Call for consensu on admission priority changes in QSPEC
"Martin Stiemerling" <[email protected]>
| Newsgroups | gmane.ietf.nsis |
|---|---|
| Message-ID | <[email protected]> |
Hi all, There have been discussions with respect to QoS NSLP QSPEC Template (draft-ietf-nsis-qspec-19.txt) on the nsis mailing list, offline between involved parties, and also during the NSIS session here at the IETF#71 in the NSIS session (see minutes). The QSPEC document shares the IANA registries for RPH priority and the admission priority with these drafts out of the DIME and TSVWG: - draft-ietf-dime-qos- parameters - draft-ietf-tsvwg-emergency-rsvp While the IANA registry for the RPH priority has not been disputed, the admission priority has been. Therefore we run a call for consensus on these three choices for the admission priority in the QSPEC. Note well, that the decision of this can have an impact to the still running IETF last call for draft-ietf-tsvwg-emergency-rsvp. Please hum on the NSIS list for one of the three options listed below until Monday, March 17th, 12pm UCT. This is to confirm the humming of the NSIS WG during the NSIS session at the IETF meeting. 1) Keep Y.2171 admission prioty in draft-ietf-nsis-qspec-19.txt as is: 2) Change in draft-ietf-nsis-qspec-19.txt to admission priority as in draft-ietf-tsvwg-emergency-rsvp-05.txt and remove Y.2171 admission prioty; 3) Keep Y.2171 admission priority and add admission priority as in tsvwg-rsvp-emergency to draft-ietf-nsis-qspec-19.txt. For choice 3) these changes are currenlty proposed: 1) the emergency-RSVP draft should pretty much stay the same. There is no "L" bit addition, and its "Admission Priority" field has local significance within a domain, and no pre-existing values are defined for it, nor is there any associated IANA registry assigned to it. 2) The QSPEC draft will have a similar "Admission Priority" field with the same characteristics as that defined above in (1). 3) The QSPEC draft will add a new field, titled "Y.2171 Admission Priority", that will be associated with the current 3 values defined in the Y.2171. This field is considered globally significant on an end- to-end basis. Note that the emergency-RSVP document will NOT have this additional field. 4) Given the presence of two admin priority fields for the QSPEC, additional rules are added to clarify their application and usage for NSIS domains. 4.1) if the source domain does NOT support Y.2171, then it sets a locally significant value in the "Admission Priority" field, and places all "1"s in the "Y.2171 Admission Priority" to indicate that this field is to be ignored. Downstream domains MUST NOT alter the value of the "Y.2171 Admission Priority" field. 4.2) if the source domain DOES support Y.2171, then a value (eg, 0, 1, or 2) is placed in the "Y.2171 Admission Priority" field AND the "Admission Priority" field. Downstream NSIS domains MUST NOT change the value in the "Y.2171 Admission Priority" field so that end-to-end consistency is maintained. Downstream routers act upon the value placed in the "Admission Priority" field. 4.2.1) Furthermore, if a downstream NSIS domain does not support Y. 2171, it MAY place a locally significant different value in the "Admission Priority" field, which is used by downstream routers within its domain. Subsequent downstream domains that do support Y.2171 make sure that the "Admission Priority" field contains the same value as the "Y.2171 Admission Priority" field. Thanks, Martin [email protected] <== NEW ADDRESS NEC Laboratories Europe - Network Research Division NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London W3 6BL | Registered in England 2832014