At 11:42 AM 9/25/01, TP wrote:
>
>EIGRP uses minimum bandwidth and total delay to compute metric (at lest in
>its default configuration).
>I'm a little bit confused: delay of each interface is inversely
>proportional to the configured interface bandwidth, so It seems that EIGRP
>metric is affected only by configured bandwidth (f(bandwidth)+ sum of
>f(1/bandwidth)): is this correct?
No, I think I misspoke. Sorry. You can configure both bandwidth and delay
for an interface. Just configuring bandwidth doesn't affect delay. You
would have to configure it separately. (Since delay is inversely
proportional to bandwidth, I would think it would change automatically when
bandwidth is changed, but it doesn't.)
The metric = [K1 * bandwidth + (K2 * bandwidth) / (256 - load) + K3 *
delay] * [K5 / (reliability + K4)]
The K values can be configured with the metric weights command, although
there's generally no need to change them.
Priscilla
>
>Please give a look to the following output.
>I see in the first subinterface BW 3264 and in the second one BW 2544, but
>the same DLY value: 80.
>
>Rome-7206vxr#sh int atm2/0.1
>ATM2/0.1 is up, line protocol is up
> Hardware is ENHANCED ATM PA
> Description: "P-to-P PVC with Milan"
> Internet address is 213.x.y.14/30
> MTU 4470 bytes, BW 3264 Kbit, DLY 80 usec,
> reliability 255/255, txload 1/255, rxload 4/255
> Encapsulation ATM
> 0 packets input, 0 bytes
> 0 packets output,0 bytes
> 99302 OAM cells input, 99302 OAM cells output
>Rome-7206vxr#sh int atm2/0.2
>ATM2/0.2 is up, line protocol is up
> Hardware is ENHANCED ATM PA
> Description: "P-to-P PVC with London"
> Internet address is 213.x.y.74/30
> MTU 4470 bytes, BW 2544 Kbit, DLY 80 usec,
> reliability 255/255, txload 1/255, rxload 4/255
> Encapsulation ATM
> 0 packets input, 0 bytes
> 0 packets output,0 bytes
> 99121 OAM cells input, 99121 OAM cells output
>
>You suggest " you can tweak the interface bandwidth" and my english is
>awful: what does it mean?
>
>Thank you and have a nide day to all,
>Teresa
> Teresa Presutto
> Grapes Italia S.p.A.
> Italy Network Development
> Via Chiana, 1 - 00198 Rome (Italy)
> Tel +39 06 84550.1, fax +39 06 84550.640, mobile +39 348 4719450
> http://www.grapesnet.com>
>>----- Original Message -----
>>From: Priscilla Oppenheimer
>>To: [EMAIL PROTECTED]
>>Sent: Monday, September 24, 2001 9:41 PM
>>Subject: Re: delay and delay [7:20926]
>>
>>With ping, delay is actually measured. The router reports how long it takes
>>to get replies.
>>
>>EIGRP delay is not dynamically measured. Delay of each interface is
>>inversely proportional to the configured interface bandwidth. Total delay
>>for an EIGRP route is a sum of each interface delay, as reported in EIGRP
>>Updates.
>>
>>If you want EIGRP's delay to be somewhat more realistic, you can tweak the
>>interface bandwidth.
>>
>>Priscilla
>>
>>At 02:06 PM 9/24/01, TP wrote:
>> >Group,
>> >what is the relation (if any) between the total delay I see in sh ip
eigrp
>> >topology and the total delay I see in a simple ping?
>> >
>> > From show eigrp topology I see total delay associated to a
point-to-point
>>atm
>> >pvc lower than total delay showed for a E1 hdlc (between the same
>> routers, 1
>> >hop).
>> >If a make an extended ping I experience a lower delay with E1 than pvc
atm
>> >(and, to be honest, this is what I'd like to see)
>> >
>> >
>> >Thanks in advace,
>> >Teresa
>>________________________
>>
>>Priscilla Oppenheimer
>>http://www.priscilla.com
>>[EMAIL PROTECTED]
________________________
Priscilla Oppenheimer
http://www.priscilla.com
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=21029&t=20926
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
