At 9:49 AM -1000 11/10/2000, Frank B. wrote:
>
>3) I've never had a need to use 2 OSPF process but Someone already
>stated it being used to transition/migrate and that seems
>reasonable...but keep in mind you'd have duplicate everything! I would
>imagine the strain on resources, say for the network to synchronize,
>would be enormous for larger networks.
Having multiple OSPF processes doesn't have to have a significant
impact on resources, and, when used appropriately, may even decrease
load.
With two processes, assuming more than one area each, you will have
an additional area 0.0.0.0 and a few unique tables, but, for the same
number of links, the workload should be comparable in some areas and
perhaps decreased in others.
OSPF workload on a router has several major components. Maintaining
neighbor and adjacency relationships, especially with short hello
timers, can be substantial. The per-router load here is a function of
the number of interfaces, the designated router status of each, and
the timer settings. In other words, it is independent of the area
size and depends purely on characteristics of the local router.
The peak processing load often comes from the recomputation of the
OSPF routing table, of which the Dijkstra algorithm is typically the
most intensive part. Dijkstra proper deals with intra-area routes,
while additional, more linear algorithms deal with inter-area and
externals.
For a single computation of the OSPF table, processing workload is
proportional to:
(number of intra-area routes * log(number of routers in area))
+ number of inter-area routes seen in the area
+ number of external routes seen in the area
The more frequent the computation, the greater the CPU load. The
larger the internal size of the area, the greater the CPU load.
So, if you can keep areas small, you reduce the probability that more
than one area will need simultaneous Dijkstra computations.
Something often forgotten is that while any given OSPF routing domain
has one and only one area 0.0.0.0, there is no reason that you can't
have multiple domains interconnected as a "backbone of backbones." I
have used multiple OSPF processes for this purpose, as well as
migration/consolidation (e.g., merging two enterprise networks).
Most commonly, however, when I use a backbone of backbones, I have
individual routers in single OSPF domains, but interconnect their
ASBR's with static routes or BGP. For example, I've built
intercontinental networks that variously had an area 0.0.0.0 for each
continent, and/or for corporate headquarters.
>
> > 3) The books and tutorials all state that "router ospf 6" defines ospf on
>> the router with a process ID of 6. They then all say that you shouldn't
>> define more than one process. Does that mean that you can have a router with
>> the following:
>>
>> router ospf 6
>> network 10.100.0.0 0.0.255.255
>>
>> router ospf 7
>> network 10.200.0.0 0.0.255.255
>>
>> If this is an allowed configuration, what kind of instances would it be used
> > for? Also, exactly what is the process ID used for?
>>
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
