Engineers like to disagree, so there likely will be a debate about the timing of the move to 100 Gigabit Ethernet. This PC World article says that the Electrical and Electronics Engineers (IEEE) is targeting 2010 for a standard. Several companies recently made news with what they called a successful test of the technology.
Currently, 10 GigE connections aren't being filled. This makes it seem strange that the industry is aiming at a higher-capacity standard that will push the upper end of the "pipe" by an order of magnitude.
It is possible to take the position that 100 GigE is premature since there is no demand for speeds that high, and hardware and software can't process data fast enough to use the capacity even if it was available. The other side could counter that the needs at the core of the Internet indeed are great enough -- or soon will be -- and that it's a good idea to get the process under way.
An interesting, though somewhat tangential, point was made in the InformationWeek article: Even if preparations for the faster speed are prudent now, starting a standards process at this comparatively early point may chill more innovative research that would occur in a less directed environment.
This all may not be rocket science, but it's pretty close. It's safe to say that the vast majority of us don't have to worry our pretty little heads one way or the other about the timing of the move to the next Ethernet speed trap. The usual factors -- vendor and service provider business case assumptions, competitive pressures, technical needs, political posturings, pure chance -- will determine when and if the new technology trumps what is in the field today.
A couple of things are important for non-engineers to understand, however. The most basic is that the networking capacity will continue to increase. It's also important for those with a rooting interest in the efficiency of these networks to understand that Ethernet is favored because it already dominates office and campus local area networks (LANs). The ability to use protocols with the same DNA in both LANs and wide area networks (WANs) tends to unify network operations and could even result in streamlined work forces.