Bringing wired and wireless LANs together is an important trend - but not without controversy
Perhaps of necessity, wireless LANs have mostly followed what we today call the overlay model – that is, the WLAN has been planned, configured, installed, managed, and used as a service apart from the underlying wired LAN that enables it. There has always been a well-established and ongoing supportive role played by the wired LAN, in the form of the interconnection of access points, controllers, location and application appliances, and all of the boxes and pieces required to get and keep a WLAN on the air. But we still fundamentally have two separate networks, which has always made little sense to me.
Sure, with wireless LANs constantly evolving (and only a hint of a slowdown here in the air today), it in fact made sense to many network managers to keep the two nets separate, event though users are migrating to the wireless LAN for default and even primary access, and even though two networks cost more (often a lot more) to operate than one. And it was that cost element that got me to thinking quite some time ago that a unified approach minimizing the boundaries between wire and wireless to the greatest degree possible would be the right way to go. The key thought here is that capital costs for equipment (capital expense, or CapEx) continue to decline because they are on an engineering/manufacturing cost curve which, driven by the familiar faster/better/cheaper phenomenon typical of high tech, continually yields more for less. But with that come both complexity, a function of feature creep and other attempts at differentiation as well as an effort to provide honest value, as well as increased usage because networks of any form only grow over time in terms of both scale and mission. And all of these add up to higher operating expense (OpEx), which usually increases both due to volume of demand and the fact that running any network is essentially labor-intensive, and people only cost more over time. So we must improve productivity (output per worker) if we are to control OpEx, because demand is only going to increase. Thus it makes sense to think in terms of operating one network, not two.
But what to unify? At the very least, we need to replace two management consoles with one. We need to unify all management, security, and related databases. We need to provision user services uniformly irrespective of access. We need to streamline help desk, training, and other support. But we might also want to unify other elements of a given solution. I had a conversation with the wireless team at 3Com (now part of HP) late last year where they described their implementation of unified networking, one of the most complete on the market today. They even go so far as to implement the same OS on all of their products, ensuring uniformity, commonality of services, and, they argue, efficiency and ultimate performance as well.
The debate rages on. So, two things: the 3Com meeting motivated our latest White Paper on Unified Networking, discussing this opportunity in much more detail, which you can find here. And second, if you’re at next week’s INTEROP conference in Las Vegas (I am once again privileged to Chair the Mobile Business Conference here), be sure to check out my session on this very topic, featuring key speakers from 3Com, Aruba, Cisco, and Motorola (thus ensuring, um, a significant diversity of opinion). But you’ll be hearing a lot about unified networking regardless – this is major trend, one that finally closes the gap between wire and wireless with profound technical and cost benefits.
