Bandwidth is like the Long Island Expressway. Add lanes, they fill up. Thank God for Gig-E.

JIM BARTHOLD By a show of hands, how many of you have heard some variation of this tune? “Cable’s broadband pipe is limitless. We don’t need to do anything ever again and we’ll still have bandwidth left over for any new service that comes along.” We’re not hearing that so much anymore, but, if you squeeze a cable exec, you’d probably hear that the networks are “adequate for the foreseeable future.” Cable’s confidence in its network — as publicly expressed — is usually accompanied by a wet raspberry aimed at telephone or satellite providers who “just don’t have enough bandwidth to really compete.” Throughout my speckled career, it’s been hammered into me that the hybrid fiber/coax (HFC) architecture is an immense conduit that can swallow anything an operator throws into it. Cable modems? No problemo — until of course the contention models get bollixed when the kids come home from school. HDTV? Bring it on — well, not every channel, and please let’s not even think about must-carry. Telephony? Hell, the phone companies can do that with twisted pair; cable has fiber and coax. Putting voice services on HFC is like dropping a Word document onto a 60-gigabit hard drive. Considerably cruder comments answer any who dare question cable’s ability to efficiently transport voice, video and data on networks that were — and here’s a semantic pothole — built for “video entertainment.” Question the flexibility of those networks, and some MSO engineering executives, assuming she/he still has a job in today’s compressed industry, will throw you a look patented by disappointed mothers, then say something like, “You poor simpleton. You’ve been deluded by those evil telephone people. Cable rules. Our networks are built for today, tomorrow and the 22nd century.” Today’s requirement is video-on-demand. Making VOD seamless for a viewer with a beer in one hand and an itchy remote trigger finger on the other would seem to require more engineering creativity than you’re likely to find in your latest SCTE handbook. Surprisingly, so far, cable engineers have been up to the task. They’ve evolved Gigabit Ethernet (Gig-E) transport to the point where they can almost — but not quite — packetize the video and dramatically improve its transport, particularly in the upstream. Gig-E isn’t quite packetized video because everyone worries that streaming video would make obsolete millions of incumbent digital set-tops that have been deployed, paid for and are now reaping dividends. On the other hand, more efficient video transport — bandwidth management — won’t happen by degrading picture quality with more advanced forms of statistical multiplexing over existing compression schemes. That was tried and found lacking during cable’s early digital rollouts. And it’s unlikely that even the most penurious operator wants to revisit that space. While some new compression schemes are being touted by eager vendors, they face the same hurdles as streaming video: an incumbent technological base. So the industry is turning to Gig-E, a sort of IP hybrid that an engineer would feel compelled to diagram on a whiteboard, or better yet, throw into a PowerPoint presentation to properly explain. Gig-E translates standard MPEG video into a packetized format. By doing so, it improves on the former standard of DVB-ASI (Digital Video Broadcast-Asynchronous Serial Interface). Although labeled gigabit, Gig-E generally delivers about 900 megabits per second of payload, which is substantially more than the 270 Mbps available via DVB-ASI. An operator can replace an ASI output in a VOD server with Gig-E, and voilà! — the transport looks just like IP. With Gig-E, MSOs can dedicate fewer servers to VOD customers and work out new contention paradigms to handle an increased signal load in previously under-used upstream bands where a plethora of new on-demand commands are now clamoring for space. It’s a relatively ingenious way to use new IP technology to handle old MPEG video. While it doesn’t solve the ongoing problem of an increased traffic load from data and voice services — to say nothing of eventually targeting the lucrative commercial space — it does pave the way for operators like Comcast, which has dedicated 2003 to VOD, to offer up anything-on-demand services. In short, the engineers have delivered what the marketers want. There is still no evidence that this iteration of VOD — anything-on-demand — will succeed any better than it has in the past. But, by using Gigabit Ethernet to tweak their networks, engineers have found the space they’ve always bragged about and have passed the first of what promises to be many hurdles as new services evolve to attract new revenue from old subscribers. Jim Barthold was technology editor of Cable World from 1996 to 2000 and more recently a senior editor with Telephony magazine, where he followed broadband.