During my first drive into Boston many years ago, it didn’t take me long to get hopelessly lost. I later learned that the labyrinthine road system was not in fact an elaborate trap for hapless outsiders, but owed its legacy to a time when horses were the primary vehicles it conveyed.

Looking back in time to when the internet was built, it’s hardly surprising that the mix of data funneling through its pipes has changed dramatically. Most striking has been the growth of video traffic, which is expected to account for 90% of consumer traffic by 2012.

While it’s a stretch to compare delivering video over the internet to an 18-wheeler truck trundling over cobblestone roads, the fact of the matter is that online publishers are still getting stomach ulcers over how to distribute their high quality video content cheaply & reliably. And it’s abundantly obvious to viewers that there’s a huge difference between watching the same show on their broadband-connected PC versus their TV. Some pundits go as far as asserting that this lousy fidelity is the root cause behind the scarcity of online video profits.

Now, there’s no shortage of startups proffering solutions for video distribution. Move Networks, Velocix, Oversi, BitGravity, RawFlow, Digital Fountain & Vusion are just a small sample of what’s out there, and while they have different approaches, they generally operate (wisely) within the confines that existing internet infrastructure imposes.

I’m curious though, given the sheer magnitude of the problem, if it makes sense to follow a more audacious path. Video traffic possesses its own unique characteristics and looks set to take over the internet – should we really be using the same stack developed for a different era to convey it?

Perhaps there is a whole new class of network device designed from the ground up for video. Maybe it’s about making changes deep down the stack between the physical & IP layers. Because it really doesn’t matter how many lanes you add to a road, or how you optimize the placement & timing of traffic lights, or how many vehicles you outfit with GPS systems – traffic will still crawl if the roads aren't properly paved in the first place.