cluis-lopez / distributed-db Goto Github PK

An Object oriented, memory based, NoSQL, distributed, scalable database with lazy eventual consistency

This project tries to emulate what guys at Google and Amazon did with their DataStores or Dynamos. Small fingerprint databases assuming read/write ratios of 10nth/1 or above where consistency is less important than scalability and data durability. Neither transactions nor ACID operations are expected. Just insert or remove objects from datastores.

• Rest API to get-data-from/insert-data into the databases.
• No complex operations into the database engine: the hard job is for the app.
• Recall objects by id or by simple search (using String fields only in first release).
• No updates allowed. Retrieve the object, modify fields, then replace the original
• Rest API to synchronize datasets between database servers
• Many-Master-2-Many-Slaves architecture. Single Master, many slaves in first release
• Easy to dockerize