* A little history lesson
Ten years ago, we were in the midst of the controversy over whether frame-based technologies, such as frame relay and IP or cell-based technologies (ATM in particular), would form the foundation of the next-generation network. Let’s take a look at how all of this has played out.
Fundamentally, there’s very little difference between frames and cells. In fact, cells are just frames that happen to be the same size. Nevertheless, countless dollars and careers were wagered on frames or cells being “better.”
To understand fully the background of this controversy, you have to remember a couple of factors that were extremely important at the time. First, transmission speeds were relatively low compared to today. Now, we have megabit services to the home via cable modem and DSL; then, we were lucky to have 56K bit/sec digital services to relatively large corporate branch offices. Also, extrapolating Moore’s Law in reverse, processing power per dollar was miniscule compared to today.
The consequence of these factors was that there was a perceived need for a technology that could provide low latency and fast switching. Thus, the interest in ATM grew because of the ease of processing fixed-length cells. At the time, processing variable-length cells was just a little too tough for most high-speed switches. Also, by keeping the cell size small (53 bytes for ATM), the freeze-out on a per-packet basis was low and predictable.
It’s interesting to note, though, that the low delay wasn’t really caused by the fixed-length nature of ATM. Rather, it was the fact that the maximum packet size – be it a frame or a cell – was small. Equivalent or better delay characteristics can be achieved with frame-based technology with a small maximum packet size.
Next time, we’ll continue this review with an assessment of where we stand today vis-à-vis frames and cells.
