Title (Goal) | Autoscaling Servers |
---|---|
Primary Actor | Infrastructure Architects, Engineers, and Moonlighters |
Scope | Filesystem |
Level | Low |
Author | Brad Spry |
Story (A paragraph or two describing what happens) | As an infrastructure engineer, I want Fedora's storage to be shareable in read/write fashion to a multiplicity of servers. Specifically, |
...
Fedora's objectStore, resourceIndex, and datastreamStore |
...
should all be asynchronous friendly |
...
. Additional servers |
...
could then be programmed to launch automatically based upon a load threshold, and in reverse, programmed to terminate |
...
when there is no load. Scale up with massive ingests, and scale down when ingest is complete. Scale up when a public digital collection suddenly becomes very popular, and scale down when popularity subsides. |