Geoplexing

From Wikipedia, the free encyclopedia
Jump to: navigation, search
For the IBM parallel infrastructure, see IBM Parallel Sysplex#Geographically Dispersed Parallel Sysplex.

Geoplexing is a computer science term relating to the duplication of computer storage and applications within a server farm over geographically diverse locations for the purpose of fault tolerance.[1][2] The name comes from a contraction of geographical multiplex.[1]

Description[edit]

In a geoplex, server clusters are duplicated over one or more geographically separate sites. Geoplexes can be "active-active" (where all of the clusters are used in tandem) or "active-passive" (where one or more of the clusters are kept as a hot spare).[1]

Data and applications are shared either via cloning or partitioning. In cloning each server in a cluster handles one or more applications, the applications and data for which is cloned to other servers in the geoplex, a load balancer then distributes requests to the cloned servers. In partitioning hardware and applications are duplicated in the geoplex and application data is divided between the servers with requests being routed to the correct server.[1]

References[edit]

  1. ^ a b c d B Devlin, J Gray, B Laing, G Spix. "Scalability terminology: Farms, clones, partitions, and packs: RACS and RAPS". arXiv:cs.AR/9912010.
  2. ^ T Barclay, J Gray, W Chong (2004). "TerraServer Bricks–A High Availability Cluster Alternative". Citeseer. CiteSeerX: 10.1.1.59.3626.