Like many of you, Darby is a busy man. Not only does he support a large healthcare network but he also spends a lot of time sharing his knowledge on-line and on the phone. Yet with a fulltime job and implementations, along with keeping his finger in the dike, he has time to blog. A prolific blogger Darby Weaver has now joined the ranks of the CCIE Flyer’s contributors for 2009. With his first article he is dangling a problem in front of our readers and will share his solutions in February. Stay tuned!
Since Darby works for a living like all of us he has interruptions and here is his response from emails when I pressed him for his first bit for the CCIE Flyer while I was away skiing;
Eman,
I've got myself in a bit of pinch this weekend. I had a change control go south that was very important and visible late Friday night and into Saturday morning.
On Sunday, I had a CCM Module throw an odd opcode to a Sup720 on a 6509 that hosts our "switchboard" for the Hospital and thus was "Priority 1".
Today, I spent doing damage control for both of those.
Tomorrow is the President's inauguration and we are expecting a lot of traffic related / QoS issues with it.
Sorry for the delay. My Video Conferencing guy is in the same boat.
I started it on Sunday but had an emergency to tend to. Let's see how the week gets off. I was exhausted by yesterday. :)
Life in the fast Lane. I have notes and traces for my progress so far from a hard drive I had to perform a "forensic data recovery on" after it died and didn't want to spin up any more...
Darby
February CCIE Flyer - Darby Weaver
How to troubleshoot and resolve Video Conferencing issues related to Quality of Service,
Part I "The Problem"
----------------------------------------------------------------------
If you know me personally then you probably have already figured out that I don't view problems like a lot of people do. Instead, I see them as opportunities to expand my knowledge to new heights and then amplify my skills exponentially as a result of my efforts. So it was with this problem.
Cisco Catalyst QOS is the basis of the problem and of the solution that I used to resolve this issue with Video Conferencing and its unwanted symtoms. I'm going to start out by describing to you what we saw as "the problem", what we saw as opportunities, and finally where we met these opportunities for improvement with the solution.
First the business need:
1. Travel is expensive. The company is geographically disperse. One half of the business is in Delaware and the other half is in Florida - all of Florida. So Video Conferencing is the solution to help mitagate these unwanted but sometimes necessary costs.
2. There is a definite quantifiable need for real-time communications between business units, business partners, and peers. Video Conferencing meets this requirement as well.
Second the infrastructure:
1. Investment in fat pipes. Metro Ethernet is the transport between major sites, 250 mbps or 150 mbps with DS 20-45 mbps as backup links over ATM.
2. The infrastructure and bandwidth are not saturated. Yet there is still a tangible need for QoS.
3. There is a consistent latency of about 25-30 ms between the main Delaware Site and the main Florida Site. In our testing to date all video conferencing calls have used this site as a part of our testing and this has not been an issue consistently. It is something we are looking into taking up with the carrier. Otherwise it is within specifications.
The symptoms:
1. The quality of the video conference drops noticeably during calls with 3 legs.
2. The Video conferencing equipment is able to detect dropped packets and when they are dropping the video quality seen by participants on all legs degrades noticeably resulting in shadows or "green monsters" or just badly pixelated screens.
3. The management staff submit surveys based on their perceptions at each meeting in which video conferencing is used. This has been helpful in helping determine our perceived improvements.
In this series, I'm going to take you through what was seen and how we incrementally resolved the issues and how we are working to resolve any new issues on a site by site basis.