RE: [kmip] Re: KMIP versioning

[Hal Lockhart <hal.lockhart@oracle.com>]

Seems like you need a metadata query which asks, what versions do you support?

 

Hal

-----Original Message-----
From: Bruce Rich [mailto:brich@us.ibm.com]
Sent: Thursday, January 27, 2011 7:38 AM
To: kmip@lists.oasis-open.org
Subject: [kmip] Re: KMIP versioning

One other point to add to the discussion:

We have a bit of a chicken-and-egg issue if we expect to use a QUERY operation to discover things about the server...since a QUERY is itself a KMIP operation, encoded like the rest, it must have a protocol version number in its encoding, so how does a v.next client encode a QUERY to a server?  Does it encode it at v.next and interpret a failure on a QUERY as "try your question again at a lower protocol level"?  This seems uncomfortably close to a "Mr Eight Ball" protocol, but instead of getting a "Try again later" response, we have to interpret a failure as a "Try again earlier" response.

And we also have to discuss how we can make versioning work in the other direction, on server-to-client ops.  But that discussion can probably wait on the client-to-server resolution.  We just need to remember that versioning needs to work duplex, although the mechanism need not be identical.

Bruce A Rich
brich at-sign us dot ibm dot com