Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

第三回月次セミナー(公開版)

1,981 views

Published on

Published in: Technology
  • Be the first to comment

第三回月次セミナー(公開版)

  1. 1. HandlerSocket
  2. 2. HandlerSocket
  3. 3. MySQL Plugin
  4. 4. Handler
  5. 5. Handler
  6. 6. Handler
  7. 7. HandlerSocket
  8. 8. HandlerSocket plugin Client for Java
  9. 9. HandlerSocket plugin Client for Java http://code.google.com/p/handlersocketforjava/
  10. 10. HandlerSocket plugin Client for Javahttps://github.com/ahiguti/HandlerSocket-Plugin-for-MySQL/blob/master/README
  11. 11. HandlerSocket
  12. 12. HandlerSocket HandlerSocket hs = new HandlerSocket(); try{ (Java) System.out.println("open connection."); hs.open(host, port); System.out.println("open index."); hs.command().openIndex(id, db, table, index, fieldList); System.out.println("operations."); for(int i = 0 ; i <= size ; i++){ if(i != 0 && i % loop == 0){ System.out.println("tcommand size : " + hs.getCommandSize() + "byte"); List<HandlerSocketResult> results = hs.execute(); System.out.println("tresult size : " + hs.getCurrentResponseSize() + "byte"); if(verbose){ System.out.println("result."); for(HandlerSocketResult result : results){ System.out.println("t" + result.toString()); } } if(i == size) break; } String[] keys = new String[]{String.valueOf(i)}; hs.command().find(id, keys); } }finally{ System.out.println("close connection."); hs.close(); }
  13. 13. HandlerSocket
  14. 14. Using MySQL as a NoSQL - A story for exceeding 750,000 qps on a commodity serverhttp://yoshinorimatsunobu.blogspot.com/2010/10/using-mysql-as-nosql-story-for.html
  15. 15. Test Case1:TableA JDBC(SQL) HandlerSocket 400 300average time(μs) 200 100 0 0 2500 5000 7500 10000 query per second
  16. 16. Test Case2:TableB JDBC(SQL) HandlerSocket 70 52.5average time(μs) 35 17.5 0 0 10000 20000 30000 40000 query per second
  17. 17. Test Case3:TableB JDBC(SQL) HandlerSocket 1400 1050average time(μs) 700 350 0 0 425 850 1275 1700 query per second
  18. 18. monitoring result(vmstat-disk io)
  19. 19. monitoring result(vmstat-context switch)
  20. 20. monitoring result (vmstat-CPU)

×