For more details, please see ourCookie Policy.


Mobility: not just a “Wireless Problem”

by dhunt on ‎01-18-2013 02:15 AM (2,966 Views)

Gartner suggests that 80% of wireless LAN’s will not be able to handle their traffic loads by 2015 in a recent article "Why Enterprise Wireless is Not Ready for the Mobile Explosion".  At the same time they also predict that over a billion smartphones and tablets will be sold in 2013, while IDC estimates that the number of remote workers dependent on wireless devices to connect will explode to 37 percent of the workforce by 2015.


Mobility is therefore a great opportunity; and a massive challenge. And critically, not just a “wireless” one.


‘Mobility’ is not simply connecting to a wireless LAN and being able to walk to the coffee machine and back without losing contact or access, regardless of wireless being the face of mobility to almost everyone.  To me, mobility is the idea of being productive by having all your resources available to you no matter where or when you decide to apply your creative juices.  And we all have plenty of wireless-enabled devices to play with.


So what is the problem, you may ask.  Where do wires come into it these days?


Traditionally wireless networks have been designed not as layer 2 networks but in a routed layer 3 design.  You have a routed distribution layer that forces wireless client traffic to be centrally switched using a centralized controller.  This creates many issues with doubling and tripling the number of times a single packet will cross the same link.  Sure, we could build larger controller interfaces, or sell our customers bigger switches to hide this problem.  Or we could step back and take a fresh look, and really help our customers.


I suggest a Distribution to Access Edge that does not use SpanningTree, for starters; which leverages the strength of locally switching packets while allowing for true layer2 and layer3 mobility.  At this point I really hope you are thinking HyperEdge.’  Or how about Software Designed Wireless LAN’s (a much more inventive approach we will discuss another time)?  Both are much better ways of bringing wired and wireless together to provide the right network infrastructure for mobility.


If we don’t address the wired architecture problem, 802.11ac and all those capable clients will make 80 percent of those wireless LAN’s deployed today unable to handle their traffic loads.  Then all the mobile devices in the world won’t be of much use!  Wired and wireless need to be reviewed as part of a holistic whole, because mobility isn’t just a wireless problem, it’s a campus problem and should be addressed as such.