Changes

Identity/AttachedServices/StorageServiceArchitecture

251 bytes removed, 05:25, 30 April 2013

no edit summary

* Each master synchronously replicates to a '''hot standby''' in the same DC, to guard against individual machine failure.

** This may not be necessary, depending on instance reliability and our uptime requirements.

* Each master asynchronously replicates to a '''warm standby''' in a separate DC, to guard against whole-DC failure.

** Or maybe to multiple independent DCs. Same same.

** The warm standby should probably be a synchronous replication pair. Because symmetry. And failover.

* All sharding logic and management lives in a stand-alone proxy process, so that it's transparent to the application.

Confirm

358

edits