Your SlideShare is downloading. ×
0
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation

1,513

Published on

Session slide at db tech showcase 2012 …

Session slide at db tech showcase 2012

How Rakuten Reduced Database Management Spending by 90% through Clustrix implementation

- About Rakuten
- Rakuten database environment and operational issues
- What is Clustrix?
- Clustrix verification results and implementation effectiveness
- Summary

0 Comments
4 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,513
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
25
Comments
0
Likes
4
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. 楽天事例紹介: Clustrix導入による DB管理コストの削減 How Rakuten Reduced DatabaseManagement Spending by 90% through Clustrix implementation October 17th, 2012 Ryutaro Yada (矢田 龍太郎) Database Platform Group Global Infrastructure Development Dept. Rakuten, Inc.
  • 2. Introduction Ryutaro Yada  First employed by Rakuten in 2008  Present job  Development of a platform database to support Rakuten  Testing and discussion of new techniques and new architecture in view of having it adopted for use.  Previous functions  Promotion of Oracle business with specified customer  Establish collaborative network with Oracle, develop and verify new solutions, etc..  LinkedIn profile: http://www.linkedin.com/pub/ryutaro-yada/32/368/4b0 1
  • 3. Agenda About Rakuten Rakuten database environment and operational issues What is Clustrix? Clustrix verification results and implementation effectiveness Summary 2
  • 4. Introduction to Rakuten About 3000 employees: (Group approx. 7000) Market / more than 40 services provided including travel More than 120,000 contracted firms; more than 80,000,000 registered products Group distribution total: 3.2 trillion yen (2011) Rakuten market
  • 5. Rakuten Global Expansion Our Goal is to become the No. 1 Internet Service in the WorldLS(UK) ★★ ★ ★ ★ ★★ ★ ★ ★★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★★ ★ ★★ ★ ★ ★ ★★ ★ ★ ★ ★★ ★ ★★ Taiwan ★ ★ ★ ★ *To be open soon ★ ★ ★ ★ ★ Ichiba (EC) ★ ★ Travel ★ Performance marketing ★ ★
  • 6. Rakuten’s Global Position Rakuten is aiming to be the world’s largest internet firm. Firm and highly flexible infrastructure is required to achieve this goal Retail / auction site global ranking 2011 based on unique (no. of) visitors 300000 250000 200000 150000 100000 50000 0 Amazon e-Bay Alibaba Apple Rakuten Walmart Source: comScore Media Metrics
  • 7. Rakuten Database Breakdown according to the number of databases: approx. 80% MySQL (more than 1100) More than 350 MySQL database servers MySQL has the largest share Oracle PostgreSQL Teraddata  No. of databases according Informix to actual environment RDBMS  Same number of databases for each STG and DEV MySQL 6
  • 8. MySQL Database Issue (1) Data Sharding Operations  Required for functionality scaling  Instance/database/table splitting, data redistribution  Correction of application code, control of database access Data Protection, HA Securing  Replication cannot realize zero data loss at failure  Switch back/switch over management takes a lot of effort 7
  • 9. MySQL Database Issue (2) Online Maintainability  Schema modification and index addition, rebuild  Lock, access concentration Number of Units Tends to Increase  Load distribution slave, redundant configuration of slave  Tendency for preparations on an individual service basis (service level differences, maintenance adjustment diversion)  CPU efficiency decreases; increases in data center costs 8
  • 10. Clustrix CharacteristicsWhat is Clustrix? Appliance-style database server Cluster database  NewSQL = LegacySQL + NoSQL  LegacySQL: SQL access, transaction consistency  NoSQL: Scalability, high performance Fault-tolerance function MySQL compatibility  Usually access is through MySQL protocol 9
  • 11. Clustrix Provision Model 2 Models 10
  • 12. Looking at Clustrix SSD Infiniband  Low latency  High performance 11
  • 13. Clustrix Operation Distributed arrangement on the physical layer Redundancy protection, auto rebalance Parallel query execution SQL SQL SQL SQL  Query, not data, is migrated (this concept differs from Oracle RAC) 12
  • 14. TPC-C Benchmark Result 13
  • 15. GUI 14
  • 16. Useful Command Interface 15
  • 17. Clustrix Implementation CasesRakuten is the first case in JapanNumerous foreign cases 16
  • 18. Verification PointsPerformanceScalabilityFault-tolerance verificationOnline schema modification 17
  • 19. OLTP Performance Results (1) Insert 45000 40000 35000 30000 25000(ops/sec) 20000 15000 10000 5000 0 p3 p12 p24 p48 p96 p192 Single Throughput 4014.703409 8350.801098 10022.32827 10448.25479 10520.08066 10213.98278 Clx 3 nodesThroughput 6301.603599 18530.31626 26182.77331 30021.30841 27581.92104 24401.28904 Clx 4 nodes Throughput 6090.513193 20584.42 30544.8252 38545.21774 36837.10176 33221.72529 18
  • 20. OLTP (2) Update 25000 20000 15000(ops/sec) 10000 5000 0 p3 p12 p24 p48 p96 p192 Single Throughput 3854.243586 8018.593249 12186.20793 13385.77834 13395.06587 11538.29668 Clx 3 nodes Throughput 3377.359372 10741.77417 16505.79652 16964.01107 16189.88416 15379.62683 Clx 4 nodes Throughput 3682.99886 12679.26966 19737.63815 22232.7357 21568.39318 21303.72872 19
  • 21. OLTP (3) Read 80000 70000 60000 50000(ops/sec) 40000 30000 20000 10000 0 p3 p12 p24 p48 p96 p192 Single Throughput 6134.386466 26773.71202 44388.78477 56144.27281 57926.62433 49362.51106 Clx 3 nodes Throughput 5050.230295 17380.6806 27803.82494 39693.75633 49822.34026 56847.77879 Clx 4 nodes Throughput 5959.900064 20794.2083 34743.31655 54382.96419 70302.27313 76000.59175 20
  • 22. OLTP (4) Mix 40000 35000 30000 25000(ops/sec) 20000 15000 10000 5000 0 p3 p12 p24 p48 p96 p192 Single Throughput 3976.841546 8587.218158 11632.64122 12946.2536 13122.33748 12769.45794 Clx 3 nodes Throughput 3113.109431 12940.99191 21264.63309 26759.75291 25976.26625 25334.18054 Clx 4 nodes Throughput 5150.537999 15220.8469 25601.67909 34647.41616 34697.737 30804.09949 21
  • 23. Complex and Heavy SQL Comparison Clustrix IA with SSD SPARC with SANJ) Count+GroupBy+OrderBy+Limit 1.9s (3.4s) 2.1s (8.5s) 3.4s (409.32s)K) Count+GroupBy+OrderBy+Limit 0.7s (1.13s) 5.9s (7.49s) 13.0s (39.41s) L) 2000 of IN+GroupBy 3.8s (8.97s) 106.5s (103.77s) 193.0s (321.68s) M) Case+OrderBy 31.0s (45.66s) 47.3s (60.9s) 22 90.5s (112.24s)
  • 24. Example of Performance Improvements Example improvements regarding a particular service  Before: 116.8ms  After: 21.4ms 23
  • 25. Fault-Tolerance Inspection Failure Test Items Downtime1 Front network (port1) No2 Front network (port2) No3 Internal network (primary) < 12s4 Internal network (standby) No Front SW1 Front SW25 MySQL instance < 4s6 Node OS < 4s 1 Online data disk 117 < 5s 2 (SSD) failure Log/work data disk8 No DB DB DB DB (SATA) failure 5,6 7,89 4 12 Infiniband switch (primary) < 12s 310 Infiniband switch (standby) No11 Front network (port1&2) < 18s 10 9 Internal network12 < 12s Infiniband SW1 Infiniband SW2 (primary & standby) 24
  • 26. Time Required for Online MaintenanceTable Rows and Size Small Medium Large Row 50,000 500,000 5,000,000 Size (byte) 113,639,424 1,063,190,528 10,696,130,560Implementation Time Small Medium Large Create Column 1.6s 13.5 149.8 Create Index 1.6s 13.0s 172.7s Drop Column 1.5s 13.8s 125.5s Drop Index 0.5s 0.5s 0.5s 25
  • 27. Impacts During Online Scheme Modification No impact on access performance in areas other than those subject to work operations Some impact on performance of access to table being subject to work operations (taking periods with little impacts, such as night service, into consideration)  Online execution – 5 million cases, total tables 10G 26
  • 28. Clustrix Implementation Impacts Release from Sharding (1)Before …… DB DB DB DB …… No more sharding!After DB DB DB DB …… +
  • 29. Clustrix Implementation Impacts Release from Sharding (2) No need for correction of application No need for DB distribution Sharding production costs reduction (over 90%) for both application engineer and DBA  In case of large-scaleas-i s sharding project, actual production costs DBA compared toto-be APP original 0 2 4 6 8 10 12 14 m an-m onth 28
  • 30. Clustrix Implementation Impacts Cost Reductions due to Consolidation (1)  Sufficient performance scalability  Fault-tolerance ready for mission critical  No data loss  High online maintainability that doesn’t affect other services  Possibility of consolidation to Clustrix of existing MySQL database 29
  • 31. Clustrix Implementation Impacts Cost Reductions due to Consolidation (2)  Consolidation of all existing MySQL within Clustrix  Number of servers will be reduced to 10%  Monthly system costs will be reduced to 40% 30
  • 32. Back-up Structure ClustrixDB DB DB … Node 1 Node 2 Node 3 Replication Slave as first backup Backup by mysqldump MySQL DB NFS NAS 31
  • 33. Data Migration Procedure  Replication to DEV for verification  Replication to PRO for migration  Conversion of application access point to PRO MySQL DB Replication Replication Clustrix DEV Clustrix PRO DB DB DBDB DB DB 32
  • 34. Other Advantages of Clustrix Auto-Defrag Cordial Support Service  Advice regarding structure  Troubleshooting  Tuning advice  Etc. 33
  • 35. Operational Issues Resolved with Clustrix Data sharding operations  Unnecessary, operational cost reduction Data protection, HA securing  Possible Online maintenance  Possible Tendency for large number of units  Consolidation possible  Cost reduction 34
  • 36. Clustrix at Rakuten An important database platform Provided as Database-as-a-Service No lead-time Usage volume rate structure 35

×