Re: [Fwd: RE: LC: draft-ietf-forces-protocol-10.txt>>ForCES message size?]

As I understood the TP flags, those were for minding together 
transactions, not for binding together pieces of messages.  Each 
message in a transaction stream is still a properly structured PL 
message.  So that does not provide fragmentation of PL messages for 
the case where the outermost TLV is "too long."

This just doesn't seem to me a big enough issue to worry about at 
this time.  If we find there is a problem, we can try to come up with 
sensible solutions later when we know what is causing the difficulty.

Yours,
Joel

At 02:20 PM 6/14/2007, Jamal Hadi Salim wrote:

>One should be ablke to use the TP flags (start/middle/end) in the main
>header to indicate multipart messages even in no-transactional actives
>such as a query response from the FE --> CE or events in the same
>direction.
>Would that solve the problem. The TML level and IP level will take care
>of their own fragmentation/segmentation.
>
>Does that address the issue?
>
>cheers,
>jamal