Real-Time Big Data Architecture Patterns
- App server writes to Shared in-memory cluster with a TTL and to queue dispatcher.
- Queue Dispatcher asynchronously writes to persistent bigdata database (With out TTL)
- Reads are first look in memory cache database. If it is missed, then look in persistent bigdata database.
BIGDATA ARCHITECTURE PATTERN2 – DISTRIBUTED MERGE & UPDATE
- Read-Modify-Write step in a single RPC
- Saves data flow over network.
BIGDATA ARCHITECTURE PATTERN3 – DISTRIBUTED MERGE & UPDATE
- Good for write once read many workload.
- Embedded database (with TTL) for local in process cache.
- Embedded database used as cache and designed to use Memory and SSD deficiently.
- Bigdata Database works for distributed writes, sharding and failover.
No comments:
Post a Comment