Active/passive replication, eventual consistency, and network partitioning

Active and Passive replication

Active Replication

⬅️ Example: Primary backup replication

This is active replication because the entire operation has been repeated on every replica where its effect is needed.

Use active replication when the updated state might be large.

Passive Replication

➡️ Example: Primary backup replication

This is passive replication because the operation is executed once on the primary node, and then the resulting state is sent to the backups.

Use passive replication when an operation might be expensive to do on each replica.

RAFT’s log replication is Active, because ****every server in RAFT is replicated state machine, commands are replicated in the same order in every machine.

Eventual consistency

<aside> 🤔 Eventual consistency is a consistency model used in distributed computing to achieve high availability that informally guarantees that, if no new updates are made to a given data item, eventually all accesses to that item (at all replicas) will return the last updated value.

</aside>

Eventual consistency

• Reads between two updates at R1 and R2 will return inconsistent values.
• Eventually both replicas will end up with the same result, i.e., the set of concurrent values in this case.
• The sets are unordered, i.e., no any of the values has priority over the other. Now it is up to the application how to resolve this conflict.

Eventual consistency is used in Amazon’s Dynamo

Network Partitioning

Network partitioning happens due to many reasons: Link is down, Router is down, Network is slow etc.

P.S: есть еще такие каракули от Шынназара с онлайн-лекции

In the presence of a network partition:

• In order to keep the replicas consistent, you need to block. (From an outside observer, the system appears to be unavailable.)
• If we still serve the requests from two partitions, then the replicas will diverge. (The system is available, but no consistency.)

The CAP theorem explains this dilemma.