>Ahhhhhhhh.......
>
>that old debate........ ( i am waiting for howard to comment )
>
>but as far as your answer is consernced ......it`s distance vector HYBRID
>
>a lot of the instructor`s i know read it like this .....
>
>EIGRP is based on IGRP ...so it`s distance vector...
EIGRP shares metrics (more or less) with IGRP, and the configuration
commands are very similar. Otherwise, they are completely different
protocols.
>EIGRP behave`s like a link state protocol ......so it`s link state
The only similarity between EIGRP and LS protocols are that they
all use hello subprotocols and send changes only. That's a historical
accident. There is absolutely no reason DV can't use change-only
updates -- it's just the concept wasn't invented at the time RIP
and IGRP were designed.
>EIGRP is cisco`s own ...uses the best of both world`s ...so it`s a HYBRID
>and should really have it`s own clssification....
>
>BUT
>
>according to the designer`s (i`m told) it IS a distance vector
>protocol..With extensions
Extensions only in the sense that OSPF and ISIS have had extensions
over the years.
DUAL was developed by JJ Garcia-Luna-Aceves while at Stanford
Research Institute, now at University of California Santa Clara. He
was not involved in Cisco's EIGRP implementation. His major area of
research is distance vector protocols, and his formal papers on DUAL
clearly identify it as a distance vector algorithm.
http://www.cse.ucsc.edu/~jj/pubs.html
Indeed, he hasn't stopped with DUAL. I've had conversations with him,
and he definitely feels there are even faster DV algorithms than DUAL.
From my book, "Designing Routing and Switching Architectures,"
The original routing protocols were distance vector. They have
continued to evolve. First-generation distance vector protocols, such
as RIP, use periodic update (and may use triggered updates), loop
detection through count to infinity, hop count metrics, and loop
prevention through split horizon and holddown. Perhaps the most
important distinction between the two generations is the way in which
they use hop count. First-generation protocols use hop count for a
wide range of functions, including metrics and loop detection.
Second-generation protocols have more intelligent metrics and
additional loop detection mechanisms.
Second-generation distance vector protocols, such as IGRP, use a
combination of periodic and triggered updates, loop detection by
detecting increasing metric, complex metrics, and loop prevention
through split horizon and holddown. Holddown can be turned off in
most cases because the loop detection mechanism will allow only
transient loops. Second-generation distance vector does not have a
hello mechanism.
Many courses and books are incorrect in the way they describe hello
subprotocols, if they say that only link state protocols have hello
mechanisms. Hello subprotocols are an alternative to using periodic
updates as a keepalive mechanism, and have nothing to do with link
state or distance vector. Historically, hello subprotocols first
appeared in link state protocols, but that is a coincidence.
...
EIGRP is another evolutionary step in distance vector. Internally,
EIGRP is a completely different protocol than IGRP. The only
similarities are in the configuration commands and the metric
computation. IGRP is upwardly compatible with EIGRP in the sense that
EIGRP can accept routing updates sent to it by IGRP, not that the
internal algorithms are at all similar. See "IGRP to EIGRP Migration"
in Chapter 12, "Special and Hierarchical Routing Topologies."
EIGRP's algorithm is called the diffusing update algorithm (DUAL). It
was developed by J. J. Garcia-Luna at Stanford Research Institute.
Garcia-Luna was not involved in Cisco's implementation of EIGRP.
Several of the Cisco product architects have made strong arguments
that properly designed distance vector algorithms have distinct
technical superiority over link state. Link state partisans point to
the evolution of their protocol family.
EIGRP completely unloaded hop count and removed it from any role. It
uses the same metric as IGRP, but has separate hello subprotocol and
reliable update mechanisms that do away with most distance vector
looping problems.
In RIP and IGRP, loops are formed by old information. A significant
reason that old information is propagated is the use of periodic
updates. When the information distribution mechanism becomes
change-only, there is no periodic updating with stale information. To
have a change-only mechanisms, you need a hello subprotocol so you
can detect dead routers. Until you can reliably detect dead routers,
you cannot distinguish between not receiving an update because there
has been no change, and not receiving an update because the router
that would have sent it is down or unreachable.
>
>in the exam`s (lets keep this relevant to why were here ..and no i don`t
>mean on the planet ....just the list)...it will be classified on the exam as
>distance vector...
I agree that Cisco exams will probably look for hybrid or enhanced
distance vector.
>
>HTH....pppppplease don`t flame me.....Awww go on then ....if it makes you
>happy
>
>steve
>
>
>>From: "CCIE TB"
>>Reply-To: "CCIE TB"
>>To: [EMAIL PROTECTED]
>>Subject: Is EIGRP a DV or LS protocol [7:10657]
>>Date: Mon, 2 Jul 2001 12:02:38 -0400
>>
>>Hi Group members,
>>
>>Is EIGRP a Distance Vector or a Link State protocol. I thought it is a Link
>>State until I read Cisco BSCN book, which classify it as both. Is that
>>possible.
>>
> >Regards to all
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=10682&t=10657
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
