Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Under the “DBMS-optimized cloud storage” section, won’t replicating the WAL across various different availability zones entail doing an atomic broadcast if you want any isolation?

It seems to me that you would need to run some sort of consensus algorithm to ensure the replication is consistent but that’s obviously very expensive in latency. Is it actually done this way?



They mention paralleling them, so it's only as expensive as the slowest write.


Aurora writes each block to 6 replicas, and waits for a quorum of 4 to acknowledge. This has small impact on commit time which you can see as IO:XactSync (https://dev.to/aws-heroes/aws-aurora-io-xactsync-is-not-a-po...) but provides good HA




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: