Successfully reported this slideshow.

Web performance: beyond load testing

5

Share

Loading in …3
×
1 of 18
1 of 18

Web performance: beyond load testing

5

Share

Download to read offline

Web Performance is a new trend in web application performance analysis and measurement which specializes in overall user experience instead of traditional approach that primarily looks at server performance. Sergey Chernyshev will describe the differences between the two approaches, show why it is important to distinguish between performance and scalability and talk about new tools that go beyond load testing.

Web Performance is a new trend in web application performance analysis and measurement which specializes in overall user experience instead of traditional approach that primarily looks at server performance. Sergey Chernyshev will describe the differences between the two approaches, show why it is important to distinguish between performance and scalability and talk about new tools that go beyond load testing.

More Related Content

Web performance: beyond load testing

  1. 1. Web Performance: beyond load testing January 27, 2012 Sergey Chernyshev Technical Director Web Systems & Applications truTV Web Services, Turner
  2. 2. NY Web Performance 1250+ members 4400+ worldwide since May 2009
  3. 3. Scalability & Performance
  4. 4. What is Web Performance? WPO improves the speed of web experience for each end user It is NOT how many users a web site can serve (Scalability) It is NOT how often a web site is down (Reliability)
  5. 5. Scalability & Performance
  6. 6. Why Web Performance? Increasing revenue vs. cost optimization • Shopzilla +7-12% conversions! & • +15% (+60M) downloads for Firefox. (-1 sec => +2.7% downloads) • Slowness is sticky (Google and Microsoft) • SEO: Google uses site's speed in ranking
  7. 7. Abandonment Rate How Loading Time Affects Your Bottom Line. KISSmetrics http://blog.kissmetrics.com/loading-time/
  8. 8. Why Web Performance? User experience "participants had to concentrate up to 50% more" "greater agitation and stress" Web user connected to EEG "40% will go to a rival web site, 37% will abandon" Web Stress. A wake up call for European business. Foviance on behalf of CA http://www.ca.com/Files/SupportingPieces/final_webstress_survey_report_229296.pdf
  9. 9. Server Time vs. Browser Time 88 requests, 6.344s only 0.968s on backend = just 15%
  10. 10. Server Time vs. Browser Time 151 requests, 6.3s, only 0.1s on backend - less then 2%
  11. 11. Page Statistics 2003 2009 Size: 93.7K 507K Objects: 25.7 64.7 Average Web Page Size Quintuples Since 2003 Andrew King (websiteoptimization.com)
  12. 12. Average Web Page Size Quintuples Since 2003 Andrew King (websiteoptimization.com)
  13. 13. Page Statistics • Load Time: +0.533 s • Time to first byte: +0.117 s • Time to start render: +0.179 s • Page Size: +48 KB • Requests: +4 just 1 year • Connections: +1 (2010) • DNS Lookups: +1 Are pages getting faster? Patrick Meenan (webpagetest.org)
  14. 14. Tools: Analysis • Yahoo's YSlow (best practices) • Google Page Speed (best practices) • WebPageTest (web based waterfalls and more) • dynaTrace AJAX Edition • REDbot (HTTP caching) • DOM Monster (DOM analysis) • *HAR format (HTTP data format)
  15. 15. Tools: Monitoring • ShowSlow (basic metrics and best practices) • WebPageTest + ShowSlow ("geo" testing) • Boomerang, Newrelic (RUM) • Catchpoint • Keynote, Gomez
  16. 16. Tools: Optimization Dev tools WPO automation • SVN-Assets • mod_pagespeed • JSMin, Closure • Yottaa Compiler • Strangeloop • CSSLint • Torbit • SpriteMe, imgmin, • Google Page Speed JPEGmini service
  17. 17. Add Web Performance to your SLAs!
  18. 18. Q&A URL: http://meetup.com/Web-Performance-NY/ Email: sergey.chernyshev@turner.com @SergeyChe @PerfPlanet

×