In a recent column, we mentioned that one of the major changes that we've witnessed over the past seven years is a dramatic drop in price per bit-per-second for WAN bandwidth. We also pointed out that it's becoming more and more difficult to cost-justify micromanaging bandwidth because it's easier to throw bandwidth at the problem than to try and solve it. This generated a couple of interesting responses from readers.
One reader pointed out that: "Your 'old school' vs. 'new school' contrast on approaches to resolving network performance issues overlooks two factors, namely latency and convergence.
"The LFN's (long fat networks) produced by high bandwidth are increasingly underutilized due to poor tuning on the end points and the reliable delivery algorithms in TCP. Converged voice/video/data networks require that different types of traffic be managed according to their unique characteristics and requirements. So there are a bunch of us 'middle schoolers' who are willing to throw cheap bandwidth at the problem but, at the same time, recognize the need for aggressive traffic management."
Of course, we agree that management is still needed and that it is of a different style.
Another reader shared his thoughts: "I found your recent piece on bandwidth very interesting, but to me the real foundation issue is monopoly control and the slow end of that long lasting disaster." And, with appropriate apologies to our reader in the U.K., we'll further share his thought that "hopefully we can go back toward the founding fathers view about the horrors of monopoly control, which in my opinion was the real reason we resisted the British back in 1776, not taxation."
The reader adds: "What we are seeing is what can happen when some real competition is allowed to happen with the resulting effect on technology." We do agree with this reader.
Maybe, now, after 20-plus years of telecoms divestiture, competition is really happening. However, we doubt that even the most idealistic dreamers in 1984 dreamt that we would be where we are today.