SlideShare a Scribd company logo

Performance: See the Whole Picture

•Download as PPTX, PDF•
•
Alexander Podelko
Alexander Podelko

This document summarizes a presentation on web performance given by Alexander Podelko at WebPerfDays New York 2013. The presentation covered performance basics, the importance of considering both front-end and back-end performance, and different approaches to performance risk mitigation. However, the presenter argued that load testing is still needed to complement these approaches, as it is the only way to verify that a system can handle expected load levels and identify potential multi-user issues. Load testing was discussed in more detail, with examples of how it can be used for performance optimization and debugging. The presentation concluded by emphasizing the importance of taking a holistic, end-to-end view of performance.

Related slideshows

Tools of the Trade: Load Testing - Ignite session at WebPerfDays NY 14
Tools of the Trade: Load Testing - Ignite session at WebPerfDays NY 14Tools of the Trade: Load Testing - Ignite session at WebPerfDays NY 14
Tools of the Trade: Load Testing - Ignite session at WebPerfDays NY 14
 
Load Testing: See a Bigger Picture
Load Testing: See a Bigger PictureLoad Testing: See a Bigger Picture
Load Testing: See a Bigger Picture
 
Performance Testing from Scratch + JMeter intro
Performance Testing from Scratch + JMeter introPerformance Testing from Scratch + JMeter intro
Performance Testing from Scratch + JMeter intro
 
1 of 20
WebPerfDays New York 2013 Performance: See the Whole Picture Alexander Podelko alex.podelko@oracle.com alexanderpodelko.com/blog @apodelko October 17, 2013 1
About Me  I have specialized in performance for the last 16 years  Currently performance testing and optimization of Hyperion products at Oracle  Board director at CMG http://cmg.org, organization of performance and capacity professionals – Conference November 4-8, 2013 in La Jolla, CA Disclaimer: The views expressed here are my personal views only and do not necessarily represent those of my current or previous employers. All brands and trademarks mentioned are the property of their owners. 2
Agenda  Performance basics to make sure that we are on the same page  Are we taking care of all parts of performance puzzle? – Where do we place load testing in web operations? 3
Performance  End-to-end, every factor impacts  Front-end – Network and client-side time  Back-end – Server time  Ignore subtleties for the 5 min talk 4
Front-End Time  Takes 80-90% of response time for modern web sites – So indeed deserves the attention it is getting now  Is constant, doesn’t change with load – Single user may investigate – Again ignore subtleties for the 5 min talk  Doesn’t matter for scalability 5
Back-End Time  Server processing  Change with load – You need apply load to investigate  Closely correlated with scalability 6
Front- vs. Back-End  Back-end response times depend on load Thanks to Andy Hawkes for the graph 7
Performance Risk Mitigation I  Single-user performance engineering – Profiling, WPO, single-user performance  Software Performance Engineering – Modeling, Performance Patterns 8
Performance Risk Mitigation II  Instrumentation / APM / Monitoring – Production system insights  [Auto] Scalable Architecture  Continuous Integration / Deployment – Ability to deploy and remove changes quickly 9
But all of them don’t replace load testing Load testing complements them an several important ways 10
Load Testing: Terminology Any testing applying multi-user synthetic load to the system – – – – – – Load testing Performance testing Stress testing Scalability testing Volume testing Reliability testing – – – – – Concurrency testing Endurance testing Longevity testing Soak testing Stability testing 11
The Stereotype  Load / Performance Testing is: – Last moment before deployment – Protocol Level Record-and-Playback – Large corporations – Expensive tools requiring special skills – Lab environment No ! It is just – Scale-down environment one case from –… numerous options ! 12
Load Testing  Any way to apply multi-user synthetic load: – Recording / playback on the protocol level – Recording / playback on the GUI level – API calls / requests scripts – System and components levels 13
Web Operations and Load Testing  “Our experiences were universal: Our software crashed or couldn’t scale” – Jesse Robbins, Foreword  Found 1 mention: – “In our opinion, the only way to make sure our production systems could handle the load was to test them, live, during the day when traffic was flowing” Brian Moon, p.134 14
What Load Testing Adds  Verification that the system handles the load  Verification of multi-user performance  Performance optimization – Exactly the same load  Debugging/verification of multi-user issues 15
Can You Handle Peak Load?  You can’t know without testing: 16
Single-User Performance Improvement  May lead to multi-user performance degradation 17
Missing Parts  Some important performance engineering areas are not covered much at Velocity – Modeling, load testing, etc.  Come to the Performance and Capacity conference by CMG http://www.cmg.org – November 4-8, 2013 in La Jolla, CA – Perfect complement to Velocity / WebPerfDays 18
Takeaways  See performance end-to-end, the whole picture  Time spent is not the only variable – Load is another, it may be different under load  Still no way to verify that the system will work under the load but load testing – Let’s talk during open space 19
Questions? Alexander Podelko alex.podelko@oracle.com alexanderpodelko.com/blog @apodelko 20

More Related Content

Performance: See the Whole Picture

  • 1. WebPerfDays New York 2013 Performance: See the Whole Picture Alexander Podelko alex.podelko@oracle.com alexanderpodelko.com/blog @apodelko October 17, 2013 1
  • 2. About Me  I have specialized in performance for the last 16 years  Currently performance testing and optimization of Hyperion products at Oracle  Board director at CMG http://cmg.org, organization of performance and capacity professionals – Conference November 4-8, 2013 in La Jolla, CA Disclaimer: The views expressed here are my personal views only and do not necessarily represent those of my current or previous employers. All brands and trademarks mentioned are the property of their owners. 2
  • 3. Agenda  Performance basics to make sure that we are on the same page  Are we taking care of all parts of performance puzzle? – Where do we place load testing in web operations? 3
  • 4. Performance  End-to-end, every factor impacts  Front-end – Network and client-side time  Back-end – Server time  Ignore subtleties for the 5 min talk 4
  • 5. Front-End Time  Takes 80-90% of response time for modern web sites – So indeed deserves the attention it is getting now  Is constant, doesn’t change with load – Single user may investigate – Again ignore subtleties for the 5 min talk  Doesn’t matter for scalability 5
  • 6. Back-End Time  Server processing  Change with load – You need apply load to investigate  Closely correlated with scalability 6
  • 7. Front- vs. Back-End  Back-end response times depend on load Thanks to Andy Hawkes for the graph 7
  • 8. Performance Risk Mitigation I  Single-user performance engineering – Profiling, WPO, single-user performance  Software Performance Engineering – Modeling, Performance Patterns 8
  • 9. Performance Risk Mitigation II  Instrumentation / APM / Monitoring – Production system insights  [Auto] Scalable Architecture  Continuous Integration / Deployment – Ability to deploy and remove changes quickly 9
  • 10. But all of them don’t replace load testing Load testing complements them an several important ways 10
  • 11. Load Testing: Terminology Any testing applying multi-user synthetic load to the system – – – – – – Load testing Performance testing Stress testing Scalability testing Volume testing Reliability testing – – – – – Concurrency testing Endurance testing Longevity testing Soak testing Stability testing 11
  • 12. The Stereotype  Load / Performance Testing is: – Last moment before deployment – Protocol Level Record-and-Playback – Large corporations – Expensive tools requiring special skills – Lab environment No ! It is just – Scale-down environment one case from –… numerous options ! 12
  • 13. Load Testing  Any way to apply multi-user synthetic load: – Recording / playback on the protocol level – Recording / playback on the GUI level – API calls / requests scripts – System and components levels 13
  • 14. Web Operations and Load Testing  “Our experiences were universal: Our software crashed or couldn’t scale” – Jesse Robbins, Foreword  Found 1 mention: – “In our opinion, the only way to make sure our production systems could handle the load was to test them, live, during the day when traffic was flowing” Brian Moon, p.134 14
  • 15. What Load Testing Adds  Verification that the system handles the load  Verification of multi-user performance  Performance optimization – Exactly the same load  Debugging/verification of multi-user issues 15
  • 16. Can You Handle Peak Load?  You can’t know without testing: 16
  • 17. Single-User Performance Improvement  May lead to multi-user performance degradation 17
  • 18. Missing Parts  Some important performance engineering areas are not covered much at Velocity – Modeling, load testing, etc.  Come to the Performance and Capacity conference by CMG http://www.cmg.org – November 4-8, 2013 in La Jolla, CA – Perfect complement to Velocity / WebPerfDays 18
  • 19. Takeaways  See performance end-to-end, the whole picture  Time spent is not the only variable – Load is another, it may be different under load  Still no way to verify that the system will work under the load but load testing – Let’s talk during open space 19