1. ■ www.balabit.com
Web Performance
Caching issues
VESZPRÉM MEETUP: 2013. 02
Presented by Boci
GUARDING YOUR BUSINESS
2. Why make performance test? ■ www.balabit.com
■ reduce server load
■ reduce latency
■ reduce network traffic
GUARDING YOUR BUSINESS
3. Magic sentences ■ www.balabit.com
„Cache is King!”
„Improve Performance by Caching”
H !
U G
„Using Web Server Cache to Boost Performance”
O
EN
„Call to improve browser caching”
T
N O
GUARDING YOUR BUSINESS
4. Empty cache? ■ www.balabit.com
■ first time users
■ cleared cache (for ex. antivirus)
■ resource got evicted
■ update - content/page changes
■ (flawed experiment)
GUARDING YOUR BUSINESS
5. Cache types ■ www.balabit.com
GUARDING YOUR BUSINESS
6. Cache – easy steps ■ www.balabit.com
GUARDING YOUR BUSINESS
7. Cache – next chance ■ www.balabit.com
GUARDING YOUR BUSINESS
8. Cache – low level ■ www.balabit.com
GUARDING YOUR BUSINESS
9. Why should we know this? ■ www.balabit.com
■ understand use cases (loading, cache warm-up, request
serving, content update) effects
■ better plan for measurement
■ define performance measurement & load test targets
GUARDING YOUR BUSINESS
10. Sample ■ www.balabit.com
GUARDING YOUR BUSINESS
11. Test with or without cache? ■ www.balabit.com
response time depend from time period and use cases:
□ server load – warm up
□ cache load (application, proxy, browser)
□ simple serve period (just dynamic content)
□ content change → partial cache load
□ private content / dynamiy content
FOCUS YOUR BUSINESS ISSUES!
make more test for check issues
GUARDING YOUR BUSINESS
12. In practice ■ www.balabit.com
■ Web server – web application benchmarking
□ analyze your server side weakness
□ analyze cache effects
□ you shouldn't forget queries, rss, dynamic contents
■ Response time analyzing
□ analyze different page response time
□ analyze different user iteration resp. time
□ analyze different user mix (first time, return, clear
cache)
GUARDING YOUR BUSINESS