From: Curtis Call (curtiscall@xxxxxxxx)
Date: Mon Apr 02 2001 - 22:55:09 GMT-3
I find it interesting that you consider OSPF to be more rigid than
IS-IS. My opinion was always the opposite. Take virtual links for
example. In OSPF, it is possible to have a backbone that is not physically
connected (it must be logically connected however). But in IS-IS, not only
is all inter-area traffic required to pass over the backbone, but the
backbone itself must be completely physically connected. In addition you
can look at the lack of variety allowed in NBMA networks. OSPF allows
NBMA, point-to-multipoint, and point-to-point whereas with IS-IS you are
stuck with either Broadcast or point-to-point (and I wouldn't recommend
broadcast). In addition, look at the non-backbone areas. In IS-IS every
area is effectively a totally stubby area, but in OSPF you can make it a
stub area, a totally-stubby area, an NSSA, or a typical area. So in my
opinion OSPF is definately less rigid and more flexible than IS-IS. I
agree that it is alot more structured than RIP or EIGRP (or any distance
vector protocol for that matter) but Link State Protocols have higher
structural demands, that is the trade off for their increased performance.
Whenever I compare OSPF and IS-IS I always think of the standards bodies
that created them. OSPF was created by the IETF the same people that gave
us such wonderful RFCs as that of transporting IP packets over Avian
Carriers, while IS-IS was created by ISO, a group that I always think of as
more political than technical after reading Radia Perlman's book and
enjoying her sarcastic remarks about them.
Just my 2 cents.
At 04:38 PM 4/2/01, you wrote:
>One of the fair criticisms of OSPF is that it is inflexible with regards to
>the requirement that all inter area traffic be routed through the backbone.
>Hence the virtual link hack that everyone is warned against, but which is in
>widespread use because OSPF prefers structure over annoying realities like
>cost of transport.
>
>All other routing protocols allow some flexibility here. All other routing
>protocols also make it a lot easier to just throw things together that work,
>but which don't scale well.
>
>IS-IS offers an interesting compromise between the rigid structure demands
>of OSPF and the wild freedom of RIP or EIGRP.
>
>Question to ponder - is it really necessary that all parts of any network
>have full visibility to all routes? Or is the stub network concept far more
>effective for a lot of reasons?
>
>Back to the salt mines. :-O
>
>Chuck
>
>
>-----Original Message-----
>From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
>Peter Van Oene
>Sent: Monday, April 02, 2001 10:23 AM
>To: ccielab@groupstudy.com
>Subject: Re: IS-IS summary route
>
>I think feature rich might be a more accurate term than robust in this case.
>I think that IS-IS is equally feature rich with the exception of OSPF's
>superior support of multipoint configurations. All L1 areas are only
>totally stubby when L2 information isn't leaked into them as is possible in
>current implementations.
>
>In all comes down to figuring out what you need the protocol to do and both
>have their strengths and weaknesses in certain implementations.
>
>Pete
>
>
>*********** REPLY SEPARATOR ***********
>
>On 4/2/2001 at 12:56 PM Steven Weber wrote:
>
> >someon eplease correct me if I'm wrong but I don't believe that IS-IS is
> >as
> >robust in the same ways that OSPF is. In IS-IS everything is centered in
> >the
> >L2 core. All L1 areas are compared to totally stubby OSPF areas so any
> >redistribution that you want to do should be done in the core.
> >HTH
> >Steve
> >----- Original Message -----
> >From: "tom cheung" <tkc9789@hotmail.com>
> >To: <ccielab@groupstudy.com>
> >Sent: Monday, April 02, 2001 12:25 PM
> >Subject: IS-IS summary route
> >
> >
> >> Hi All,
> >> I'm stumped with an IS-IS route summarization problem and needed some
> >help.
> >> Here's the physical connectivity:
> >>
> >> area1 isis area2 rip
> >> ----------r3------------r4------------r5--
> >> isis vlsm e0 s1
> >>
> >> I want to summarize the different isis area 1 subnets and inject them
> >into
> >> RIP. I've tried using "summary-address a.b.c.d w.x.y.z" under router
> >isis
> >> in R4. But the summary route does not show up in r4's route table and no
> >> routes are injected into RIP's domain when I try to redistribute between
> >the
> >> two routing protocols. What seems to be the problem??
> >>
> >> Config for R4:
> >> router isis
> >> summary-address 152.1.1.0 255.255.255.0 level-1-2
> >> redistribute rip metric 1 metric-type internal level-2
> >> net 00.0002.0000.0000.0040.00
> >> !
> >> router rip
> >> redistribute isis level-1-2 metric 2
> >> passive-interface BRI0
> >> passive-interface Ethernet0
> >> network 152.1.0.0
> >> default-metric 2
> >>
> >> Route table:
> >> r4# sh ip ro
> >> Codes: C - connected, S - static, I - IGRP, R - RIP, M - mobile, B - BGP
> >> D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area
> >> N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2
> >> E1 - OSPF external type 1, E2 - OSPF external type 2, E - EGP
> >> i - IS-IS, L1 - IS-IS level-1, L2 - IS-IS level-2, * - candidate
> >> default
> >> U - per-user static route, o - ODR
> >>
> >> Gateway of last resort is not set
> >>
> >> 152.1.0.0/16 is variably subnetted, 8 subnets, 5 masks
> >> i L2 152.1.1.128/25 [115/30] via 152.1.1.65, Ethernet0
> >> i L2 152.1.1.8/30 [115/20] via 152.1.1.65, Ethernet0
> >> i L2 152.1.1.1/32 [115/20] via 152.1.1.65, Ethernet0
> >> i L2 152.1.1.2/32 [115/30] via 152.1.1.65, Ethernet0
> >> C 152.1.2.0/24 is directly connected, Serial1
> >> i L2 152.1.1.4/30 [115/20] via 152.1.1.65, Ethernet0
> >> R 152.1.4.0/24 [120/1] via 152.1.2.2, 00:00:04, Serial1
> >> C 152.1.1.64/26 is directly connected, Ethernet0
> >> r4#
> >>
> >>
> >>
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:29:38 GMT-3