Did you purposely reverse the 80/20 rule? I ask this in all seriousness. We
used to design campus networks with the assumption that 80% of traffic
stays local and 20% is destined for a different LAN. That may or may not be
valid for your network. With intranets and server farms, the rule became
somewhat antiquated. It might be valid to reverse the rule, in fact, but
you would have to study your traffic src/dest pairs to know.
To determine your network utilization, use the show interface command. You
mentioned that you are using show processes cpu, but that might not help
you find what you need. On the other hand, if show interface displays input
queue drops, show processes cpu might help you determine the protocol which
is congesting the input queue.
You also mentioned that you are using show ip traffic. I think that command
is only good for looking at statistics for traffic destined to the router,
not traffic forwarded by the router.
Anyway, there's no cut-and-dry answer to your question. Your math looks OK,
but it may not be based on the actual amount of traffic your end stations
are pumping out and where that traffic is destined to go.
Good luck! Keep us posted on what you learn. Thanks,
Priscilla
At 12:16 PM 3/21/01, ciscosis wrote:
>Hello all,
>
> Can anyone advise/help with this?
>
>Recently we installed a switched network using the Cisco collapsed core
>model.
>Our core/distribution switches are 2 cat 6506's connected to each other
>via a gbit etherchannel providing 2 gbits of bandwidth between the core
>switches.
>Our trunk links from the core are fast etherchannel cat 5 providing 200 mbts
>of bandwidth to each access switch.
>We have 5x cat 5500's providing access points into the network for our end
>stations. Each access switch has approximately 50 end stations connected and
>each end station has a full duplex 100mbits switched connection.
>
>My maths is petty shaky but does this not mean that each trunk link from
>each access switch should ideally be capable of providing 50x100mbits 3D
>5Gbits - 20% for the 20/80 rule 4GBits of bandwidth?
> which, If I am correct means our trunk links are hugely over subscribed!!
>
>Show proc returns an average CPU utilization on the access switches of give
>or take 8%
>The cores show approx 2%
> Show traffic returns very low peak traffic stats (all switches less than
>8%)
>response times are very good.
>
>now for the question .... should I be concerned? If my maths is correct why
>such low traffic/CPU stats, is there a command on the switches to see how
>much bandwidth is being used on the trunks?
>
>TVMIA
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>_________________________________
>FAQ, list archives, and subscription info:
>http://www.groupstudy.com/list/cisco.html
>Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
>
>
>_________________________________
>FAQ, list archives, and subscription info:
>http://www.groupstudy.com/list/cisco.html
>Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
________________________
Priscilla Oppenheimer
http://www.priscilla.com
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
