Suuchi, toolkit tobuild distributedsystems
View Slide
About meSriram@ Indix
About Indix- 840M Products- 1.5B Offers- 50K+ Brands- ~7.5K Categoriesand counting...
Handling Scale
Layered Scaling
Buck stops with DB Scaling
Essentially - Scaling datasystems
rise of KVsDistributed - Replicated -Fault-tolerant - Sorted**optional
Distributed Data Stores &Distributed ComputePlatforms
Unify them?
Existing SolutionsCoProc with HBase
Applications should OWNeverything
What if we can “enable”building them?
Distributed System Primitives- Communication- Membership- Routing / Sharding / Partitioning- Replication- Handling Failure
CommunicationgRPC
MembershipStatic | DynamicConfig-Driven | Atomix
Sharding / Partitioning /RoutingConsistent Hash Ring, Pluggable
Handling Failures
ReplicationSync | Async
Storage*RocksDB | Pluggable* optional
Getting Started- gRPC Service using Protobuf2- Generate stubs & implement them- Connect using Suuchi “Server” abstraction
Server Abstraction- Pluggable membership mechanism- Pluggable routing strategy- Pluggable replication method