[Distributed File System] hash tree

Hash tree

The hash tree can also be used for file synchronization in a distributed file system. Once a node changes, its hash changes.

Quickly spread to the root node. The system to be synchronized only needs to misquery the hash of the root node.

The tree structure can be used to locate the changed content at the logn level and then synchronize it.

This function is useful for multiple subsystems to share some data, and for regularly updating backups.

You only need to adjust the regular query of the hash tree and then check which nodes need to be updated synchronously.

It seems that there is another question: Is it instant synchronization or a certain delay allowed?

The Distributed File System has a cap principle:

Distributed, good scalability (P)

Eventual consistency (c), that is, data inconsistency in a short period of time

High Availability (A), failure of some nodes does not affect the service