maven-proxy is a lightweight servlet engine that sits on your main server and imitates ibiblio (for instance). When it gets a request for a jar it doesn't have, it gets the jar from the upstream servers and stores it into it's cache before returning it to the requester.
How does this differ from a normal proxy cache?
- The directory structure of the cache mimics a local repository
- It can talk to multiple upstream repositories - ibiblio, codehaus, open symphony etc, providing a unified view for local applications.
- It is persistent, most caches expire their content
- It is simpler than setting up a full blown proxy cache
- It is quicker than mirroring the entire of ibiblio (not to consider the waste)
- It is maven aware - eg. SNAPSHOTs can be given a schedule for removal etc.
- It is self contained - it's not a general purpose proxy