mpbxyz / hydrakv

hydrakv

29/11 update

• Focus moved from distribution of work to the storage and retrieval aspects of db design
• Will look to implement a WAL and scalable storage platform

<

HydraKV is a simple key value store for learning purposes.

Longer term learning opportunities include having read only distributed slaves

Operations

• PUT -> Put key and value into dataset
• GET -> Get value from dataset
• DELETE -> Delete k,v from dataset
• GETALL -> Get all keys in dataset

TODO

• Write ahead log
• Remove JSON and have hydra storage 'engine'
• Separate DB server elements from public API elements
• Multiple Datasets
• LOAD from JSON
• DUMP to JSON
• Put operator
  • If you put an existing key what happens?
    • Should overwrite, TODO
• Get operator
• Delete operator
• Figure out how to do operators while checkpointing to file
  • Optimise?
• Get all operator
• Master write
• Slave read

Distribution

Master

• Handles all PUT, DELETE operations
• Interfaces client connection and sends to slave for read
• Sends WAL stream

Slave

• Handles all read operations
• Register with Master
• Reads WAL stream and implements on local db