Whether your organization has two sites or two hundred sites, figuring out how to manage the data communications between those sites can be a real challenge. With a vast array of options, it's often hard to determine what's best for your organization. Here are three key evaluation criteria you can use to make the right decisions for your connectivity.
In most cases the reliability of the network is a critical component to how useful it is. Even in organizations where the ability to communicate between sites isn't core to the business because there are no central databases to be accessed and no real-time need to share information, loss of connectivity is still very disruptive.
Reliability is measured by the frequency, the duration, and the completeness of outages.
- Frequency — This can vary. For instance, in the US we're used to minor power outages through the summer as thunderstorms roll through. We may expect a frequency of a power outage in the summer months to be nearly once per month — more if you live in a rural area.
- Duration — If a power outage is only long enough to require a reboot, it may not even be recognized.
- Completeness — It may be that the system doesn't completely fail, rather it just experiences significant slow downs or congestion. Depending upon your needs, occasional congestion or slow downs may not be a significant issue for you.
When considering a data communications transport you should ask what kind of reliability ratings they have and what service level agreements (SLAs) that they have. SLAs outline the commitment to provide a certain reliability of service, the responsiveness to service problems, and the compensation your organization should expect if they fail to meet their commitment.
The next concern with a data communications strategy is the scalability. In other words, when you need more bandwidth, how easy is it to acquire? Is it as simple as placing an order, or are there physical or contractual limitations that will make getting more bandwidth difficult or time consuming? Scalability can be achieved in three different ways: Expand, Replicate, or Replace.
Expanding is simply increasing the bandwidth that is being paid for. If you're using a DSL line at 1.4 Mbps download and 256 Kbps upload and you have the ability to upgrade to a 3Mbps download and 768 Kbps upload, that's expanding the capacity. In this case you're not adding new hardware, connecting new lines, or doing anything other than asking the provider to change some software settings. The net effect is that expanding connectivity is very easy and very quick. It's what we strive for when creating a scalable data communications network.
In the replicate model of scalability you add another of the same type of communications to the mix. This is most often used with T1s (or DS-1s, if you prefer). You buy a router with ports for two T1 lines but you only plug in one T1 line to start. When you need the additional bandwidth, you add the line and connect it to the existing hardware. This option is good in that you get to continue to use the same hardware and infrastructure; however, it generally has a lead time associated with it and therefore it's not as good an option as expanding an existing line.
The final scalability option is to replace. If there is no scalability option, the hardware and connection lines must be replaced for an entirely different data communications transport. Obviously this is least desirable since it involves both potentially long lead times and entails the high cost of replacement and reconfiguration of hardware. Complicating this option are potential contractual limitations which may require early termination fees for the existing service.
The final consideration for a data communications strategy is complexity. Complexity is a necessary evil for constraining overall costs. However, there is the opportunity to introduce too much complexity into a data communications scenario, which results in your management costs going up — and your overall reliability going down. Every additional component must be evaluated based on its management cost and its impact to reliability.
I was working with a consumer service organization with locations in a tri-state area. They had managed over the course of the years to connect each of their sites with a different technology. Most of the locations were on DSL and Cable, however, some were dial up and others were on frame relay connections. You name it, they had it. Complicating matters was that each location had a different router. He result was an extremely complex network. They didn't know who to call for problems and they couldn't build experience with troubleshooting their own issues. This organization ultimately decided to invest in standardized routers and began a process to reduce the number of data communications providers they worked with. Both efforts helped reduce the complexity of their network and ultimately create an environment with higher reliability.