Study Some hints and tips on load-testing your website... 1 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    2. Agenda  Introductions  Before you start  Designing the test  On the Day…  Analysing the results and taking action  The vendor bit…  Questions? 2 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    3. SERITI CONSULTING Before you start Working out what to test and how much to test it 5 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    4. Why do you want to load test?  New website still under development?  Existing website about to head into a business peak period?  New marketing campaign about to launch?  The Met Office is predicting snow... 6 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    5. “Load test” vs. “Performance Test”  Performance vs Load Test?  This affects: How you will do it What you will do during the test Who needs to be involved What information you record! Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 7
    6. Load vs. Performance?  Load Test = \"Black Box\" I only care if the whole system can handle the load If it can't I probably don't have a lot of clues as to why not...  Performance Test = \"White Box\" Looking inside the App and Infrastructure architecture to identify the key elements Designing tests to stress each element Monitoring those key elements during the test Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 8
    7. Test In-house or use 3rd party?  Project triangle decision Time Money Scope  Resources required... Testing software e.g. LoadRunner, Selenium, Site Confidence People to script it Hardware & network to generate the load Accessibility of test environments (or use Prod) Reporting interface to get the right stats & graphs People to interpret the results... 9 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    8. SERITI CONSULTING Designing the test Working out what to test and how much to test it 1 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 0
    9. Decide on your critical user journeys  Home page  Search pages  Details pages  Upload/download  Checkout/e-commerce  SEO Crawl – Google can hammer sites! Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 12
    10. Set your targets  How many transactions/sec?  How many concurrent users?  What is the mix of the user journeys? Don't underestimate the impact of concurrence...  What's your \"safety margin\" We used 40% Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 13
    11. TIP - Web Analytics is your BFF  Leverage your site metrics! Google Analytics, WebTrends or Omniture etc Use the information to identify your peak loads across different user journeys It is very common for people to test their sites at very unrealistic rates... Or focus on the wrong parts… • Don’t forget deep linking! Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 14
    12. TIP – Management Speak..  The Management tends to ask “how many users will the site support”…  How many users do what, how often? Users on playing online poker are different to users reading “The Register”…  I find it best to talk about “page impressions” since that is a metrics they are probably familiar with by now… Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 15
    13. Concurrent users & “think time”  Take the number of page impression per hour  Divide by 3600  That gives you the page impressions (\"transactions\") per second target  Then take the number of \"concurrent users\" and divide by the \"think time\" (dwell time) to give you transactions a second  1200 users with a 60s think time = 20 transactions /sec  600 users with 30s think time = 20 transactions /sec  300 users with 15s think time = 20 transactions /sec  100 users with a 5s think time = 20 transactions /sec Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 16
    14. Design the test… Scripts Profiles Scenarios  Scripts - the actual user journey (“clicks”)  Profiles - how many, how fast, how long etc  Scenarios - combining profiles to match the traffic patterns on your site… The “mix” of user journeys at any one time Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 17
    15. What does a Profile look like? Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 19
    16. 3 Quick Tips…  Baseline your network first  Do a “GET” for a 100K JPEG and see how many you get before you saturate your network  Do “single profile” scenarios first  Work from simple to complex  Test each element before you try combinations  Think about “keyhole” tests to isolate specific application functions  Rather than simulate the “click by click” of the user just “POST” (form submit) directly to the site  This enables you to focus on the app performance rather than the UI elements (you need to do both though!) Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 21
    17. One more tip…  DON’T FORGET YOUR BATCH PROCESSES  How does the system perform during… • Backups • Re-indexes • ETL activities • Any other batch process that stresses the system…  Many a site dies when these processes run  And if you are truly 24x7 online – when is your window? Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 22
    18. On the day – running the test! Who does what, where & when? 23 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    19. Plan out the timetable!  Make the most of your testing window…  Plan carefully but BE REALISTIC!  Allow a lot of time for review between each scenario run...  You WILL change your plans on the day... Be flexible!  Get the right people – Operations, QA, Network, Database etc Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 24
    20. On the day...  DON’T - Suddenly remember that you forgot to tell your hosting provider...  And watch their IDS system think it was a DDOS attack and drop all your packets...  DO – keep an eye on the time (and ensure that time is sync’s across the systems)  You will want to correlate events!  DO – use a meaningful name for each run and add comments  So you know how many users are added, over what ramp up period, and run for how long at a glance  DO – think about cleaning up after you!  If you have been running against Prod you might have created a lot of dummy transactions (and your web logs will be HUGE!) Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 25
    21. SERITI CONSULTING Analysing the results and taking action Data -> Information -> Knowledge -> Action 26 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved
    22. Lots of questions…  Did you hit your targets?  Where there any odd peaks/troughs?  Was “mixed scenario” performance different?  Was performance consistent across test runs?  Did your releases make things better (or worse)?  What was the Max/Mean/Min values  And what was the variance? 2 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 7
    23. Before and After a Network upgrade… Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 28
    24. Lower & longer is better Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 29
    25. Faster response, with more users… Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 30
    26. Taking action…  Use the performance stat (if you gathered any) to isolate the weak links Upgrade/re-factor & re-test  Use the (bad) stats to get more budget to fix the problem…  Use the (good) stats to get more a bonus because you did such a good job… 3 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 1
    27. SERITI CONSULTING The vendor bit… Over to James…  3 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 2
    28. Site Confidence Clients… Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 34
    29. Contact Details – Site Confidence  Talk to Site Confidence on their stand at the TestExpo after the presentation or if you would like a copy of the presentation  Email James – James.Bavington@SiteConfidence.com  Visit their website - www.siteconfidence.com 3 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 5
    30. Contact Details – Seriti Consulting  If you have any questions around  web hosting (IIS, Cloud)  Performance tools (Appmetrics, Avicode)  Operations & ITIL (service-now.com)  Configuration Management (Configuresoft ECM)  Near-shore development, functional test (Endava)  Email me – Stephen.Thair@SeritiConsulting.com  Connect with me on LinkedIn  http://www.linkedin.com/in/stephenthair 3 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 6
    31. Any Questions? 3 Commercial in Confidence - © Copyright 2008 NCC Group plc - all rights reserved 7

    + sthairsthair, 7 months ago

    custom

    534 views, 0 favs, 0 embeds more stats

    Presentation given at the 2009 TestExpo in London w more

    More info about this document

    © All Rights Reserved

    Go to text version

    • Total Views 534
      • 534 on SlideShare
      • 0 from embeds
    • Comments 0
    • Favorites 0
    • Downloads 7
    Most viewed embeds

    more

    All embeds

    less

    Flagged as inappropriate Flag as inappropriate
    Flag as inappropriate

    Select your reason for flagging this presentation as inappropriate. If needed, use the feedback form to let us know more details.

    Cancel
    File a copyright complaint
    Having problems? Go to our helpdesk?

    Categories