56K: Breaking the speed limit

This article originally appeared on a CMP web site in 1997.

In the fall of 1962, an event took place that would eventually reshape the world of communications. The first digital telephone call was made from the AT&T central office in Skokie, Illinois to Chicago, and it marked the first practical application of digital communications.

Until this call, the human voice had been carried over phone lines using entirely analog transmission. But instead of electrical currents, this new technology converted the voice to bits, or digital pulses. These digital pulses were suddenly talking a language that computers understood, which meant that computers could now "talk" over telephone lines, as well. A few years after the first digital phone call was placed, a little thing known as AARPANET was born, and, well, you know the rest.

The white paper that launched 1,000 chips
In the fall of 1996, another leap forward in digital communications took place. Rockwell Semiconductors released a white paper detailing a technology that allowed modems to connect to the Internet at speeds approaching 56Kbps. Shortly after this, U.S. Robotics announced they would start shipping modems using their version of 56K technology, X2.

The race toward a standard was on.

But wait. Surely modem speeds of 56Kbps weren't possible? Only a few short months before the Rockwell white paper, experts had suggested we'd reached the end of the line for conventional analog modems. V.34 33.6Kbps modems were the fastest we could hope for, prevailing wisdom said; better get ISDN or cable or one of the DSL technologies if you want to surf the Internet at speeds higher than that.

We now know 56Kbps is possible, and here's why. When data is converted from analog to digital (or vice versa), quantization noise is added. If there's too much noise, the data can lose integrity. That's why the V.34 standard was set at that level; anything faster was thought to add too much noise.

However, because data traveling from the Internet to your home is already digital (just about all the major Internet service providers use digitally terminated lines to bring data off the public-switched telephone network (PSTN) and onto the Internet), it only has to make one conversion from digital to analog when it hits your twisted pair copper wire phone line.

56K modems are basically a hybrid of analog and digital technolgies. They utilize what's known as pulse-code modulation (PCM) technology, which has been in use by ISDN local loops for a while now, to receive more data with less quantization noise. Less noise means a faster download time, which means a much faster Internet experience. And, while it is possible to use this analog/digital hybrid model to push modem technology even further, up to 64K, such speeds are at least a year away, if not more.

Close but no cigar

So basically, 56K technology operates under two primary assumptions. One, it assumes a digitally terminated connection -- the Internet to you. And two, it assumes most Internet surfers are doing just that--surfing. In other words, you're following the asymmetrical pattern of the majority of Internet use and receiving data (in the form of graphics or files), not sending it.

Therefore, if you don't upload a lot of files to the Internet and you don't need to connect to other analog modems on a regular basis, 56K is your answer. However, Paul Kraska, product marketing manger for Multi-Tech, a leading modem manufacturer based in Minnesota, cautions that users won't be able to realize 56Kbps speeds right away. "It appears that this technology is going to work well in the high forties," he explains. "When you think about it, that's a wonderful advancement in modem technology. If you operate one of these things, it screams. Compared to a V.34 modem, 45-48Kbps is phenomenal. However, we need to be careful not to set the user's expectations too high."

So why can't you achieve the full 56Kbps? Well, there are several reasons. For one thing, your phone lines may be old and noisy. Plus, if you don't live in an urban area, your lines might have to travel a long way to reach a central office (CO) connection, which might mean they travel through what are known as suburban subscriber link concentrators (SLCs). These SLCs add noise, forcing your 56K modem to "fall back" to slower speeds. And if you're connecting through a PBX system, found in most hotels and many businesses (if you have to dial a 9 to get an outside line, you're using PBX), you won't get 56K, either.

However, as Lisa Pelgrim, a senior analyst for Dataquest Inc., in San Jose, Calif., points out: "Not achieving the top speed of the modem is not a new occurrence When V.34 modems became available, this same argument was used and V.34 still became the most popular type of modem."

Not so fast...
Technical speed bumps notwithstanding, there exists another, more significant problem facing a potential 56K modem customer -- there's more than one 56K protocol available. Last fall, after Rockwell released its white paper detailing its new 56Kbps technology, U.S. Robotics wasn't far behind in announcing its Courier V.Everything X2 and Sportster 56K Faxmodem modems that utilized USR's proprietary X2 technology. Lucent Technologies had developed a third 56K technology, but they quickly aligned with Rockwell and folded their technology into what is now known as K56flex.

What are the differences between the competing protocols? Well, unless you eat modem algorithms for breakfast, the differences are fairly subtle. The biggest difference boils down to access: Which Internet service provider supports which flavor of 56K?

"How confusing is this for the consumer? Very, very confusing." That's according to Michael Henderson, marketing director for Rockwell's Network Access Division. "Our customers really do not undertsand the difference between X2 and K56lfex. I recently worked in a retail store for a couple of days in the modem aisle. We wanted to get a feel for what kind of things were happening in the retail channel. We found out that there's an enormous amount of confusion. People need to know that they have to pick a technology that matches their ISP."

Dataquest expects most ISPs will align with the technology that is compatible with their infrastructure, a strategy that a recent Boardwatch article likened to "dance with the one that brung ya." And since the digital remote access infrastructure already in place at most ISPs across the country uses Rockwell chip sets, it's not rocket science to figure out that these providers will support K56flex. The top five ISPs -- AOL, CompuServe, MSN, Prodigy and EarthLink -- have all pledged varying degrees of support for both technolgies.

The number of ISPs supporting K56flex is growing every day, so the best way to find out if your ISP is listed is to check our ISP database. That way, you'll know whether your ISP supports K56flex before you buy the wrong type of modem. You can also find a list of modem manufacturers here that use K56flex technolgy in their products.

Can't we all just get along?
The best way to end this confusion, of course, is to implement a standard. The International Telecommunication Union, a United Nations agency that coordinates telecommunications issues, has created an Expert, or Rapporteur, group that has been charged with reviewing vendor submissions and developing a recommendation on PCM modem technolgy. The ITU is expected to rely heavily on input from the Telecommunications Industry Association (TIA) TR30 committee, a U.S.-based standards group that has already started work on a US Interim Standard for PCM technology.

"The current plan is to have all the technical details agreed and documented by September of this year," says ITU vice-chairman, John Magill, VP with Lucent Technologies. "A draft ITU-T recommendation (will be) proposed at a meeting in mid-September, which will start the ITU-T approval process. The final approval will then take place at the meeting of ITU-T Study Group 16 in Geneva in January, 1998. This schedule was set in March this year, and we are still on schedule."

Of course, January 1998 might seem a long way off when 56K modems are in stores now. However, Magill is quick to defend the ITU's deliberate pace. "The ITU is always under pressure to respond quickly to perceived market needs, and indeed has changed its procedures to facilitate this," he explains. "(But) no matter what happens in the standards arena, companies will always bring new technologies to market early in an effort to gain market share and sales. Standards bodies cannot and should not interfere with this enthusiasm. In general, the standards process benefits from the experiences of those early in the market."

Industry experts are quick to point out that such "pre-standard" sales are nothing new -- the exact same thing happened with the V.34 standard--28.8 modems were on store shelves long before the ITU stamp of approval was on the box. Dataquest's Pelgrim explains: "Because most 56K modems...will be upgraded to the eventual standard, modem vendors have to get in the market now in order to make the sale."

However, once a standard is set, it won't take the modem manufacturers long to get up to speed. "Already ISPs are installing equipment capable of very simple and quick upgrades once the standard emerges," Magill points out. "The same applies at the client end -- many products on the market today are capable of a software upgrade. We could see literally millions of users within months of the standard emerging."

"Ultimately, we really need a standard," Henderson admits. "The ISPs aren't happy about this situation, so the quicker we can get a standard the better. Of course, what may happen here is that in the short term, the ISPs become a channel for the modems. That way, the customer gains confidence that the modem they're buying works with their ISP."

Back to top


© Lisa M. Moore
May not be reproduced in whole or part without my written permission.