Skip to main content

Multilocation Use Cases

Multilocations allow for the creation of a single, unified cluster that can be physically distributed. This is particularly useful when you need to maintain a single namespace while managing data across different geographical areas or functional groups of servers.

Below are the primary scenarios where Multilocations provide significant architectural advantages.

1. Primary and Secondary Data Centers (Business Continuity)​

In this configuration, active operations are conducted in a Primary Data Center, while a Secondary Data Center maintains a real-time synchronized copy of all data.

  • Benefit: In the event of a total failure at the primary site, operations can be switched to the secondary site with minimal delay and zero data loss.
  • Mechanic: Storage classes are configured to ensure that for every chunk created in Location A, a replica must exist in Location B.

2. Geographically Distributed Cluster (Local Access)​

This use case applies to organizations spread across multiple cities or regions that require a single, unified namespace but want to prioritize local performance.

  • Scenario: Users in London and users in New York access the same MooseFS cluster.
  • Mechanic: Data frequently used by London employees is pinned to the "London" location for high-speed local access, but it remains visible and accessible to New York users over the WAN when needed.

3. Disaster Recovery (DR) Backups​

Unlike the active/passive setup, this scenario focuses on long-term data safety where one main data center handles all active workloads and one or more backup centers act as "Disaster Recovery" sites.

  • Configuration: You can have multiple backup locations.
  • Mechanic: Only essential, non-temporary data is mirrored to the backup locations. Temporary or scratch data can be restricted to the primary work data center to save bandwidth and storage costs in the DR sites.

4. Content Distribution and Streaming​

This is an ideal setup for services that need to ingest data once and serve it to a global audience.

  • Workflow:
    1. Data is initially recorded to a single "Ingest" location.
    2. The cluster automatically replicates this data to multiple "Edge" locations.
    3. Clients connect to their nearest local location for read-only access (e.g., video streaming).
  • Benefit: Reduces latency for the end-user and distributes the network load across multiple geographical points.
Summary

Whether you are looking for High Availability (HA), Disaster Recovery (DR), or Reduced Latency, Multilocations provide the granular control necessary to manage physical data placement within a single logical filesystem.