GPRS Q&A

All answers have been held as short as possible and require an understanding and study of the corresponding chapter of the book.

Answer 1:
When data is transferred over a circuit switched channel, a dedicated connection is established between two parties. Data is sent without any overhead like lower level addressing. Bandwidth and delay are constant. In a packet switched network on the other hand, there is no direct connection between the endpoints of a session. Resources in the network are only used for the connection when data is sent. Data is sent in packets which have to contain a source and destination address in order to be transported through the network. This also enables N:N connections in the network, i.e. a subscriber can communicate with any subscriber without establishing a physical connection first. Depending on the load of the network, bandwidth and delay for a connection can vary. This is a clear disadvantage compared to a circuit switched channel. Due to the bursty nature of many information exchanges the advantage of the packet switched approach on the other hand is to use more bandwidth during the burst which decreases transmission time.

Answer 2:
As GPRS is a packet switched network, resources or the air interface are only assigned to a user when data is actually sent. This tremendously increases the capacity of the network especially for applications such as web surfing which only send and receive data at irregular intervals. Several timeslots can be assigned to a subscriber simultaneously to increase throughput. If the physical connection to the network is lost (e.g. due to bad reception quality) the logical connection persists. As soon as the physical connection has been reestablished, data transfer on higher layers resumes without the user having to reestablish another channel manually as would be the case for a circuit switched connection.

Answer 3:
Dynamic coding schemes allow to adapt the ratio of error correction and detection bits to user data bits. For good transmission conditions the redundancy information in a block can be reduced which in turn increases the overall transmission speed of the user data. During times of bad reception, more error detection and correction bits are inserted which ensures that the link remains stable.

Answer 4:
While in GPRS ready state the SGSN can send data to the mobile terminal without delay. In this state, the SGSN is aware of the cell which the subscriber uses to communicate and thus can forward incoming packets directly to the PCU responsible for this cell. The PCU does not need to page the subscriber and can immediately assign resources on the air interface. When changing the cell in ready state the mobile station has to send a cell update message to the SGSN. Once the mobile station is in GPRS standby state, the SGSN is only aware of the location are of the subscriber, as the mobile station only has to report cell changes when a location area boundary is crossed. This reduces the mobile’s energy consumption. In order to send data frames to a mobile in standby state, the SGSN has to page the subscriber first. The mobile station responds with an empty frame and thus implicitly changes into the ready state again. Once the SGSN receives the empty frame it is aware again of the cell the mobile station uses and can then forward the frame.

Answer 5:
In practice, no handovers are performed for GPRS today (Network Control Order = 0). The mobile station has to perform cell changes on its own. In case a cell change has to be performed during an ongoing data transfer due to deteriorating reception conditions it is necessary to interrupt the transmission and perform the cell change. Afterward, the mobile station reports to the SGSN from the new cell by continuing to send data. The SGSN detects the cell change as the cell global ID is part of every incoming frame and can thus change its routing of incoming Internet packets to the new cell.

Answer 6:
GPRS requires the following network nodes: A) The serving GPRS support node (SGSN) which is responsible for mobility management and session management (GMM/SM). B) The gateway GPRS support node (GGSN) which is the interface between the GPRS network and the Internet. The GGSN is responsible for the assignment of IP addresses to the mobile subscribers and hides subscriber mobility from the Internet. C) The packet control unit (PCU) is the interface between the GPRS core network and the radio network. The PCU is responsible for packet scheduling, assignment of timeslots to the subscribers and terminates the RLC/MAC protocol.

Answer 7:
GPRS assigns resources (timeslots) to a subscriber only for the time required to send the data. Furthermore, timeslots are not exclusively assigned to a single subscriber but only in blocks of four bursts. This way, timeslots can be used to transfer data to several subscribers at the same time. The temporary block flow with the temporary block identifier describes which data blocks are addressed to which device currently listening on a timeslot.

Answer 8:
An Inter-SGSN routing area update is performed if the mobile device roams into a cell which is connected to a new SGSN. As the new cell belongs to a new routing area, the mobile device attempts a routing area update. The new SGSN then detects that the mobile device is currently registered with a different SGSN and thus sends a message to the previous SGSN to retrieve authentication information. After authenticating the mobile station, the HLR is informed that the subscriber has changed its location to the new SGSN. Furthermore, the GGSN is informed of the position change so it can forward incoming packets to the new SGSN in the future. Once all of these actions are performed, the routing area update in the core network is complete and the subscriber gets a confirmation from the SGSN that the operation was performed successfully.

Answer 9:
The GPRS core network between the SGSN and GGSN use the IP protocol for routing the IP data frames of subscribers. These are not transferred directly, however, but are encapsulated into GPRS tunneling protocol (GTP) frames. Part of the encapsulated frame is the IP address of the mobile device and the source/destination of the frame. Thus, a GTP frame contains two source and two destination IP addresses. This mechanism has the advantage that no routing table updates are required in routers between these two network components if the user is roaming into the area of another SGSN. In addition, the GPRS core network is decoupled from the Internet and the GPRS user as it is not possible to directly access these components from outside the local GPRS core network.

Answer 10:
The user does not have to change any settings on his/her device for international roaming. All packets that are sent and received are always routed through the GGSN in the subscriber’s home network. This is possible as the access point name (APN) is a qualified domain name and the SGSN inserts the mobile country code (MCC) and the mobile network code (MNC) as well as a top level domain (‘.gprs’) to the APN string received by the subscriber during the connection establishment. This domain name is then sent to a DNS server which resolves the domain name into the IP address of the GGSN in the subscriber’s home network.

Answer 11:
During a GPRS attach, the mobile device registers with the network. Afterward, the network is aware that the device has been switched on and in which routing area it is located. Up to this point no IP address has been assigned to the mobile device and no data can be transmitted. The IP address is only assigned to a mobile device during the PDP context activation procedure. Billing is also only invoked during the activation of a PDP context.

Answer 12:
In order to transfer data via GPRS to and from the Internet, a PDP context has to be established between the mobile device and the GPRS network. During the establishment of a PDP context the mobile device sends the access point name which identifies the GGSN and profile to be used to establish a connection to the internet. (Also see answer 10)

Answer 13 (removed in the 3rd edition as MMS is no longer covered):
MMS messages are always exchanged between a mobile device and the MMS gateway which is located behind the GGSN. For sending an MMS the mobile device first establishes a GPRS connection and uses the APN which the network operator has foreseen for the MMS service. Usually it is only possible to reach the MMS gateway via this APN. Afterward, the MMS, which has many similarities to an eMail, is sent by using an HTTP-PUT push command. This command is also used by web browsers to send the input the user has made on a web page in text fields, etc. to the web server. Once the MMS is received by the MMS gateway it is stored and an attempt is made to deliver the message to the destination. If the destination is a mobile subscriber, an SMS is sent to inform the mobile device of the waiting MMS message. Depending on the configuration of the mobile device it either establishes a GPRS connection immediately or queries the user first before doing so. To receive the MMS message the mobile device uses the HTTP-GET command. This command is also used by web browsers to request web pages from a web server.

Answer 14 (removed in the 3rd edition as MMS is no longer covered):
An MMS message has many similarities to an eMail. The header for example is structured in a similar way as an eMail header and just contains additional X-MMS-tags which contain MMS specific information. Text and pictures are sent in the “body” of the MMS and are separated by Multipurpose Internet Mail Extension (MIME) separators. The first part of an MMS body is the description of the general layout of the message. SMIL, an XML language, is used for this purpose. Further MIME parts of the MMS then contain the text, pictures, videos, etc.

GSM Q&A

All answers have been held as short as possible and require an understanding and study of the corresponding chapter of the book.

Answer 1:
In a circuit switched digital telecommunication network a speech channel usually uses a 64 kbit/s timeslot. The pulse code modulation (PCM) algorithm is used to convert an analog voice signal for digital transmission.

Answer 2:
The GSM NSS consists of at least of the following network components:

  • MSC: The Mobile Switching Center, which includes the Visitor Location Register, is responsible for connecting calls and mobility management.
  • The Home Location Register, which stores subscriber information
  • Service Control Points, which handle services like prepaid or location based billing applications
  • SMS Service Center
  • Voice Mail Systems

Answer 3:
The GSM BSS consists of the following network components:

  • The Base Station Controller, which controls the channels on the air interface. It is also responsible for power control, timing advance control and handovers.
  • Transcoding and Rate Adaptation units convert the speech codecs used in the BSS into 64 kbit/s PCM coded channels used in the core network. (Note: The TRAU is usually co-located with the MSC in order to minimize the number of required transmission links between the MSCs and the BSCs).
  • The Base Transceiver Stations (BTS): A high number of base stations are connected to a base station controller. They are responsible for transferring data streams over the air interface.

Answer 4:
Three methods are used: Several calls can be handled on the same frequency. This is done by splitting a channel into 7 timeslots. This is called Time Division Multiple Access (TDMA). Capacity can be further increased by using several carrier frequencies per base station. Typical GSM base stations today use one to three carriers. This is called Frequency Division Multiple Access (FDMA). Additionally, a further increase can be achieved by splitting the coverage area of a base station into several sectors and by using different carrier frequencies in each sector. Typical base stations today use either two or three sectors.

Answer 5:
The first step is to digitize the voice signal recorded by the microphone by using the PCM codec. This digital signal is then used as input for the speech coder (Full Rate, Enhanced Full Rate, AMR) which compresses the data generated by the PCM codec. This data stream is then sent to the channel coder which adds redundancy (error detection and error correction bits). Afterwards, the Interleaver changes the order of the bits and spreads consecutive bits over a wider area of the data stream. This allows spreading transmission errors which usually affect several consecutive bits over a larger area which helps the receiver to detect and correct errors. To protect the communication from eavesdropping and tampering the resulting data stream is then ciphered. Finally the now encrypted data stream is modulated onto an analog carrier frequency and transmitted.

Answer 6:
A handover changes the routing of an ongoing voice call from one cell to another. This is necessary if a subscriber leaves the coverage area of a cell and enters the coverage area of another. Network elements involved in the handover are the mobile station, the old and the new base station and the base station controller. If the current base station controller is not responsible for the new cell, the mobile switching center and the BSC controlling the new cell are also part of the handover procedure.

Answer 7:

For an incoming call the Gateway MSC queries the home location register of the subscriber to ask for its location (Send Routing Information). The HLR is aware of the current MSC/VLR (Visited MSC) of the subscriber and requests a temporary identifier, the Mobile Station Roaming Number (MSRN). This number is returned by the HLR to the Gateway MSC. From there the call can then be forwarded to the Visited MSC. At this stage of the call routing process, standard ISUP signaling is used which is already known from fixed line networks. Instead of using the phone number as identifier, the MSRN is used to route the call. On the Visited MSC, the temporary MSRN is used to correlate the call to the correct subscriber. As the V-MSC is only aware of the subscriber’s current location area, the subscriber has to be paged in order to establish the cell to which the call has to be forwarded to.

Answer 8:
In order to authenticate a subscriber, the MSC queries the Authentication Center for the Authentication Triplets of the subscriber. Elements of each triplet are a random number (RAND) and a response value (SRES) which is generated from the random number with an authentication algorithm. The MSC then sends the random number to the mobile station. The MS forwards the random number to the SIM card which computes the SRES. This value is then sent back to the network. The MSC then compares the SRES computed by the authentication center and the SRES computed by the SIM card. The two values can only match if the authentication center and the SIM card have used the same secret key to generate the signed response from the same random number.

Answer 9:
A mobile phone always sends an SMS to the SMS Service Center (SMSC) which is a network node in the core network. The SMSC then uses the phone number contained in the short message to query the HLR for the current MSC of the destination subscriber. It then forwards the message to this MSC. If the subscriber is not reachable, the Message Waiting Flag is set in the MSC/VLR and HLR, and the SMS is stored in the SMSC. Once the subscriber becomes reachable again the SMSC is informed and another delivery attempt is undertaken.

Answer 10:
In downlink direction (network to mobile phone) the DSP is used for the analysis of the incoming data stream. The training sequence of a burst is used to calculate a channel approximation which is then applied to the main parts of each burst. This improves compensation of external effects on the transmission. Furthermore, the DSP compresses and decompresses the speech data by using the Full Rate / Enhanced Full Rate / AMR codec. The RISC processor on the other hand deals with channel coding and decoding, interleaving/de-interleaving and ciphering/de-ciphering of the incoming and outgoing data stream. Additionally, the RISC processor also manages the user interface, the overall control of a connection (Mobility Management / Session Management), user programs and external interfaces (Bluetooth, USB, etc.).

Answer 11:
Data is stored on the SIM card in a non-volatile, re-writable memory. To the outside, the microcontroller on the SIM card presents the memory as a directory tree and files. While PCs use filenames to identify files and directories, the SIM card instead uses 4 digit hexadecimal numbers. Each file has its set of individual read and write permissions. Some files are readable only, such as the file that contains the IMSI of the user. Some files can neither be read nor written to from the outside, like for example the file which contains the secret key used for authentication. Directories are referred to as ‚dedicated files’, ordinary files are called ‚elementary files’.

Answer 12:
CAMEL is the abbreviation for Customized Applications for Mobile Enhanced Logic. It describes how databases and applications can communicate over network boundaries and interfaces via MSCs, SGSNs and GGSNs. CAMEL is used for services such as prepaid or location based services. As CAMEL is not restricted to the home network of a user, foreign MSCs are able to contact databases in the home network of a subscriber for services such as prepaid billing when a subscriber is abroad. CAMEL specifies both the protocol between the network nodes and a state model which describes the different phases of a call by using Detection Points. If a certain detection point is activated for a subscriber, the treatment of a call is suspended at this point and the database and service responsible for the subscriber is contacted for further instructions of how to proceed with the call.

Mobile Monday Paris April – Pictures and Report

1076_030420060081079_03042006011
1080_03042006012
1083_03042006015 

My first Mobile Monday in Paris. I was very positively surprised to see so many people there and had many good discussions. As you can see in the pictures the number of people has easily surpassed 200. Here are some details of the event as it happend.

Presentations: (here’s a link to most presentations)

The first presentation was by Vincent Veran of Axalto, one of the leading SIM card manufacturers. Their latest ideas they are working on are large memory SIM cards for preloaded content and storage of content such as eMail, a SIM card based web server for applications such as easy access to the operator portal and contact less applications embedded in the SIM. For the later, applications such as contact less payment in the metro (similar to the Felica service in Japan) come to mind.

The second presentation was given by Alex Kummerman of Clicmobile and he gave an overview of a current proof of concept application his company is trialing in Geneva at the moment which deals with finding out if some of your friends are near you while being at a certain location.

Mobile Gaming was the theme of the third presentation of Nicolas Caris of Acute about Urban Rivals. urban-rivals.com is an online game which seems to be an adaptation of a game initially launched back in 2001 in Japan (my interpretation as in 2001, iMode was only available in Japan). How I would have liked to have that in my pocket back then in school 🙂

Finally, the fourth presentation was given by Valérie Beaudouin of France Telecom about perceived and real usage of mobile and mobile services.

Discussions:  Glad to have met Stuart Mudie at the MoMo to discuss many on and of topic things. Other people I have spoken to include someone who promotes tracedog.fr, a cool utility for dog owners how get tired of searching their dog. I really like the application, check out their web site!

Operator and supplier attendance: Thanks for hanging out the list of attendees and separating the participants in operators, suppliers, mobile media, etc. Interesting to see many people in the the operators category, lots of people from Bouygues and Orange, only few from SFR though. To my surprise only few people from suppliers at the meeting. Some from Ericsson and only one or two from Nortel and Alcatel. I find that somewhat strange.

Summary: All in all a great evening, a big thank you to the organizers and the speakers of the evening. Looking forward to the next MoMo Paris on the 5th of June.