Pfrields/AWS ideas

Template

Idea: Description; Cost: (include $$ estimate, but also any code development or process changes required); Benefit: What portion of the community or project benefits?; Risk: What happens if the service goes away later?; Time Horizon: How long would it reasonably take to make this work? (including code/process time)

EC2 internal Fedora mirrors: One or more mirror instances inside EC2 using traditional FTP/HTTP services. Possibly some metalink work?; Cost: Unknown, but likely in the hundreds of dollars a month to start. Since this currently costs AWS users money, it's a chicken and egg problem. May require commitment from someone at Amazon to maintain.; Benefit: Very fast updates for people running current Fedora AMIs; Risk: Low to project. Would potentially impact users accustomed to the benefit and possibly drive them to other solutions.; Time Horizon: Short.

Proxy or other servers: One or more proxies or other services for APAC/EMEA.; Cost: Likely high, several thousand dollars a month per proxy?; Benefit: Increased

Running a full test suite on Cloud AMI: Cost: ? (higher compute time); Benefit: Higher quality of release to all Cloud edition users; Risk: Must fall back on manual testing