Multi-Version Concurrency Control
Concurrency Control
In a typical relational database when we modify a table, we put a lock - and all other clients that want to access the table are queued
- This sequential execution of tasks wastes a lot of processor's power and time:
- under high load it may spend a lot of time trying to figure out whose turn is next
MVCC, Multi-Version Concurrency Control, is one of the Concurrency Control techniques.
- it is used to manage concurrent access to the data
- it runs effectively even under high load, without worrying about queuing requests
(figure source: [1])
This is used in:
Ways to Achieve
B-Tree Storage Engine
This way it is achieved in CouchDB
B-Tree is used everywhere, also for internal data: documents and views
- Usage of this data structure imposes an important restriction: can access only by key.
- Reason: to be make huge performance gains
Modification in B-Trees to support MVCC:
- reads and writes without locking the system
- writes do not block reads
- this is because of append-only design
Append-only design
- old versions are not deleted
- every time something is updated, a new node is created, and a new root as well
- but old reads still have references to the old root,
- so they are able to continue reading without being interrupted,
- i.e. have old consistent state
- data never lost and never corrupted
Idea:
- When a new request comes, it uses the current root for querying
- If the root changes by other transaction, this request is still active and will get the old valid and consistent value