Provisioning Part II: @Home-pty Dumpty Sat On a Wall …
by Leslie Ellis // January 21 2002
Ask any cable operator what they’ve been up to lately, and unless you’re talking to Time Warner Cable, the words “Excite@Home” will invariably and wearily follow. The high-speed data service goes dark for good in about five weeks, which means that affiliated MSOs are up to their eyeballs in details and sequence about how to manage customer transitions as painlessly as possible.
The cutovers that must occur before E@H goes dark are complicated – an understatement – and require considerable planning, human resources and cross-departmental cooperation. In a larger sense, the movement of subscribers from E@H to those of cable providers is the biggest provisioning (translated in the Dec. 17 edition) effort ever undertaken.
To review: The term “provisioning” generally describes how to activate an account for service, as automatically as possible. What’s different in the case of the E@H mess is the customers themselves: In most cases, they’re not new. But because E@H holds most of the electronic information that represents them to back office systems, like billing, the customers have to be treated, behind the scenes, as though they are new. They must be re-provisioned, in a sense.
Cutting customers off the spine of E@H, and grafting them, vertebrae by vertebrae, onto a new, wholly cable spine, is just as surgical, bloody and delicate as it sounds. Just ask AT&T Broadband, the first of the MSOs darkened by Excite@Home. Employees involved in the round-the-clock effort, which spanned the first six days of December 2001, liken the experience to working in an air traffic control center, or to launching a space shuttle. A collective adrenalin rush, similar to what relief workers feel during any major crisis, caused participants to work 36-hour+ shifts, then rue ducking away for a nap. @Humpty Dumpty had fallen off the wall, and needed to be put back together again.
Thinking through a provisioning project this vast involves at least five categories: 1) The backbone, 2) the regional network, 3) the databases, 4) the coordination with the field, and 5) the internal and external lines of communication.
Provisioning the backbone involves the almost-always dreary task of ordering high-speed circuits – DS-3s, in data parlance, which haul data between the backbone and headends at a zippy 44 Mbps. Ordering circuits is dreary in the same way getting tickets to The Producers is dreary: Insist that you need them tomorrow, and you get laughed at. Getting circuits at all requires dogged persistence. (AT&T Broadband had a somewhat easier time with this part, since daddy owns and operates so many of them.)
In most cases, the backbone lines don’t drop right into cable headends. For reasons of economics and logistics, a regional network almost always exists to shuttle traffic from individual headends, to the place where the high-speed backbone (the long distance lines) picks up and drops off. That part was always Excite@Home’s job, which means it needs to be redesigned before Feb. 28.
Then there are the databases. Perhaps the surest way to proceed is to secure an up-to-the-minute database replicate of Excite@Home’s “back office system,” or “BOS.” From it, customer information (name, address) and cable modem information (“MAC,” or “Media Access Control” address) can be parsed into the billing system. Similarly, e-mail and login information can be poured into a new e-mail server. That way, customers don’t have to create a new login sequence, but instead only have to remember the new domain name (that’s the part after the “@” sign).
Constant coordination with field employees is also a critical part of cutover and re-provisioning. It’s sort of hard to physically remove the cables from Excite@Home’s gear, plug them into the replacement gear, and make sure everything works, without a few smart people onsite. Especially considering that every market cutover almost always involves visiting multiple headends.
Massive methods for communication, both internally and “customer-facing,” are also a big part of re-provisioning former E@H subscribers. Bridge lines, e-mail updates, toll-free lines, instant messages – all are useful for the cutover itself. E-mails, snail mail reminders and phone calls go far in easing subscriber anxieties about the switch.
One of the many useful processes AT&T Broadband created, for example, was a sort of round-robin communication pool among markets. If Dallas cut over on a Monday, and San Francisco was next, employees in Dallas and San Francisco got on the phone to talk about what did and didn’t work. San Francisco then carried that knowledge forward to the next market, and so on.
In the end, it is possible to put this @Humpty Dumpty together again – but not without triage-type provisioning, planning and coordination.
This column originally appeared in the Broadband Week section of Multichannel News.
‘Grooming’ Means Customizing Digital Video Lineups
by Leslie Ellis // January 07 2002
A letter showed up in the e-mailbox a few weeks ago, seeking translation on a term that snuck into digital video lingo about five years ago: “Grooming.” It resurged just before the Western Show, when a supplier company, BigBand Networks, said it would outfit “grooming” gear to two Time Warner Cable systems.
“Grooming,” in the digital video sense, has nothing to do with standing in front of mirror with a comb, a toothbrush and a bar of soap. Instead, grooming has to do with customizing digital video channel lineups. Tactically, grooming usually happens after a clump of digital video channels enter the headend, but before the customized channel clump moves over the plant to subscribing homes.
Technically, “grooming” is also known as “add-drop remultiplexing.” It is performed by a piece of headend equipment, made by a (very) short list of vendors. In a signal flow sense, grooming gear generally sits after the satellite receiver and before the encryption and modulation gear.
Most people who use grooming equipment nowadays use it to add or remove programs from AT&T Broadband’s “Headend In the Sky (HITS)” feeds. Maybe a HITS affiliate wants some of the channels in the feed, but not others. (The obvious example is the removal of risque programming in deeply religious areas.) The process of breaking the feed open, plucking out the desired channels, and repackaging it all together, is what necessitates the “grooming” gear.
Most businesses have some form of “grooming.” They just don’t call it that. Consider a gourmet food store, in Vermont, that does a brisk business in gift baskets. Perhaps one of its bigger basket suppliers always seems to include piña colada mix, when a tin of hot cocoa would probably have more appeal (given the season.) When the proprietor removes the colada mix and substitutes the cocoa, the basket has thus been repackaged – “groomed” – for sale.
The mother of “grooming” gear is the statistical multiplexer. Statistical multiplexers and talk about “grooming” always seem to go together, mostly because one (statistical multiplexing) begat the other (grooming).
Technologists like statistical multiplexers, or “statmuxers,” as they’re sometimes called, because they extend to the passageway what MPEG-2 video compression aimed to do in the first place: Eliminate wasteful, redundant bits. Nip and tuck. Be efficient.
Sometimes, redundant bits in a video image are hard to find. Smoky scenes, for example, as well as high-action sports scenes and explosions, are complicated to compress, and require lots of bits for a given amount of time. At the other extreme, a newscaster’s head and shoulders don’t change much, frame to frame, yielding lots of duplication that could be removed.
Those are two compression extremes. In between is a suitable-enough distribution of bit rates and complexities to apply statistical analysis (the “stat” in “statmux”), which yields averages and means and all sorts of intricate, useful ways to create a package of squeezed, digital channels (the “mux” in “statmux”) more efficiently.
Think of the need for statistical multiplexing in terms of any congested highway, at rush hour. If that highway is a 6 MHz channel, the statistical multiplexer is what assures that the fast lane is going uniformly fast. Ditto for the slow lane.
In video terms, that statmuxer is finding the most efficient way for digital video channels of differing bit rates to ride efficiently within the same 6 MHz channel. The amalgam of the compressed programs is the “multiplex,” or “mux,” as engineers say. Grooming is the method and equipment used to customize that mux, before it goes to set-top boxes.
Statmuxing and grooming are in the lexicon right now because digital video is reality. It won’t be long, in fact, before other digitized material – high-speed Internet, telephony – will be candidates for statmuxing and grooming, too. That’s certainly the intent of the supplier companies active in the field.
Cable technologists have long mulled the day when they don’t have to deal with anything analog, and can then treat the entire 750 MHz channel-load as one, gigantic channel – instead of a series of 6 MHz channels – where content is muxed in and out at peak efficiency. In a phone call where there’s silence – “Hang on, somebody’s at the door” – that bandwidth gets reclaimed for use elsewhere, until the conversation resumes.
That sort of thing will require even more sophisticated “grooming” devices. Right now, it’s mostly just MPEG-2 digital video that gets groomed. Over time, watch for it to expand to Internet Protocol (IP) and other digital signal types.
This column originally appeared in the Broadband Week section of Multichannel News.