HA of Plasma

There are some design that very useful for the redfish data modeling by using etcd as persistent storage for Plasma.

1) Key-value design and JSON ready
This design and modeling is suit for the redfish/json modeling we use in the plasma/PODM , we don’t need to care how to design a SQL schema to satisify redfish/json case, they all a json object and can put into etcd object.
Example
https://futurestud.io/blog/coreos-read-write-values-to-etcd
https://coreos.com/etcd/docs/latest/api.html

2) Watching
This help us to monitor redfish resources state without using polling way
3) Replication
This is native supported in etcd including active-active model if we need multiple persistent storage for plasma backend.
4) write per seconds to 10000 transaction
http://thenewstack.io/intel-gives-the-etcd-key-value-store-a-needed-boost/