database - Replication Modes Definitions? -


I am checking various NoSQL and RDBMSes in relation to various replication capabilities to create distributed systems.

Reading through several papers and books, I feel that some vendors or writers have used their definitions regarding the terms

  • Master-Master Replication (Replication Between Two Servers)
  • Master-Slave Replication (Replication Between Mutliple Server to Increase Reading Speed, Writes It Is Only Enabled for Master Server)
  • Multi-Master Replication (= Peer-to-P
  • Merge recurrence (?) Ul>

    Example: Some mix defines master-master and peer-to-peer as the same, whereas for example in Maxil Docus, I have Master Master and Multi-Master (= Peer-to-Peer ?? ?) The difference between replication is considered.

    Where is the difference in multi-master and peer-to-peer copy? Is the use of multi-master replication more oriented towards case clustering, while the peer-to-peer target distributed content is on distributed content?

    I would like to sort things out and make sure that I understand these conditions correctly, so perhaps here discussion will help to merge some knowledge.

    Regards, Chris

    Edit: Merge replication added to the list and some explanations such as I understand them ...

    < About couchDB, the story is simple:

    Only one replication mode for CouchDB The source copies all of your data to the goal, under optional yes / no filter, I have mentioned in another question. The main point is that "Replication" is simply a DB customer connecting it with both coaches, reading from source, and writing the target.

    Any other large picture architecture (peer-to-peer, multi-master, master-slave) is the only implementation of developers or system administrators, for example, if GET s Distributed in coaches, but POST goes to a central couch that replicates others, which is effectively master-slave, if you put a couchbedi in every big city for performance , And they are with each other Ide repeat, which is multi-master replication.

    A "peer-to-peer" replica is a concept within CouchDB community, and especially from Chris Anderson's projects and presentations, where CouchDB is everywhere: mobile phones, data centers, telephone poles and without central authority or architecture, there is a replicated duplication in a decentralized manner.

Comments

Post a Comment