Re: draft-ietf-ldup-replica-req-12.txt
"Timothy Hahn" <[email protected]>
| Newsgroups | gmane.ietf.ldup |
|---|---|
| Message-ID | <[email protected]> |
Patrik, Thanks for the info. I was of the opinion that a "well-defined" replica topology (discernable by the set of replication agreements created), would address potential congestion. There is no requirement that each replica communicate directly with every other replica. Replicas are required to "forward" updates received to other replicas that have not yet "seen" the updates it knows about. Replicas, then, could act as "gateways", thus using less network bandwidth than if every replica had to communicate with every other replica. But this would not have been understandable from just reading the replica requirements draft. Regards, Tim Hahn Internet: [email protected] Internal: Timothy Hahn/Endicott/IBM@IBMUS or IBMUSM00(HAHNT) phone: 607.752.6388 tie-line: 8/852.6388 fax: 607.752.3681 Patrik Fältström <[email protected]> To: [email protected] Sent by: cc: owner-ietf-ldup@m Subject: draft-ietf-ldup-replica-req-12.txt ail.imc.org 06/13/2002 12:57 PM The document was today approved by IESG, but, the IESG have concerns about avoidance of congestion and over-chattiness of the replication protocol. Today, LDAP is TCP only, and therefore LDUP, but, if LDUP is UDP-based or something else, this have to be taking into consideration. paf