![]() |
Comm
Corner
56 Kbps Modems Where do we stand? by John Woody |
| This is the third 56 Kbps analog modem article in this
series. We learned in previous articles that there
are two standards (protocols) being pushed by the modem manufacturers and
that the international standards committee has not as yet selected the
final standard. The name for this standard is V.PCM for Pulse Code Modulation
and will operate at 56 Kbps down-link and at V.34 standard (33.6 Kbps)
up-link.
We determined that there is a lot of money at stake in the sale of new modems to satisfy our thrust for speed in Cyberspace. We discovered that the modem manufacturers were fighting the standards' acceptance in the market place. We also learned that 56 Kbps may be an impossible rate to obtain on the current analog telephone lines for several reasons. Lastly, we learned that the 56 Kbps rate applies only to the down-link; the up-link operates at V.34 speed. What's a person to do? We all rushed out and purchased one of those new 56 Kbps modems anyway. We had little assurance other than the manufacturer's marketing hype that these modems would be software upgradable to the final standard once it is approved. In addition, we have had to make extra decisions as to which vendor to buy from.
ReviewTwo standards are being pushed in the market. On the one hand, 3 Com/U. S. Robotics, as the 800 pound gorilla, is pushing it's x2 technology as the standard for this class of modem. In the other camp, Rockwell, Lucent, Motorola, and the rest of the analog modem industry are pushing 56Flex as the standard. The 56Flex camp is pushing for an interim standard governed from the United States Telecommunications industry Association (TIA) to bridge the gap until the international group, International Telecommunications Union (ITU), makes its standard known.Both x2 and 56Flex are backward compatible with previous protocols, ie., both revert to V.34 or lower if the modems can not agree to operate at the higher speed. 56 Kbps technology seems to give higher speed performance even if you previously could not achieve the V.34bis 33.6 Kbps speeds. The technology is asymmetric, which means, essentially, that sending and receiving are at different speeds. The technology takes advantage of digital connections that most ISPs have or are installing in their connections to the telco public network. The download data actually are passed over the telephone line at 56 (53) Kbps using a new technique of encoding data for transmission over the phone line. Compression or modulation are not directly involved in this technology, even though compression can be used within the modem or by the sender. The standards do not mix well. That is what they say. More on this later. One has to be careful in that purchase, because the ISP must have equipment which will handle each protocol at its end.. The ISPs have had to set up separate receiver modem banks for each of the standards which can impact buying decisions. After we purchased that new super fast modem, we discovered that the manufacturer's marketing hype did not tell us that 56 Kbps is not legally obtainable in the U. S. at the current time. The current legal analog speed is 53.3 Kbps by FCC code. And the manufacturers did not tell us that only about 25 percent of the existing telco telephone lines are capable of even 53 Kbps. Most copper line installations have too much interference to provide even 53 Kbps. Moreover, the vendors did not tell us that the standard works through only one telco switch. If your line is routed through more than one switch, the protocol reverts back to V.34. The ISPs have had to install digital switches which are compatible with the particular connection they are offering, ie., x2 is usually connected to a US Robotics Total Control NetServer8 in order to provide the V.PCM download throughput. More little "gotchas" have come out of the closet as we have gotten into V.PCM, such as 16C550 UART-controlled serial ports when an external modem is installed. The internal 56 Kbps modems have these UARTs built-in. The modem chip set manufacturers are Rockwell, US Robotics, Lucent, and Texas Instruments. USR manufactures its own modem chip sets with x2. The others manufacture chip sets for the 56Flex. All 56Flex modems currently use the Rockwell chip set. Lucent supports 56Flex, but has its own symmetrical 56 Kbps technology chip sets in the wings. Lucent support is seen as a method of making sure everything is compatible at the final standard issue. Lucent's technology is said to be capable of both sending and receiving data at 45 Kbps. Now there is one other "little" "gotcha" which must be overcome in the form of a law suit concerning property rights. It seems that one individual owns the patent-rights to the current Pulse Code Modulation scheme being used.
The Standards Approval CycleThe standards approval group, the International Telecommunication Union (ITU), was to have its review work completed in about January 1998. Its original schedule had the group moving at an extra fast pace to obtain the world group approval for this standard. The whole process started in September 1996 with Rockwell's announcement of the K56 protocol. US Robotics (x2) and Lucent (V.flex2) announced their protocols in October 1996.Rockwell and Lucent announced 56Flex in November and Lucent shipped the first chip sets. US Robotics started beta testing in January 1997 and discovered the 53 Kbps downstream limitation, after which it started shipping modems in February 1997. Rockwell started shipping chip sets in February 1997 and stopped because of bugs in March. Rockwell re-started shipping chip sets in April 1997 and other vendors started shipping various modems. US Robotics guaranteed free upgrade to ITU standard in May 1997. I obtained a Hayes Accura 56Flex internal modem in July 1997 for testing. The ITU working group started meeting in March 1997. and was supposed to have a draft standard completed by September. This draft by the ITU was to be the first ITU recommendation and was to cover the 56 Kbps down-link and V.34 up-link. A second ITU recommendation which will provide full-duplex V.PCM in both directions is being drafted for the future. There has been a delay in the approval cycle due to intellectual property issues and technical roadblocks. 3Com/US Robotics sabotaged the standards process by announcing its intention to charge $9.00 for each 56 Kbps modem in a remote access server. This means that all remote access server manufacturers would have to pay 3Com/US Robotics a royalty if its technology or the 56 Kbps technology it licensed from a third party becomes part of the adopted standard. This all came about because of an agreement 3Com/US Robotics signed with one Brent Townshend, "the inventor of the 56 Kbps technology". The ITU will not issue any standard until these property rights are settled. It seems that Mr. Townshend came to Rockwell and US Robotics in 1995 with his technology ideas and was well received. US Robotics signed an agreement with him, but Rockwell did not. 3Com/US Robotics now say they have exclusive right to license 56 Kbps technology, which it claims is at the core of any 56 Kbps modem. The 3Com/US Robotics license will be $100,000 for each manufacturer or $150,000 per-piece limit total fee. Mr. Townshend has brought a law suit against Rockwell which could have the effect of stopping any further 56Flex chip set shipments until the case is over. This would place 3Com/US Robotics in an unfair advantage. Meanwhile, Lucent declining comment on the lawsuit, has stated that it holds the earliest and most basic 56 Kbps patented technology. Lucent claims it holds three patents which are basic to the V.PCM specification. And Motorola has stepped into the licensing fray by stating that it wants a royalty for its property rights. The big money stake is at the ISP end of the connection in that each ISP has had to upgrade its Point of Presence (POP) to proprietary technologies, only to have to re-upgrade them when the ITU standard arrives. This holds for local as well as national ISPs. The ITU standard will be delayed.
My 56Flex ExperienceI purchased a Hayes Accura 56Flex internal modem in August 1997. This is important because of the bugs in the early Rockwell chip sets. I installed it in a 486/100 MHZ with Windows 95 as the OS. The physical installation went without problems. When I first ran it with Win95 with Win95 Dial-Up Adapter, I got a shock in that my connections were all at speeds lower than the old Hayes Accura V.34 internal modem, an unacceptable situation.Calls to Hayes did not solve anything, but a call to my ISP did. We added addition AT commands to the Win95 modem setup Initialization string which solved the low speed problem. The modem connected properly each time and operated at high V.34 speeds, ie., 33.4 to 38.2 Kbps on downloads. Then I took it to the Resource Center and could not connect at all through the ISP 56Flex number. Again, the ISP Technical Help came to the rescue. We changed the dial-up telephone number from the 56Flex line to the x2 line. Success again. I have continued to use the x2 line wherever I take this computer, and it works. Amazing. Connection performance has been excellent upon connection. The modem holds the connection at the Resource Center without drops. In the past, we have consistently had drops with the V.34 modems we have used. There has been some failure to connect times at the Resource Center. I suspect these failures are ISP and Windows 95 related. During these failures, the modem dials, shows "dialing..." dialog box; then, "verifying user name and password...." dialog box; and fails to complete the connection. I have to shut the computer completely down and restart. Download speeds are consistently in the 33.4 to 42.2 Kbps range. I have never achieved 53 Kbps consistently. Short bursts of download speed go as high as 54 Kbps. It seems that the analog connection is dependent upon the other end and what the data must go through, as with earlier analog modems. My Courier I-modem ISDN terminal adapter (TA) is much more consistent at 40.4 to 64 Kbps downloads. It runs at 50 to 60 Kbps nearly all the time.
How to Choose a 56 Kbps ModemThe very first step in selecting a 56 Kbps analog modem is to talk to your Internet Service Provider (ISP) to determine what connection they offer. Make sure of the connections they support, either x2 or 56Flex or both. If the answer is both, then do some shopping around for the best price in a name brand.The name brand manufactured modem is key to this technology at the current time. Ascend Technologies, Boca Research, Diamond Multimedia, Hayes Microcomputer Products, Lucent Technologies, Motorola, MultiTech, Shiva, TDK, Transcend, Xircom, US Robotics, and Zoom are among the modem vendors who have introduced 56 Kbps modems. US Robotics seems to be the sales leader. The second step is to select one of the name brand 56 Kbps modems which range in price from about $119.00 to $205.00. The Hayes modem I have cost approximately $169.00. Laptop PC Card 56 kbps modems are becoming available from some of the above list. PC Card modems are in the $275.00 range. Upgrades to some high end V.34 modems such as the US Robotics Courier is possible at $60.00. This is the technique I am going through with my Courier. Remember that the external modems require the high speed 16C550 UARTs which may mean obtaining a serial I/O card with that UART installed. The third step is to install it, using the Windows 95 dialer along with the Windows 95 TCP/IP setup. Happy 56 Kbps surfing. John Woody is a telecommunications consultant specializing in small business communications, networks, and Internet business training. |