How Riak S2 works
1. Riak S2 is built on the Riak NoSQL
database. 2. Try to upload a database to Riak S2. This is what happens: a. Riak S2 breaks the object into smaller chunks or blocks so they can be streamed, stored and replicated in the underlying Riak cluster. b. Each block has meta data associated with it. c. Meta data is used for retrieval purposes. d. Data is replicated across nodes. e. Nodes automatically take over the responsibilities of nodes that go offline. f. Hence, data remains available even in failure. 3. Any node is capable of responding to client requests. 4. There’s no master node architecture here, and each node has the same responsibilities. 5. Three replicas per object (data) are created, by default. 6. For failed or non-communicative nodes, other nodes automatically take over the responsibility to make data available and, hence, high availability and fault tolerance is achieved.