Page tree

The Global System for Mobile Communications (formerly Groupe Spécial Mobile, GSM) is a mobile radio standard for fully digital mobile radio networks, which is mainly used for telephony, but also for circuit-switched and packet-switched data transmission and short messages. It is the first standard of the so-called second generation ("2G") as successor of the analog systems of the first generation (in Germany: A-net, B-net and C-net) and is the most widely used mobile radio standard worldwide.

GSM was created with the aim of offering a mobile telephone system that allowed subscribers mobility throughout Europe and offered voice services compatible with ISDN or conventional analogue telephone networks.

In Germany, GSM is the technical basis of D and E networks. GSM was introduced here in 1992, which led to the rapid spread of mobile phones in the 1990s. Today, the standard is used as the mobile communications standard in 670 GSM mobile communications networks in around 200 countries and regions of the world; this corresponds to around 78 percent of all mobile communications customers. Later additions to the standard such as HSCSD, GPRS and EDGE have been added for faster data transfer.



In contrast to the fixed network, a mobile network has various additional requirements:

  • subscriber authentication
  • Channel access method
  • Mobility management (HLR, VLR, location update, handover, roaming)
  • The subscribers are mobile and can therefore switch from one radio cell to another. If this happens during a call or a data connection, the call connection must be transferred from one base station to the next (handover) so that the mobile phone always gets its radio connection to the most suitable base station. In exceptional cases, the call can also be routed via an adjacent base station to avoid overloading.
  • efficient resource utilisation
  • Since a lower data transmission rate is available on the radio interface than in the fixed network, the user data must be compressed more strongly. In order to keep the proportion of data transmission rate that must be used for signalling processes small, the signalling messages were specified bit-precisely in order to keep them as short as possible.
  • Mobile phones have a limited battery capacity that should be used sparingly. In general, transmitting costs more energy than receiving. Therefore, the amount of data sent and status messages should be kept as low as possible in standby mode.
  • Use of external networks (roaming)

Some important functions within mobile networks


→ main article: handover

Inter-Cell handover

One of the most important basic functions in cellular mobile radio networks is the cell change initiated by the network during an ongoing call. This can be necessary for various reasons. Decisive factors include the quality of the radio link, but also the traffic load of the cell. For example, a call can be transferred to a remote cell to avoid overloading.

Intra-cell handover

Here, for example, a new channel is assigned to the MS within a cell due to the channel quality.


→ main article: roaming

As many mobile phone operators from different countries have signed roaming agreements, it is possible to use the mobile phone in other countries and to continue to be reachable under one's own number and to make calls.

Voice transmission

Over the years, several codecs have been standardized for voice transmission in GSM. The usual speech codecs, which typically manage with a data rate of less than 20 kbit/s, perform a feature extraction adapted to human language, making them usable only for the transmission of speech. Music or other noises can therefore only be transmitted with lower quality. The following is a brief summary of the voice codecs used in the GSM network:

Full Rate Codec (FR)

The first GSM voice codec was the full-rate codec (FR). It has a net data rate of only 13 kbit/s (in contrast to G.711 64 kbit/s with ISDN). The audio signals must therefore be highly compressed, but still achieve an acceptable voice quality. The FR codec uses a mixture of long-term and short-term prediction that enables effective compression (RPE/LTP-LPC voice compression: linear predictive coding, long-term prediction, regular pulse excitation).

Technically, each 20 ms language is sampled and buffered, then subjected to the speech codec (13 kbit/s). For forward error correction (FEC), the 260 bits of such a block are divided into three classes, according to the extent to which a bit error would affect the speech signal. 50 bits of the block are divided into class Ia. They have the highest protection and receive a CRC checksum of 3 bits for error detection and error concealment. Together with 132 bits of class Ib, which are slightly less protectable, they are subjected to a convolutional code that generates 378 output bits from the 185 input bits. The remaining 78 bits are transmitted unprotected. This turns 260 bits of user data into 456 bits of error-protected data, increasing the required bit rate to 22.8 kbps.


The 456 bits are divided by interleaving into eight half bursts of 57 bits each. After deinterleaving in the receiver, short-term interference (e.g. one burst long) is minimized by the error spread. By combining the different error protection procedures in the GSM, good voice quality is often achieved even though the radio channel is extremely error-prone.

Half Rate Codec (HR)

With the introduction of the half rate codec it became possible to handle not only one but two calls simultaneously on one time slot of the air interface. As the name says, only half the data rate is available for HR as for the FR codec. To achieve a usable voice quality, a vector quantization is used instead of the scalar quantization used in the FR codec. This means that coding requires approximately three to four times the processing power of the FR codec. Because voice quality is still rather poor, HR is only used by mobile network operators when a radio cell is overloaded.

Enhanced Full Rate Codec (EFR)

EFR works with a data rate similar to the Full Rate Codec, namely 12.2 kbit/s. Due to a more powerful algorithm (CELP) a better voice quality was achieved compared to the full-rate codec, which corresponds approximately to the level of ISDN telephone calls (G.711a) with a good radio channel.

Adaptive Multirate Codec (AMR)

AMR is a parameterizable codec with different data rates between 4.75 and 12.2 kbit/s. In the 12.2 kbit/s setting, it largely corresponds to the GSM-EFR codec in terms of both algorithm and audio quality. The lower the data rate of the voice data, the more bits are available for channel coding and thus for error correction. Thus, the 4.75 kbps codec is described as the most robust, because an understandable conversation is still possible despite the high bit error frequency during radio transmission. During a call, the mobile network measures the bit error frequency and selects the most suitable codec from a list, the Active Codec Set (ACS). The code rate used is thus continuously adapted to the channel quality.

data transmission

If a GSM channel is used for data transmission, a usable data rate of 9.6 kbit/s is obtained after the decoding steps. This type of transmission is called Circuit Switched Data (CSD). Advanced channel coding also enables 14.4 kbit/s, but causes many block errors in poor radio conditions, so that the "download rate" can actually be lower than with increased security on the radio path. Therefore, depending on the bit error frequency, between 9.6 and 14.4 kbit/s is network-controlled (=Automatic Link Adaptation, ALA).

However, both are too few for many Internet and multimedia applications, so that extensions under the names HSCSD and GPRS have been created that enable a higher data rate by allowing more bursts per time unit to be used for transmission. HSCSD uses a fixed assignment of several channel slots, GPRS dynamically uses radio slots for the connected logical connections (better for Internet access). E-GPRS is a further development of GPRS. This is the use of EDGE for packet data transmission.

Extensions and further developments of GSM

GSM was originally designed mainly for telephone calls, faxes and data transmissions at a constant data rate. Burst-type data transmissions with strongly fluctuating data rates, as is usual with the Internet, were not scheduled.

With the success of the Internet, the so-called "evolution of GSM" began, in which the GSM network was completely downwardly compatible with possibilities for packet-oriented data transmission. In addition, only minimal costs should be incurred through the replacement of frequently used components.


Speeds of up to 14.4 kBit/s are achieved with Circuit Switched Data.


By coupling several channels, HSCSD achieves a higher data rate overall, maximum 115.2 kbit/s. In order to use HSCSD, a compatible mobile phone is required; on the part of the network operator, hardware and software changes are required for components within the base stations and the core network. In Germany, only Vodafone and E-Plus support HSCSD.


GPRS allowed packet-switched data transmission for the first time. The actual data throughput depends on the network load and is a maximum of 171.2 kbit/s. At low load a user can use several time slots in parallel, while at high network load each GPRS time slot can also be used by several users. However, GPRS requires additional components (the GPRS packet core) from the network operator within the core network.


With EDGE, a new modulation (8PSK) made it possible to increase the data rate. It is a maximum of 384 kbit/s. EDGE expands GPRS to E-GPRS (Enhanced GPRS) and HSCSD to ECSD (Enhanced Circuit Switched Data).

Frequencies used

GSM works with different frequencies for uplink (from mobile phone to network) and downlink (from network to mobile phone). The following frequency bands can be used by the mobile operator:

Band designationRangeUplink (MHz)Downlink (MHz)ARFCNContinentNotes
T-GSM 380GSM 400380,2 – 389,8390,2 – 399,8dynamic

T-GSM 410GSM 400410,2 – 419,8420,2 – 429,8dynamic

GSM 450GSM 400450,4 – 457,6460,4 – 467,6259 – 293

GSM 480GSM 400478,8 – 486,0488,8 – 496,0306 – 340

GSM 710GSM 700698,0 – 716,0728,0 – 746,0dynamic

GSM 750GSM 700747,0 – 762,0777,0 – 792,0438 – 511

T-GSM 810
806,0 – 821,0851,0 – 866,0dynamic

GSM 850GSM 850824,0 – 849,0869,0 – 894,0128 – 251America
P-GSMGSM 900890,0 – 915,0935,0 – 960,01 – 124Africa, America, Asia, Australia, Oceania, Europe
E-GSMGSM 900880,0 – 915,0925,0 – 960,00 – 124, 975 – 1023Africa, America, Asia, Australia, Oceania, Europe
R-GSMGSM 900876,0 – 915,0921,0 – 960,00 – 124, 955 – 1023Africa, Asia, Europe
T-GSM 900GSM 900870,4 – 876,0915,4 – 921,0dynamic

DCS 1800GSM 18001710,0 – 1785,01805,0 – 1880,0512 – 885Africa, America, Asia, Australia, Oceania, Europe
PCS 1900GSM 19001850,0 – 1910,01930,0 – 1990,0512 – 810America
  •     Frequency bands 2 and 5 (blue background color) are commercially used in America.
  •     Frequency bands 3 and 8 (yellow background color) are used commercially in Europe, Africa, Asia, Australia, Oceania and partly in America.
  •     All other frequency bands are not used commercially in public mobile networks.
  •     There is no public GSM mobile phone network in South Korea and Japan.
  •     A mobile phone that supports the GSM and UMTS FDD frequency bands 5 (850 MHz), 8 (900 MHz), 2 (1900 MHz) and 1 (2100 MHz) is suitable for worldwide use.

For cost reasons, the construction of new mobile radio networks (e.g. Australia/Telstra) or mobile radio network extensions (e.g. Switzerland/Swisscom) was only carried out with the newer UMTS mobile radio technology. New mobile phone stations are increasingly sending out only one UMTS and LTE signal.

GSM disconnection

GSM is expected to be replaced by successor standards in the long term. While Australia and Singapore have already decided to switch off in 2017, Germany and Austria, for example, have not yet set a shutdown date, but in Switzerland no public GSM mobile phone network will probably be available from 2021.