Add File Caching

[lharzenetter]

Currently, the repositories are quering the file system for each request. To speed up the requests, we should include a chaching mechanism.

[koppor]

IMHO, this needs an ADR for the storage: file-based, document store, SQL database, graph database. When starting the implementation, I assumed, the file system is fast enough. a) SSDs are fast enough, b) file system caches itself, c) the number of requests per second is low.

May I ask about the decision driver here?

There is also a Winery implementation based on graph database. Should be somewhere archived at the IAAS. Maybe, this implementation can be included?

Oliver Kopp <[email protected]> on behalf of MBition GmbH.

[lharzenetter]

The problem is, that with a growing repository size, the Topology Modeler's loading time increases rapidly. Additionally, as the Container now also uses Winery as its repository provider, the generation of plans got notably slower because each requests to the repository loads the contents from the files directly. Hence, we should maybe not call it file cache but rather object cache.