Re: [rbridge] TRILL WG last call on draft-ietf-trill-adj
Radia Perlman <radiaperlman <at> gmail.com>
2011-04-07 13:47:49 GMT
I agree that Donald (and the document) are correct. Section 4.1 is
about suspending ports when multiple things are using the same MAC
address on a links, and section 8 is about an RBridge detecting that
multiple of its (unsuspended) ports are on the same link. These can
be used for load splitting and do not need to be suspended (since they
are not using the same MAC address).
On Thu, Apr 7, 2011 at 6:18 AM, Donald Eastlake <d3e3e3 <at> gmail.com> wrote:
> On Thu, Apr 7, 2011 at 5:31 AM, sujay gupta <sujay.ietf <at> gmail.com> wrote:
>> I have read the draft and support it.
>> There is some clarity required at Section 4.1
>> Operation of the port is suspended because there is a higher
>> priority port on the link with the same MAC address (SNPA).
>> It should have been the "with the same LAN ID" as also mentioned in
> I do not agree.
> The LAN ID is a seven (technically length of System ID +1) byte field
> in the fixed header portion of IS-IS Hello PDUs. It can be viewed as
> the name of the link as assigned by the Designated Intermediate System
> (DIS / DRB). In the steady state, every Intermediate System / RBridge
> port on the link that is sending Hellos will be sending those Hellos
> with the LAN ID field set to the value that is being dictated by the
> DIS / DRB. Thus, if a port was suspended when it received a Hello from
> a higher priority RBridge with the same LAN ID as it was sending, all
> ports would be suspended except for the highest priority port of the
> DRB. Since a suspended port cannot send or receive TRILL Data frames,
> it would be impossible to send data in the campus except between end
> stations that happened to be attached to the same RBridge.
> Thus, I believe that MAC or Sub Network Point of Attachment address is correct.
>> Section 8, RB1 can detect
>> this condition based on receiving TRILL LAN Hello messages with the
>> same LAN ID on multiple ports.
> The condition referred to is one RBridge having multiple ports on a
> link. (This is not the same as there being two or more ports on the
> link with the same MAC/SNPA address, which is what suspension is based
> on.) Since, in the stable state, every RBridge port on the link will
> be sending Hellos with the same LAN ID, the text you quote from
> Section 8 is one correct way for an RBridge to determine if it has two
> or more ports on the same link.
> Donald E. Eastlake 3rd +1-508-333-2270 (cell)
> 155 Beaver Street
> Milford, MA 01757 USA
> d3e3e3 <at> gmail.com
>> On Tue, Apr 5, 2011 at 11:21 PM, Erik Nordmark <nordmark <at> acm.org> wrote:
>>> We've received last call comments from a few people, and as a result the
>>> draft has been revised over the last three months, and we've also
>>> created draft-perlman-trill-rbridge-af-00.txt with some separate
>>> BUT I'd like to make sure we have sufficient number of reviewers for
>>> draft-ietf-trill-adj. If you've read it during the last call process
>>> (whether the -01 we started with, the most recent -05, or a version in
>>> between), please send a note to the mailing list.
>>> Please do this even if you didn't find any issues with the document.
> Isis-wg mailing list
> Isis-wg <at> ietf.org