Chat with us, powered by LiveChat

Blog

Back

Coordinated Peer-to-Peer Architecture

4 Aug 2014 by anth

Coordinated Peer-to-Peer Architecture

ArchiveShuttle employs a coordinated peer-to-peer architecture that is designed for maximum scalability and resiliency.

The coordination is carried out by a CORE server which passes instructions to peer-level agents that are responsible for extraction, ingestion, synchronization, switching etc. This approach ensures central coordination, control and reporting while providing a totally scalable solution with multiple agents carrying out the heavy lifting.

ArchiveShuttle - Architecture

In the simplified example above, which represents the migration of a single archive from Enterprise Vault to Microsoft Exchange 2013, a workflow will be executed as follows:

1) The CORE will identify an available source agent and send an export command detailing the export criteria
2) The source agent will export data to a staging area as instructed
3) Upon successful completion the source agent will send a notification to the CORE
4) The CORE will identify an available target agent and send an import command detailing the import criteria
5) The target agent will import the data directly from the staging area
6) Upon successful completion the target agent will notify the core

At this point the data has been successfully synchronized, the CORE will now issue the necessary commands to source and target agents to initiate the switch, so that the “Sync’n’Switch” is complete.

This architecture is well suited to on-premise and hybrid cloud deployment where the CORE can be located remotely. It also lends itself well to very fast performance using a multi-threaded approach. Real world migration speeds for a  migration of Enterprise Vault from Centera to Netapp are shown below. In this case three export and three import agents were used under the control of a single CORE.

ArchiveShuttle - Extraction Speed