The present situation in the modem market is confused. All the modem manufacturers have released 56K designs, based on one system or the other. Some companies have even hedged their bets by releasing both X2 and K56flex devices. This all adds extra complications for the prospective purchaser, whether a user or an ISP. Each is trying to guess which system the other will choose. Will the majority of users follow USR? Will most ISPs upgrade their existing equipment to K56flex? For both, the fear is that their investment will become obsolete if they make the wrong choice.
US Robotics' X2 is now available in the US and Europe. The European launch was delayed by some months due to technical problems - the design worked correctly with the US' �-law companders, but not with Europe's a-law companders. As a result, it was July before X2 became available in the UK. US Robotics has recently undergone a merger with the large networking company, 3Com. This led to speculation of a change in USR's marketing, as 3Com is a member of the K56flex consortium. So far, however, there appears to have been no change to either company's strategy. The combined company, still called 3Com, may be maintaining its presence in both camps in order to reduce the market risks.
Lucent and Rockwell's development has been lagging USR's for some time. Their situation is somewhat different, however, as they are not producing modems directly. Both companies supply integrated chip-sets to other manufacturers, such as Hayes and Motorola, who then develop the final product. The chip-sets have now been released, and K56flex modems are available from a number of different companies. Lucent also claims to be able to take the technology further. Late last year, they announced the development of a system that would allow an upstream data rate of 40 kbit/s, as well as the 56 kbit/s downstream rate. There was no further discussion of this until the beginning of August, when the project resurfaced in a new guise. Lucent is now developing a symmetrical 45 kbit/s modem, for release in 1998. Rather than Internet access, this modem will be aimed at applications requiring equal data rates, such as video conferencing.
Which system will dominate is difficult to predict. USR reached the market first with X2, but this lead was offset by the technical problems in Europe. Rockwell already has a large number of people using its existing chipsets, and they could potentially upgrade to K56flex. This user-base includes ISPs, as several manufacturers of rack-mounted modems, such as Ascend, use the Rockwell chipsets.
Further complicating the choice is the promise of an international standard. Both camps have submitted their systems to the International Telecommunications Union (ITU), hoping that they will form the basis of a new standard. Given the usual speed of the standards process, nothing is expected until 1998: little assistance to those trying to make a purchasing decision. The latest news is that the standard may be further delayed. At a recent meeting of the ITU, it emerged that there are several Intellectual Property and Patent problems to be overcome before a common coding method can be set. In an attempt to boost sales, USR has promised to upgrade its X2 modems to the future ITU standard. This is a gamble: there is an assumption that the standard will only require a software upgrade, and that the existing hardware will be capable of whatever new processing is required. Many other manufacturers have followed this lead - promising an upgrade is now standard, rather than an exception. The modem producers clearly believe that this market will be important, and are willing to fight hard for market share.
Initial performance tests suggest that there is little difference between the two systems. Several magazines, mainly in the US, have published group tests. Whilst these can be useful guides, it should be remembered that the telephone system in the US is significantly different to that in Europe, with the result that performance may not be the same. This is not necessarily bad news: the US system suffers from a couple of unique restrictions. The important consideration for European users is how well the manufacturers have implemented their a-law companders, an area that has already caused problems for USR.
The general consensus among the reviews is that both systems are capable of greater than 50 kbit/s on noise-free lines. However, both are badly affected by noise, and it is here that some manufacturers' devices stand out. The real-world performance appears to be more dependent on good modem design than on whether X2 or K56flex is used. Under identical test conditions, some modems are forced back to V.34 speeds, whilst others still maintain 40 kbit/s or better. A recent survey revealed that users of the new modems considered a data rate of around 45 kbit/s to be a good connection.
Another important consideration facing purchasers is the lifetime of the 56 kbit/s modem. The improved speed would seem less impressive if there was a better, faster system just around the corner. An entirely new, high-bandwidth network does seem unlikely, due to the huge investment required. A small number of people have access to Cable Television networks, which may provide a higher-speed connection in the future. Despite enthusiastic reports on its performance, however, availability is very restricted, even in the US. The satellite networks are several years away - if they prove to be technically feasible at all - and ISDN has been sidelined by its cost.
The most likely contender for future internet access is one of the Digital Subscriber Line (DSL) techniques. The most favoured variation seems to be Asymmetric Digital Subscriber Line (ADSL), which promises to provide faster data transfer through the existing telephone wiring. It will, however, require major investment at the exchanges. It was mentioned earlier in this article that the bandwidth of the telephone line is artificially restricted; ADSL requires the removal of this restriction, together with the installation of new Switches that can handle the faster data. Many telephone companies see it as a convenient interim solution - it defers the expensive replacement of analogue lines, whilst hopefully satisfying the users' demands for more bandwidth.
However, ADSL still calls for substantial investment, which has not yet been forthcoming. Barring further innovation, 56 kbit/s is likely to be the highest speed generally available until this happens. The modem companies appear concerned that the possibility of ADSL in the near future may deter those considering purchases now. Several manufacturers, including USR, are believed to be working on devices capable of both 56 kbit/s and ADSL. Early 1998 was considered to be the likely release date for USR's development, although it is unlikely that an ADSL network will have been created by then.
Back to The Upstream Channel - Contents - On to Conclusion