From HPC Wiki
Jump to: navigation, search

There are cases in which on-premise resources may not be sufficient for a researcher's needs. Important factors are :

  • Nature of the research
  • Time frame allocated for results
  • Collaboration constraints
  • Data sharing constraints

Such cases may include insufficient on-premise :

  • Number of cores
  • Available RAM
  • Storage
  • Node interconnect speed
  • Disk access speed
  • Visualization capabilities

Off-premise resources

There are several vendors that provide off-premise computational and/or storage resources, for a fee. Whether use of such resources makes sense for the researcher is dependent on many variables, as listed above.

In making a decision on using off-premise resources in addition to, or instead of, on-premise resources, researchers are encouraged to consult with Academic and Research Computing Systems,

Resource type Example
computation, storage Amazon Web Service (AWS)
computation, storage Microsoft Azure
computation, storage Google Cloud
computation, storage Rackspace
storage only Box
storage only Dropbox
computation grid Xsede

Research grants from cloud providers

Several companies that offer cloud computing have programs that fund research grants. CloudResources