On the basic building blocks for GrepTimeDB

[v0y4g3r]

As a summary of the first discussion in our team, there will be currently 3 basic building blocks for GrepTimeDB:

• Proxy: ingests data and serves query requests. Proxy nodes are also aware of data sharding rules so request can be routed to correct data nodes;
• DataNode: handles logical plans generated by proxy nodes, produces physical plans and read/write data from/to storage. Data nodes should provides scalability (clustering), elasticity and reliability(data replication);
• Storage: abstraction layer that hides the subtle difference between various objection storage service on different cloud and provides unified APIs for data storage and retrieving.