Upgraded GSM Makes UMTS an Unsafe Bet.
There are today 280 million GSM users. It is a robust, tried-and-tested piece of technology which is, however, very bad for handling data traffic. (8/23/2000)
In its basic version, GSM offers two alternatives for data transfer. One is CSD (Circuit Switched Data), with a maximum transfer speed of 9.6 kbps. The disadvantages are long uplink times, and that billing is at a rate per minute, irrespective of whether you are downloading data or not.
The other alternative is SMS, which transmits data in very small portions, at a maximum of 160 characters per packet. In the long-term, it is the third generation of mobile systems that will solve both these problems. But it will be a long time before UMTS is widely available. The first version of the standard was released just a few short months ago.
It is therefore hardly likely that UMTS operators will be offering their services earlier than 2005. Five years is an extremely long time in the world of IT, so there is little point in waiting for UMTS before launching mobile data services. So for the time being, we must do what we can with the GSM network.
GSM + High Speed
The first step is to make some small changes in the way data is coded; to bring the circuit switched data speed up from 9.6 to 14.4 kbps. Telephones manufactured from 1999 onwards can usually manage 14.4 kbps. A little more speed can be coaxed out of the system. In the GSM system, each frequency has sixteen time slots, with each call using two slots. Using more time slots will increase the data transfer speed. This is known as HSCSD (High Speed Circuit Switched Data) is, but it is referred to by both manufacturers and operators simply as High Speed.
With up to four time slots per connection, 4 times 14.4 give 57.6 kbps. If there are suddenly a large number of calls and the system needs time slots, the speed is stepped down without any connections being broken.
The problem is that as yet there are no telephones.
"This is a general problem. The same applies to Wap and Dual Band. The technology is ready before the terminals are," complains Jan Karmakar of Telia.
"We have a PCCard for computers. High Speed telephones will arrive in the third quarter [2000]", says Thomas Jönsson of Nokia.
GSM + GPRS
If GSM were capable of packet traffic, the mobile telephone would work as a computer terminal with a permanent connection to the mobile network. This solution is called GPRS (General Package Radio Service). With GPRS, data can be sent in packets over the mobile network in more or less the same way as an IP packet is sent over the Internet. The packets do not reserve any capacity, but drop into a free time slot which they only use as long as it takes to transfer the data. At 21.4 kbps, the data capacity per time slot is a little higher for GPRS than for circuit switched data transfer. But as everyone shares all the time slots, a theoretical transfer speed of 171 kbps can be achieved if you are the sole user of that frequency.
"In purely practical terms, each user may reach 20 kbps with this service. The great advantage of GPRS is not speed, although the speed is higher, but that you are always connected to the network," says Nokia's Kaj Hagros.
GPRS requires a larger network installation. In theory, a complete parallel network for data traffic must be built. In practice, it is enough to upgrade each base station controller (BSC), and it is not necessary to upgrade each separate base station. Today, both Ericsson and Nokia have pilot installations of their GPRS equipment. They expect to have complete telephones and systems on the market by the end of the year.
GSM + Edge
Edge can be used to reach even higher data speeds. By using phase modulation, edge introduces more signal levels, and can thus fit three times more data per time slot. GPRS and Edge give speeds of up to 450 kbps (once again, capacity that you have to share with ordinary telephone traffic and all other data traffic).
The disadvantage of the Edge modulation method is that the range is smaller, with the cell radius being perhaps 15% less. Installing Edge is also a fairly extensive process, requiring the replacement of all base station transceivers.
Nokia and Ericsson expect to have telephones and equipment on the market almost exactly one year after GPRS, i.e. at the end of 2001. Count on at least one year before the operators have it in full use.
"We have just decided to use GPRS in our network. We have made no decision on Edge as yet; first we shall have to see how GPRS sells," comments Jan Karmakar of Telia.
UMTS Requires Too Many Base Stations
Now we have a fully hotted-up GSM network, with HSCSD, GPRS, and Edge, giving data speeds of 450 kbps per frequency. Why, then, would we want UMTS?
UMTS only gives data speeds that are four times higher, and a high price has to be paid. A completely new mobile network must be built (although much of the GSM infrastructure can be used), and this network still cannot replace GSM. The fact is that UMTS can never offer nationwide coverage. Its higher frequencies mean that base station range is substantially less than that of GSM. For full speed, the user cannot be more than 500 meters from the base station, and the maximum range is 1.7 kilometers.
Is UMTS really worth the cost of building a new mobile network? Christoffer Andersson at Ericsson thinks so:
"Data speed is just a small part of all the advantages of UMTS. UMTS has traffic prioritization, making it possible to introduce a range of completely new services. And the network has a far higher capacity. A typical GSM base station can handle 300 calls, while even the first release of UMTS can handle 2800 calls."
It remains to be seen whether consumers will be equally convinced of the benefits of UMTS. A lot can happen in five years.
mformobile.com |
