From: Scott Smith (hioctane@gmail.com)
Date: Fri Aug 31 2007 - 19:48:01 ART
Exactly, Gary. This "extra" step only comes into play when the type2s
metric is a tie.
Just for some addtional info, this exact situation is mentioned in:
http://www.cisco.com/warp/public/104/1.html#t32
"If the external routes are both type 2 routes and the external costs
to the destination network are equal, then the path with the lowest
cost to the ASBR is selected as the best path"
I had a quick peek at the RFC and came up with this... *hopefully* I
correctly interpreted RFC into English :-)
We use type1 if we care about the internal cost to the external route.
In other words, the internal cost is more important than the external
cost.
We use type2 if we don't care about the internal cost. In other words,
the external cost is more important.
If you all need anything deeper than that or to verify my
interpretation you're going to have to suffer through the RFC :-)
One more note here... there are other "tie breakers" within OSPF which
are used for other situations. For example, a router has 3 equal
internal routes to a destination but the max-paths only allows 2 to be
installed in the routing table. So, which 2 of the 3 get installed in
the routing table and why? If you don't know I strongly suggest that
you look into it :-)
-- Scott CCIE #17040 (R&S)On 8/31/07, Gary Duncanson <gary.duncanson@googlemail.com> wrote: > > > > From the looks of things..(Im sure Scott will correct me)..that is the whole > point *except* when the E2 route metrics are the same, in which case the > cost to the ASBR for the advertised routes is also checked to see if they > are also the same giving a tie. > > In the scenario where the E2 metrics are different the cost to ASBR wouldn't > need to be checked. > > With E1 routes the cost to ASBR is always added to the metric. > > At least that's how it looks to me so far. > > Scott? > > ----- Original Message ----- > From: Herbert Maosa > To: Scott Smith > Cc: Gary Duncanson ; ccielab@groupstudy.com > Sent: Friday, August 31, 2007 9:35 PM > Subject: Re: OSPF external type-2 route load balance > > Thrown a little offset here. I thought the whole point of type-2 external > route was that the cost to ASBR was irrelevant ? If type-2 is looking at the > cost to the ASBR, then what is the difference with type-1 ? > > > On 8/31/07, Scott Smith <hioctane@gmail.com> wrote: > > While that is true that type1 metric changes along the path and the > > type2 is constant this isn't the point of the show output. > > > > What I've pointed out is there is another metric in play with > > externals aside from the metric of the route. The router also > > considers the cost to the ASBR when deciding if externals are truly > > equal. For two externals to be considered equal everything must be > > equal. Both the metric of the route as seen via "sh ip route" and the > > metric to the ASBR seen via "sh ip ospf border-routers" must be equal. > > > > -- > > Scott > > CCIE #17040 (R&S) > > > > > > On 8/31/07, Gary Duncanson <gary.duncanson@googlemail.com > wrote: > > > Thanks for that Scott. > > > > > > Looks like this demonstrates this rather well, at least how I read it. > > > > > > From OSPF Design Guide on CCO. > > > > > > sic 'External routes fall under two categories, external type 1 and > external > > > type 2. The difference between the two is in the way the cost (metric) > of > > > the route is being calculated. The cost of a type 2 route is always the > > > external cost, irrespective of the interior cost to reach that route.' > > > > > > Rgds > > > Gary > > > ----- Original Message ----- > > > From: "Scott Smith" <hioctane@gmail.com> > > > To: "Sadiq Yakasai" < sadiqtanko@gmail.com> > > > Cc: "Joseph Brunner" <joe@affirmedsystems.com>; "Herbert Maosa" > > > < asawilunda@googlemail.com>; "ISolveSystems" > <support@isolvesystems.com>; > > > "Cisco certification" < ccielab@groupstudy.com> > > > Sent: Friday, August 31, 2007 7:49 PM > > > Subject: Re: OSPF external type-2 route load balance > > > > > > > > > > Bingo! > > > > > > > > A1 learns 100.100.100.0 from both 10.43.1.61 & 10.43.1.73. > > > > > > > > Before - BW 1536 to both ASBRs > > > > > > > > A1(config-router)#do sh ip route ospf > > > > 100.0.0.0/24 is subnetted, 1 subnets > > > > O E2 100.100.100.0 [110/20] via 10.43.1.61, 00:00:36, Serial1/0.1 > > > > [110/20] via > 10.43.1.73, 00:00:36, > > > > Serial1/0.2 > > > > > > > > > > > > After - Dropped BW to 512 on s1/0.1 > > > > > > > > A1(config-subif)#do sh ip route ospf > > > > 100.0.0.0/24 is subnetted, 1 subnets > > > > O E2 100.100.100.0 [110/20] via 10.43.1.73, 00:00:06, Serial1/0.2 > > > > > > > > > > > > Then changed BW back to 1536 on s1/0.1 (now both are equal again) > > > > > > > > A1(config-subif)#do sh ip route ospf > > > > 100.0.0.0/24 is subnetted, 1 subnets > > > > O E2 100.100.100.0 [110/20] via 10.43.1.73, 00:00:02, Serial1/0.2 > > > > [110/20] via > 10.43.1.61, 00:00:02, > > > > Serial1/0.1 > > > > > > > > The metric is a constant 20, however, for the externals OSPF also uses > > > > the cost to the ASBR. > > > > > > > > -- > > > > Scott > > > > CCIE #17040 (R&S) > > > > > > > > > > > > On 8/31/07, Sadiq Yakasai < sadiqtanko@gmail.com> wrote: > > > >> Can you show the Type 4 LSA database as well? > > > >> > > > >> I think the cost to the ASBR has a role to play in selecting 2 Type 5 > > > >> LSA's reporting the same prefix. > > > >> > > > >> I am not quite sure about this though. > > > >> > > > >> > _______________________________________________________________________ > > > >> Subscription information may be found at: > > > >> http://www.groupstudy.com/list/CCIELab.html > > > > > > > > > _______________________________________________________________________ > > > > Subscription information may be found at: > > > > http://www.groupstudy.com/list/CCIELab.html > > > > > _______________________________________________________________________ > > Subscription information may be found at: > > http://www.groupstudy.com/list/CCIELab.html > > > > > > -- > Kindest regards, > hm
This archive was generated by hypermail 2.1.4 : Sat Sep 01 2007 - 11:32:14 ART