POC using the MEMSQL

Could Please someone throw the light on this topic. This is a POC that am doing on the memSql.
It would be a great help if someone can give the logic on how it can be done.

  1. Create tables in MemSQL during load each table in MemSQL will act as a separate cache of its own. There would be around 10-15 tables created initially.
  2. Each table can have two columns.
    a. Key – This key will be used to add in the table and used to retrieve. This would be of data type string.
    b. Value – The value associated with the key. The data type can be a blob,text,xml, binary as needed
  3. The value in the tables should get deleted automatically after it reaches the expiry time. The expiry time should be a configurable parameter.
  4. Every time the MemSQL instance is restarted, there should not be any entries. We can start with a clean state.
  5. Ability to add entries into the table once created.
  6. Ability to retrieve entries from the table if the key exists
  7. Ability to delete entries manually or if the expiry time is reached.

Thanks
Phaneendra

1 Like

In production, what would your data size be?

How many concurrent operations would you do per minute?

If you lose an node, are you okay resetting the data, or do you want high availability that can tolerate loss of one node?

This POC is for one of the biggest bank in USA and so the data volume will be in trillions.
if we lose a node, we are ok to reset the data.

Thanks

Hello @Phaneendra. Thanks for the post with your questions. Since this gets deep into specific areas, I am going to reach out directly to discuss and get further details about your use case to see how we can help.

Mike