Determining WAN T-1 bandwidth application requirements

[mbroberson1]

Hey guy's just curious for some advice on favorite methods for calculating WAN voice and application bandwidth requirements over a T-1.

For example:

You need to come up with the bandwidth reqiurements for VoIP and a high priority application for implementing a QOS policy.

Say you have a site connected by a full point to point T1 and the remote site runs some VoIP and has a highly important Web based application.Traffic from both of these comes back across the WAN link. You are familiar with setting up QoS you just need to prioritize traffic, but need to calculate the bandwidth requirements for Voice and the web application. The web application you will prioritize based on the servers ip addresses being hit.

Current tools available:

Solarwinds

[Collin Clark]

You forgot your most important tool, IOS! You can use auto-qos discovery. It even creates the policy for you.

http://www.cisco.com/en/US/docs/ios/12_3t/12_3t7/feature/guide/ftautoq2.html

If you prefer to use SW, you could use Netflow (also available in IOS of course).

Hope that helps.

[Joseph W. Doherty]

If your question really is about calculating bandwidth before deployment, it's rather complex to do so correctly, and for traffic supporting a web application, unless someone can predefine how the web application should behave with regard to its bandwidth needs, it can be about impossible to precompute necessary bandwidth.

If your question is about bandwidth requirements after deployment, there are many options. One of the simplest is checking whether your seeing interface outbound drops. Assuming both LLQ for VoIP and the usual default queues sizes for other traffic, no or little drops likely indicates sufficient (or too much) bandwidth. If there are many drops, then you need to investigate further.

If your IOS supports it (and you licensed for it?), the "QoS Bandwidth Estimation" offers some interesting information. See http://www.cisco.com/en/US/docs/ios/qos/configuration/guide/qos_bndwidth_est_ps6350_TSD_Products_Configuration_Guide_Chapter.html. BTW: one snapshot of the stats is often insufficient since the stats are based on current traffic flowing thorough the interface.

Another method to analyze bandwidth, on the router itself, is to use a 3 level policer that passes all traffic, but the stats tell us what different bandwidth needs are.

e.g.

policy-map x

class Voip

priority 1024

police 256000 conform-action transmit exceed-action transmit violate-action transmit

Router#sh policy-map i

FastEthernet0/0

Service-policy output: x

Class-map: Voip (match-any)

0 packets, 0 bytes

5 minute offered rate 0 bps, drop rate 0 bps

Match: none

0 packets, 0 bytes

5 minute rate 0 bps

Queueing

Strict Priority

Output Queue: Conversation 264

Bandwidth 1024 (kbps) Burst 25600 (Bytes)

(pkts matched/bytes matched) 0/0

(total drops/bytes drops) 0/0

police:

cir 256000 bps, bc 8000 bytes, be 8000 bytes

conformed 0 packets, 0 bytes; actions:

transmit

exceeded 0 packets, 0 bytes; actions:

transmit

violated 0 packets, 0 bytes; actions:

transmit

conformed 0 bps, exceed 0 bps, violate 0 bps

Class-map: class-default (match-any)

45 packets, 2997 bytes

5 minute offered rate 1000 bps, drop rate 0 bps

Match: any

Otherwise, you can perform either SNMP stat polling or Netflow stat analysis, but both, I believe, require more effort, and both, I also believe, don't always clearly show the need for instantaneous bandwidth, which both the "QoS Bandwidth Estimation" and multi-level policer try to capture.