NetflixAWS         Cassandra
Netflix API – Growth in requests                                 25Requests in Billions (per day)                         ...
us-west-2                    1c                                          1c     1a        1e                         1d   ...
23
24
25
26
Client Writes/s by node count – Replication Factor = 31200000                                                             ...
50          Number of production clusters15          Number of multi-region clusters4           Max regions, one cluster10...
http://techblog.netflix.comhttp://netflix.github.com
We are sincerely eager tohear your FEEDBACK on thispresentation and on re:Invent. Please fill out an evaluation   form whe...
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Optimizing Cassandra in AWS
Upcoming SlideShare
Loading in …5
×

Optimizing Cassandra in AWS

716 views

Published on

Optimizing Cassandra in AWS presentation from 2012 AWS Reinvent conference.

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
716
On SlideShare
0
From Embeds
0
Number of Embeds
5
Actions
Shares
0
Downloads
10
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide
  • … Developer in house …Quickly find problems by looking into codeDocumentation/tools for troubleshooting are scarce… repairs …Affect entire replication set, cause very high latency in I/O constrained environment… multi-tenant …Hard to track changes being madeShared resources mean that one service can affect another oneIndividual usage only growsMoving services to a new cluster with the service live is non-trivial… smaller per-node data …Instance level operations (bootstrap, compact, etc) are faster
  • Optimizing Cassandra in AWS

    1. 1. NetflixAWS Cassandra
    2. 2. Netflix API – Growth in requests 25Requests in Billions (per day) 20 15 10 5 0 Data Center Capacity
    3. 3. us-west-2 1c 1c 1a 1e 1d S3 1b 1d 1e 1b 1c Durability (Quorum)1c One instance:  1c 1a 1a Availability zone:  Replica set:  1e 1d 1c 1b 1d 1e 1b 1c 1c S3 1a S3 us-east-1 eu-west-1
    4. 4. 23
    5. 5. 24
    6. 6. 25
    7. 7. 26
    8. 8. Client Writes/s by node count – Replication Factor = 31200000 10998371000000800000600000 537172400000 366828200000 174373 0 0 50 100 150 200 250 300 350
    9. 9. 50 Number of production clusters15 Number of multi-region clusters4 Max regions, one cluster101 Total TB of data across all clusters780 Number of Cassandra nodes72/32 Largest Cassandra cluster (nodes/data in TB)250k/800k Max read/writes per second on a single cluster
    10. 10. http://techblog.netflix.comhttp://netflix.github.com
    11. 11. We are sincerely eager tohear your FEEDBACK on thispresentation and on re:Invent. Please fill out an evaluation form when you have a chance.

    ×