bit.ly/MZa8Rz
Easy to start a                     No extra costAmazon Linux for Premium Support                      bit.ly/MZa8Rz
Easy to start a                     No extra costAmazon Linux for Premium Support                      bit.ly/MZa8Rz
Easy to start a                     No extra costAmazon Linux for Premium Support                      bit.ly/MZa8Rz
Simple:         Whitepaper - - - >         Marketplace         Cloudformationmedia.amazonwebservices.com/AWS_NoSQL_MongoDB...
NestedNodes + StorageConfigurableAWS CloudFormation:Infrastructure belongs in yoursource control    mongodb.org/display/DO...
3.757.515       Where your data should be17.13034.268.460.5
3.75 m1.medium7.5 m1.large15 m1.xlarge17.1 m2.xlarge     In Memory!!30 m3.2xlarge34.2 m2.2xlarge68.4 m2.4xlarge60.5 cc2.4x...
New Storage Options        • Amazon EBS – Provisioned IOPS volumesEBS     • Deliver predictable, high performance for I/O-...
New Storage Options                   Testing: random 4k reads        One volume: ~200 MongoOPS with some variability, <1m...
New Storage Options         Testing: random 4k reads                                    SSD                             + ...
Stability Tips• Ext4 or XFS, nodiratime, noatime                                     Stable• Raise file descriptor limits•...
• 10 Million records• 44 GB collection• 2,000 ops• 1000% expected  growth
• 10 Million records• 44 GB collection• 2,000 ops• 1000% expected  growth
1000% expected growth  • Current Setup: Replicated
1000% expected growth  Shard  Shard  Shard  Shard
1000% expected growth.. Quarterly  Shard  Shard  Shard  Shard
1000% expected growth.. Quarterly  Shard  Shard  Shard  Shard
Scale Both Ways
SSD Scale                    SSD   SSD     SSD                      SSD       SSD   SSD                    SSD   SSD     S...
Shrink   quickly   too Go from m2.4xlarge to m1.medium in..
Shrink   quickly   too Go from m2.4xlarge to m1.medium in three easy steps so you can save...
Shrink   quickly   too Go from m2.4xlarge to m1.medium in three easy steps so you can save... 90% of the compute cost when...
We are sincerely eager tohear your FEEDBACK on thispresentation and on re:Invent. Please fill out an evaluation   form whe...
Miles Ward – Solutions Architect - @milesward – miward@amazon.com
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012
Upcoming SlideShare
Loading in …5
×

DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012

1,729 views

Published on

MongoDB is one of the fastest growing NoSQL workloads on AWS due to its simplicity and scalability, and recent product additions by the AWS team have only improved those traits. Join us for a deep-dive on MongoDB best practices, including installation, configuration, orchestration, performance, and durability optimization, as well as operational management using tools from AWS and 10gen.

1 Comment
4 Likes
Statistics
Notes
  • Well developed tech. & skills are to be used for human victory.
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total views
1,729
On SlideShare
0
From Embeds
0
Number of Embeds
6
Actions
Shares
0
Downloads
0
Comments
1
Likes
4
Embeds 0
No embeds

No notes for slide

DAT203 Optimizing Your MongoDB Database on AWS - AWS re: Invent 2012

  1. 1. bit.ly/MZa8Rz
  2. 2. Easy to start a No extra costAmazon Linux for Premium Support bit.ly/MZa8Rz
  3. 3. Easy to start a No extra costAmazon Linux for Premium Support bit.ly/MZa8Rz
  4. 4. Easy to start a No extra costAmazon Linux for Premium Support bit.ly/MZa8Rz
  5. 5. Simple: Whitepaper - - - > Marketplace Cloudformationmedia.amazonwebservices.com/AWS_NoSQL_MongoDB.pdf
  6. 6. NestedNodes + StorageConfigurableAWS CloudFormation:Infrastructure belongs in yoursource control mongodb.org/display/DOCS/Automating+Deployment+with+CloudFormation
  7. 7. 3.757.515 Where your data should be17.13034.268.460.5
  8. 8. 3.75 m1.medium7.5 m1.large15 m1.xlarge17.1 m2.xlarge In Memory!!30 m3.2xlarge34.2 m2.2xlarge68.4 m2.4xlarge60.5 cc2.4xlarge
  9. 9. New Storage Options • Amazon EBS – Provisioned IOPS volumesEBS • Deliver predictable, high performance for I/O-intensive workloadsPIOPS • Specify IOPS required upfront, and Amazon EBS provisions for lifetime of volume – 2000 IOPS per volume, can stripe to get thousands of IOPS to EC2 instance • High IO Instances – hi1.4xlargeSSD • For applications that require 10s of thousands of IOPS • Eliminates network latency as a performance constraint to storage + • EBS Optimized Instances • Enables instances to fully utilize IOPS provisioned on an Amazon EBS volume • Dedicated instance throughput to Amazon EBS – 500 Mbps and 1000 Mbps
  10. 10. New Storage Options Testing: random 4k reads One volume: ~200 MongoOPS with some variability, <1mb/sEBS Loaded instance: ~ 1000 MongoOPS with some variability <10mb/s +PIOPS One Volume: 2000 MongoOPS with <1% variability, 3mb/s Loaded Instance: 20000 MongoOPS with <1% variability, 60mb/sSSD Hi1.4xlarge ephemeral: ~64,000 MongoOPS with low variability, ~245mb/s
  11. 11. New Storage Options Testing: random 4k reads SSD + PIOPSEBS Stable
  12. 12. Stability Tips• Ext4 or XFS, nodiratime, noatime Stable• Raise file descriptor limits• Set disk read-ahead• No large virtual memory pages• SNAPSHOT SNAPSHOT SNAPSHOT
  13. 13. • 10 Million records• 44 GB collection• 2,000 ops• 1000% expected growth
  14. 14. • 10 Million records• 44 GB collection• 2,000 ops• 1000% expected growth
  15. 15. 1000% expected growth • Current Setup: Replicated
  16. 16. 1000% expected growth Shard Shard Shard Shard
  17. 17. 1000% expected growth.. Quarterly Shard Shard Shard Shard
  18. 18. 1000% expected growth.. Quarterly Shard Shard Shard Shard
  19. 19. Scale Both Ways
  20. 20. SSD Scale SSD SSD SSD SSD SSD SSD SSD SSD SSDThat’s over½ a million IOPS…
  21. 21. Shrink quickly too Go from m2.4xlarge to m1.medium in..
  22. 22. Shrink quickly too Go from m2.4xlarge to m1.medium in three easy steps so you can save...
  23. 23. Shrink quickly too Go from m2.4xlarge to m1.medium in three easy steps so you can save... 90% of the compute cost when you don’t need it..
  24. 24. We are sincerely eager tohear your FEEDBACK on thispresentation and on re:Invent. Please fill out an evaluation form when you have a chance.
  25. 25. Miles Ward – Solutions Architect - @milesward – miward@amazon.com

×