Re: Addr: Re: BGP-4 - revised I-D

[Tony Li <[e-mail]>]

   >Yes it is absolutely neccesary.  The order in which routes are
   >compared impacts the outcome of the comparison.

   How is the text already in RFC 1771 section 9.1 not sufficient to clearly
   specify the order?  To summarize, it says that when selecting a route you
   should consider, *in this order*:

   - Local_Pref
   - MED (adding the note that you can only compare MED between like ASes)
   - IGP distance
   - EBGP better than IBGP
   - Router ID

If I understand what Curtis is getting at (50% chance, max), there is some
ambiguity (or at least possibility of misinterpretation) in the MED
comparison step.  One interpretation is:

- group the routes by neighboring AS
- within each group, select the routes with the best MED

which is also what I think is the intended interpretation.  Alternate
interpretations include:

i)
- group the routes by neighboring AS
- within each group, select _a_ route with the best MED

ii)
- across all routes, select the routes with the best MED
[yes, this violates the comparison]

iii) - if route A has a better MED than route B, but both are incomparable
to route C, and route B has a better IGP distance than C, while A has a
worse IGP distance than C, select route B.  [I believe that cisco's
implementation had this bug for a while and that this is now fixed.]

   As far as I can tell this specifies everything which is needed for
   loop-free operation (I do understand the issue and agree with you).  I
   think that this:

   >   One route from the neighboring
   >   AS is selected as best.  The best routes from each of these
   >   neighboring AS are then compared.

   goes from specifying the algorithm (necessary) to specifying implementation
   (unneccessary).  Perhaps 9.1.2.1 could be made even more explicit as to the
   order in which the tie-breaking criteria must be applied, and there could
   be a forward reference from 5.1.4 to make sure people don't miss this.

I would suggest that while it is more specific, it possibly introduces more
bugs (do you select ONE route? what happens if there are multiple routes
with identical best MED?).  Yes, I think that this all boils down to
wordsmithing some more explicit text...  I look forward to the concrete
proposals.

Tony