...
See the archives for the complete discussion thread: avoiding dynamic metadata queriesTBD
- The Metadata Query Protocol is a scheme for addressing metadata resources, it is not a metadata query language per se.
- Everything of interest is known in advance and can therefore be pre-computed. In particular, signed per-entity metadata can be pre-computed.
- The current MDQ protocol spec mandates that an MDQ server MUST serve up an "all entities" aggregate at a pre-defined endpoint. At the protocol level, this requirement should be relaxed: An "all entities" aggregate MUST be served at such-and-such an endpoint location but a server is NOT REQUIRED to provide such an aggregate.
- Our initial deployment of an MDQ server should not expose an "all entities" aggregate. We want deployers to migrate to the MDQ server to leverage per-entity metadata, not a full aggregate. Those deployments that need or want a full aggregate should continue to get it from md.incommon.org, at least for the foreseeable future.
Subject: implementing a cache on the client
...