Hello George,
Two of the new vistas that have been opened up as a result of photonic transparency are those which surround the long distance deployments of Gigabit Ethernet (GbE) and Fibre Channel (FC). 10 Gbe is also in the works now, and promises to be an alternative to OC-192s much sooner than was previously thought.
I think it's safe to say that we can prepare ourselves to see some new religious wars erupt here among competing in-house factions within enterprises (data center vs. telecomms, cabling facilities vs. LAN/WAN, design&construction vs everyone, etc.), alongside the older wars that have been waging between ATM and IP for the past eight years. This time the new battle lines will be drawn along those of SONET vs wavelength, and FC vs Gbe. And to a lesser extent, the older dispute between ATM and IP will continue to wage, as well.
[I once wrote a piece here in SI about a year ago, I believe, which I cant find now... that focused on the secondary effects of removing distance constraints as a result of dwdm technologies. I think it was over in the MRVC thread, although we covered this topic in LMT and elsewhere, as well. If anyone has these tomes bookmarked, please post te links. I do wish SI's search features would improve. In any event, a major point which I brought up in those earlier writings was the impact that fiber will have on in house practices and habits in the workplace, especially where the EIA/TIA distance parameters for LANs and structured cabling systems are concerned. The latter have been responsible for limiting the distances of end points to within 90 meters of fully conditioned closets and data centers, in the past. The implications on architectural and construction costs (and time for readiness) of this one attribute alone, are astounding. These issues, and the vesting that they represent to those who have capitalized on them in the past, are also seriously threatened now by the transparency allowed by dwdm/fiber.]
Of course, IP can still ride atop any of the lower layer protocols we've mentioned, so IP needn't be the dominant issue of concern (because it's never really fully excluded, i.e., it should not be considered mutually exclusive to any of the above protocols. What we're discussing here is a matter which concerns the physical and data link layers, or Layers 1 and 2, whereas TCP/IP sits at Layers ~3 and ~4.
When distances are removed, one can plot the topology of their network as though it was a LAN. I think that GbE offers an excellent solution for "flattening" said LAN, and can be used as the bridging protocol between remote routers in support of IP. I also think there's a place for GbE for storage area networks (SANs) and network-attached storage (NASs) over much greater distances than previously possible, although, here, FC is quite formidable, still, with no signs that its popularity is going anywhere but up, as well. To wit, note the demand for ANCR and others, recently.
I'd like to spend more time exploring both of these alternatives here, and would welcome more comments/questions and rebuttal from yourself and others.
Regards, Frank Coluccio |
