Suuchi - Application Layer Sharding

Talk presented as part of Indian Linux User Group - Chennai (ILUGC) December '16 and Digital Ocean January '17 Meetups

References - https://github.com/ashwanthkumar/suuchi-sharding-talk

Transcript

1. ashwanth kumar @_ashwanthkumar principal engineer suuchi - toolkit for application layer sharding
2. from simple to unmanageable beasts how we scale systems? an opinionated view
3. Service
4. Service
5. Service Service
6. Service Service Service
7. Service Service Service
8. Service Service Service
9. Service Service Service
10. rise of KV stores distributed, replicated, fault-tolerant (optionally sorted) 2006 BigTable from Google 2007 Dynamo from Amazon 2009 VoldemortDB from LinkedIn Cassandra from facebook 2008
11. Distributed (NoSQL) Datastores Service Service Service
12. Distributed (NoSQL) Datastores Service Service Service Network Latency
13. data locality for low latency / data intensive applications
14. Service Service Service Co-locate data to improve performance
15. Sharded and replicated to improve throughput Service Service Service
16. Service Service Service Deal with complex distributed system problems at the application layer
17. suuchi github.com/ashwanthkumar/suuchi toolkit for application layer sharding
18. components - transport uses http/2 with streaming
19. components - membership static dynamic fault tolerance in case of node/process failure scaling up/down needs downtime of the system
20. components - request routing Consistent hashing and random trees: Distributed caching protocols for relieving hot spots on the World Wide Web node 2 node 1 node 3 node 4
21. components - request routing Consistent hashing and random trees: Distributed caching protocols for relieving hot spots on the World Wide Web node 2 node 1 node 3 node 4
22. components - request routing Consistent hashing and random trees: Distributed caching protocols for relieving hot spots on the World Wide Web node 2 node 1 node 3 node 4
23. components - request routing Consistent hashing and random trees: Distributed caching protocols for relieving hot spots on the World Wide Web node 2 node 3 node 4
24. - Peer to Peer system - no single point of contact - Each node handles or forwards requests transparently - Uses pluggable partitioner scheme - Can be customized as weighted distribution / Rendezvous Hash etc. components - request routing
25. - define a gRPC service using proto2 (or proto3) - generate the stubs in java / scala - implement the services - connect them together using Suuchi - Server abstraction getting started
26. let’s see some code see reference for actual links
27. suuchi.proto
28. suuchi.proto
29. implement proto service
30. connect using Server abstraction
31. - HTML Archive System - Handles 1000+ rps - write heavy system - Stores ~120TB of url and timestamp indexed HTML pages - Stats (as Monoids) Storage System* - All we want was approximate aggregates real-time - Real-time scheduler for our crawlers* - Finds out which of the 20 urls to crawl now out of 3+ billion urls - Helps crawler crawl 20+ million urls everyday suuchi @indix
32. idea behind suuchi membership, request routing / sharding 2011 Gizzard from Twitter 2016 Suuchi 2016 Slicer from Google 2015 RingPop from Uber
33. questions? references available at github.com/ashwanthkumar/suuchi-sharding-talk