There are really two approaches to optimizing bandwidth economics with little or no up-front investment. Both work and they are not mutually exclusive.
The first is strictly competitive. Telecommunication line contracts are based on monthly lease commitments for 1, 2 or 3 years. During the course of the contract term, the competitive landscape can change dramatically. What cost $1,000 a few years ago might be had for $500 or less today. That scales all the way from T1 voice and data lines up to optical carrier services. Can you save 25 to 50% or more by simply switching vendors? It's done everyday. The most dramatic savings are seen in areas that have only recently seen competitive carriers move in or in situations where contracts have been automatically renewed over and over with little or no price negotiation.
The second approach is to switch technologies. That may seem like a daunting prospect. After all, you've spend a lot of time stabilizing your network. The last thing you need is to throw some technical incompatibility into the mix.
But transport technologies are more compatible than you might think. Most network traffic today is based on packets. Ethernet packets to be precise. Older metropolitan and long haul networks were developed around telephone technology and based on digitized voice channels. There's no reason that a T1 or OC3 connection can't transport packets instead of voice streams. You've been able to get T1 and SONET data lines for years. To make this work, there is circuitry that performs a protocol conversion so that packet data can be sent as collections of bits the right size for the line technology and then reassembled at the other end.
All of this conversion from one network protocol to another happens transparently and so fast that you are unaware of it as a user. You know that you have a dedicated connection to the Internet or a private data pipe from point to point. How it all works is unimportant. Or is it?
A network designed from the ground up to be compatible with the core protocol that you are running is bound to have less overhead and be more efficient. That's the idea behind IP core networks and Carrier Ethernet. Competitive carriers without a legacy investment in telephony networks have been building out all-new IP networks serving regional areas and even the entire country. This is where you'll find Ethernet services for metro and long haul applications.
But does switching to Ethernet from TDM technologies like T1, DS3 and OCx really save anything. In many cases, yes. No, you won't be able to get a 10 Mbps Ethernet service for the same price as a traditional T1 line, but the cost per Mbps might be 1/3. It gets even more dramatic when you move up to 100 Mbps Fast Ethernet and 1000 Mbps Gigabit Ethernet. Depending on your location, Fast Ethernet even get you twice the bandwidth for less money overall than a standard DS3 service.
A third approach is to add hardware based bandwidth optimization to squeeze the most out of the lines you have. This is more of a trade-off arrangement versus an outright cost savings. There's a cost for the WAN optimization appliances that needs to be offset by the costs avoided by not upgrading to faster connections. If you are in a situation where it would be really expensive to move up the next bandwidth increment due to limited service choices, bandwidth optimization can buy you the breathing room you need for your business activities.
