Policing the new network traffic

Adam Davison looks at the future of the network in light of the rise of commercial off-the-shelf hardware

I don't know how those chipset boffins do it, keeping up with Moore's Law. It's as if a blood oath was sworn all those years ago, forcing them to create faster, smaller and cheaper IT processing engines just so a quiet, unassuming academic never has to be proven wrong.

PCs, notebooks and tablets have become so replaceable. Paying a (modest) premium is no real obstacle to users getting their hands on next generation, faster hardware. Meanwhile, the economics of Moore's Law mean enterprises still benefit, and the channel always has new capabilities to sell.

In networks, however, costs and disruption make enterprises rather less eager to replace computational intelligence annually. The result is 3-7 year life cycles that enable enterprises to get more from their kit even while performance levels diminish.

In carrier networks, where major equipment brands have accumulated hundreds of billions of dollars in revenue for their long term, proprietary solutions, there is a massive change underway to what is called commercial off-the-shelf (COTS) hardware.

With so much traffic now migrated from legacy networks to IP, the myriad switching, routing, signalling, shaping and other network siloes are merging into homogenised hardware environments. This means that network operators can ‘ride the wave' of Moore's Law, routinely upgrading wholesale massive swathes of standard hardware - rather than having to manage each hardware supplier relationship, product evolution and procurement cycle separately.

This saves these businesses huge sums, and it has encouraged the equipment manufacturers to focus more on their intellectual property smarts - after all, everything is an application now - and less on developing specialised hardware footprints.

In enterprise networks, of course, it's the virtualisation trend that is changing the face of things to come. Will we see the complete commoditisation of hardware, a transformation into a COTS-like state? And where would this leave the typical reseller, used to pitching, providing and supporting specialised hardware footprints as an extension of the familiar technology brands?

Having worked in WAN acceleration and performance optimisation for the past decade, I'm aware of all the racing car analogies used down the years to entice resellers to sell the offering and help them communicate the vendor's message to customers.
WAN optimisation started out as a network add-on, designed to improve the performance of what you already had. Today, most resellers already have a WAN optimisation arrow in their quiver.

But as we've seen, the network is changing. At the risk of sounding somewhat over the top, where exactly the hell is the network these days? The world has moved on to a new generation of IT applications and working habits that have affected the networking question.

Network use is now more fractured, unpredictable and user-controlled, instead of being determined or anticipated by the IT department. And enterprises understandably struggle to optimise what they cannot see or control.

Analysts are saying WAN optimisation is a notable growth area, so I guess that's something for resellers to think about. But network optimisation sales don't arise if you don't know where the network is or where it's going, any more than enterprise CIOs can gain network optimisation improvements without knowledge of the same crucial facts.

Adam Davison is vice president of sales EMEA at Exinda