WAN Transformation is a Huge Project

WAN Transformation Includes Multiple New Technologies

Building on last week's WAN Transformation blog, our WAN Transformation project grew out of our experience earlier this year with WAN acceleration. We realized that we still wanted WAN acceleration, but we really needed significantly more bandwidth to make it useful. Thus, the two underlying technologies of WAN Transformation are significantly more bandwidth and WAN Acceleration. However, there are five major new technology changes. First, bandwidth. As I've highlighted above, this is a significant bandwidth upgrade. Typical US field sites are going from 6 Mbps to 36 Mbps total bandwidth. Plus, we are designing the WAN circuits (using BGP changes) as active/standby, so during the 99.9% of the time the primary circuit is up, all user sessions are using the 30 Mbps circuit. With per-session load balancing and WAN acceleration, this is a better design than two 18 Mbps. This maximizes the available bandwidth to each user session. This way we give the WAN acceleration hardware the most WAN bandwidth we can to optimize COLD transfers per session. When the primary circuit is down, the site has a 6 Mbps backup circuit which is no different than the bandwidth they have today. QoS is there to protect critical applications. This design is being replicated to our international sites also. And, not to be forgotten, the big sites are getting an even bigger upgrade. Our larger field sites are getting between 60 and 100 Mbps total bandwidth, while the corporate hub sites, where the data centers are, will be getting 600 Mbps total bandwidth. We're also upgrading the Internet pipes at the two hub sites since everyone loves their Faceboo....err....cloud based business applications. Now that we have the bandwidth, we need to use that bandwidth. This is where the WAN acceleration equipment comes in. Each site will be getting an appropriately sized WAN accelerator to speed user applications. The WAN accelerators will be able to exploit the bandwidth we have and make the user experience much better than today. Bandwidth + WAN Acceleration = user productivity. That is what we learned earlier this year. You cannot have one without the other two. This forms the basis of WAN Transformation, but it's not the only two things. In fact, to make the ROI work - which I'll detail next week - we had to include many other technologies, particularly voice. After bandwidth and WAN acceleration, we will also be updating our QoS model. Our MPLS carrier provides more queues now and we plan to use them. Plus, with more overall bandwidth, we can move more applications into the critical application queue, removing them from the "bulk" QoS queue. There's also an opportunity to separate Internet traffic from backups, which will make users happier. Finally, we already have a queue for video conferencing, but this will need to be enhanced since one of the biggest selling points of this project was the ability for each site to have Cisco Telepresence. With a 30 Mbps circuit, every site could easily handle a Cisco Telepresence System 1000 while the larger sites could get a 3000 series. The massive 600 Mbps circuits at the hub sites will allow for multipoint meetings via a Cisco Telepresence Multipoint Switch. To make the network changes possible, we needed a way to drive other costs down to invest in more bandwidth and WAN acceleration. This was done with voice changes. The two new technologies we will be using are IP Conferencing and SIP Trunking. IP Conferencing leverages VoIP from your offices to reach the conferencing provider. This takes our per minutes costs from maybe 12 cents per minute to 2 cents per minute (I'm changing the numbers for confidentiality). Being a large user of conferencing calling, our company can reap substantial savings by moving to IP Conferencing. Not all calls to the conferencing bridge will use IP since home users and customer will still use the PSTN (at say 12 cents per minute), but any calls from our offices will use IP. Thus, we expect 40% of conferencing calling minutes to be subject to the IP rate. Finally, we are moving some US field offices to SIP Trunking. These offices will now use IP not just for conferencing calling, but also for PSTN calls. This allows us to remove local PRIs at each field site and rely on a bucket of calls via the MPLS provider. Since it's a pooling and bursting model, there are economies of scale from sharing the pool. This cuts our telecom costs which we can use to purchase more bandwidth and WAN acceleration. These are the five major changes included in our WAN Transformation project. This will be a major initiative and take nearly a year. Next week I'll detail the financial ROI that made the project possible.

More >From the Field blog entries:

WAN Transformation is a Go!

Cisco Unified Computing System is 75% Networking - Who Knew?

Congress Needs to Step In for Net Neutrality...Really? Seriously?

It's Time to Start Loving Tunnels

Special Cisco Live Contest - Hottest Booth Girl

Best Sign at Cisco Live

  Go to Cisco Subnet for more Cisco news, blogs, discussion forums, security alerts, book giveaways, and more.

Copyright © 2010 IDG Communications, Inc.

The 10 most powerful companies in enterprise networking 2022