Patent Application
20070104094
The present invention relates to a method for controlling data connections in a data network having a plurality of data network nodes according to claim 1, and a corresponding device according to claim 9.
In data networks having a plurality of data network nodes, a data connection takes place from a source to a destination via a route or a link in the data network which includes a certain number of data network nodes. In an end-to-end connection in such a data network, therefore, data packets run from the source to the destination via a plurality of data network nodes. These data network nodes work like relay stations.
In a data network having dynamic or mobile data network nodes, for example, a radio communications network having mobile radio terminals as data network nodes, the transmission quality of a data connection fluctuates as a rule. Such data networks will gain increasing importance in the future, in particular because of the wide distribution of wireless LAN networks occurring in the meantime, and the introduction of UMTS (universal mobile telecommunications system). Therefore, an important goal when working with such data networks is to achieve a high data transmission rate, in spite of the fluctuating transmission quality.
In radio communications networks, a distinction is made between subscriber networks and ad hoc networks. Ad hoc networks permit a self-organizing organization. Since, as a rule, the radio range in radio systems having a high data rate is limited, so-called multihop data connections are increasingly used in data networks. For example, in an ad hoc radio communications network, individual laptop computers may be used as data network nodes which are provided with a radio interface operated in ad hoc mode. Multihop data connections are used, above all, when the destination is not located in the direct radio range of the source. The connection setup, the connection finding and the restoration of connections are ensured by special routing protocols.
As soon as a data connection between a source and a destination has been found and set up, it may be used for data transport. Depending on the dynamics of the data network nodes participating in the data connection, the quality of the data connection may now change continuously. In the worst case, the data connection may be interrupted if a data network node participating in the data connection moves out of the radio range of its adjacent data network node. In such a case, as a rule, a data network node adjacent to the breaking point informs the source about the breaking point in the data connection. The source then tries again to set up a new connection to the destination. To do this, the source initiates the renewed transmission of a so-called connection search message, which is sent through the entire data network. In this case, one speaks of “flooding” the data network, since the connection search message “floods” the entire data network. The disadvantage with this “flooding” is the high requirement for network capacity. In addition, the network capacity required for the “flooding” is not available to other network connections.
Moreover, upon breaking a data connection, no data transport is possible any longer between source and destination. Data packets which were already sent from the source to the destination, but which could not pass the breaking point in time, are lost. Thus, such data packets must be sent again by the source to the destination, resulting in an increase in packet delay as well as a waste of capacity in packet-oriented data networks. In transport protocols, such as TCP (transport control protocol) used on the Internet, these great delays lower the throughput of a data connection, and therewith of the entire data network.
In order to achieve an especially high quality in a data connection, the continuity of an end-to-end connection is therefore of substantial importance. This applies especially if commercial services are to be offered over such data networks.
It is therefore the object of the present invention to provide a method and a device for controlling data connections in a data network having a plurality of data network nodes, which are able to ensure an end-to-end data connection that is as free of interruptions as possible.
This objective is attained by a method for controlling data connections in a data network possessing a plurality of data network nodes having the features of claim 1, and by a corresponding device having the features of claim 9. Further refinements of the invention are derived from the dependent claims.
An important idea, on which the present invention is based, is to judge the quality of a data connection in a data network having a plurality of data network nodes according to the coding used. In this patent, the term coding is used for the combinations of physical transmission methods employed. As an example, we mention here the combination of modulation and puncturation in the OFDM method, or the various chip rates in the CDMA method. In this context, a high coding corresponds to a high data rate. As a rule, the data are transmitted in differently coded fashion over a data connection so as to assure a transmission that is as secure and interference-resistant as possible, as a function of channel quality. Especially in radio communications networks that are used as data networks, coding is of great importance for a sufficient transmission quality. In most data networks, various coding methods are used, which may be adapted dynamically to the quality of a data connection. This means that, when there is a deterioration in the quality of a data connection, the coding method is adapted to the deteriorated data connection. This is especially the case in radio communications networks such as WLAN's, according to the IEEE 802.11 standard or the ETSI HIPERPLAN/2 standard in GSM/GPRS, CDMA and UMTS mobile radio networks. Thus, the coding method used for a data connection reflects the quality of the data connection. By evaluating the coding methods used for a data connection, particularly the characteristic of the coding methods used, it is now possible to assess the quality of a first data connection, and depending on that, a second data connection may be set up which is used as a substitute data connection in case of the breakdown or in the case of a severe interference in the first data connection. Thus, the method according to the present invention detects the breakdown of a data connection, or a threatening interruption in the data connection, in timely fashion, and ensures that an alternative data connection is searched for and found. At the moment of a disconnection of the first data connection, or even earlier, it is thereby possible to change to a second data connection without any delay. This leads to a reduction of the control signals in the data network, especially of connection search messages which, up to now, were sent out by a source after the disconnection of a data connection for the purpose of finding a new data connection. In addition, packet losses are minimized or considerably reduced, if not really completely eliminated. Finally, data or data packets are also prevented from having to be transmitted again in response to a disconnection of the data connection.
Specifically, now, the present invention relates to a method for controlling data connections in a data network having a plurality of data network nodes, in which the quality of a first data connection is assessed in light of the adaptations of one or more codings used for the first data connection, and, depending on the assessment, a second data connection is set up that is used as replacement for the first data connection. Radio terminals that act like data relay stations may serve here as data network nodes. It is essential for the method according to the present invention that the coding of data of a data connection be dynamic, that is, can be adapted as a function of the transmission quality of the data connection used. By the assessment of the adaptations of one or even several codings, in principle, a prediction may be made as to whether the first data connection will be interrupted in future.
Preferably, the assessment of the quality of the first data connection is made in that, in light of the codings used, at least one first data network node, participating in the first data connection, makes a statement as to whether the first data connection is becoming unstable. For instance, in a radio communications network, a first radio terminal, acting as a data network node, is able to evaluate the quality of the data connection to an adjacent second radio terminal by the coding method used for the data transmission between the two radio terminals. If the radio connection deteriorates, the transmitting terminal changes the coding for its partial connection. If this occurs several times in succession, it can be assumed that the data connection between these two radio terminals will possibly be interrupted in the future. Since the receiving terminal also detects these changes in light of the coding that the transmitting terminal is using, each of the two terminals is able to make the statement as to whether the data connection will become unstable.
In particular, in response to a statement that the data connection will become unstable, it is able to initiate the search for a second data connection. The search may be initiated, for example, by the emitting of a first message by the at least one first data network node that has made the statement that the first data connection is becoming unstable. In this context, the first message is sent to second data network nodes, for instance, to the data network nodes adjacent to the first data network node.
The message preferably includes the number of transmissions, the address of a transmitter and the address of a receiver of data of the first data connection. The number of transmissions (“time to life”: TTL) is understood to mean the number of partial transmissions for which data packets of the first data connection are able to be transmitted and maintained until they are discarded or become invalid.
Preferably, the message is sent with a different coding than the coding(s) used in the first data connection. Because of this, only those data network nodes which are able to receive and decode the differently encoded messages, are able to process the message. For example, data network nodes are not able to process the message if they are outside the reception range of the coding used. This procedure can effectively prevent data network nodes of the first data connection from becoming a part of the second data network connection.
In particular, the new coding is a higher coding than the coding(s) used in the first data connection, which means the message is not able to be transmitted over connections between data network nodes that have a bad transmission quality. This avoids routes that are not considered stable.
The second data connection is preferably utilized when the first data connection is disconnected. Alternatively, the second data connection may also already be utilized if a prespecified coding is used for the first data connection, for instance, an especially robust, that is, low coding at a very poor transmission quality.
According to one additional aspect, the present invention relates to a device for controlling data connections in a data network having a plurality of data network nodes, which are designed in such a way that the quality of a first data connection is assessed in light of the adaptations of one or more codings used for the first data connection, and, depending on the assessment, a second data connection is set up that is used as replacement for the first data connection.
Preferably, at least one first data network node participating in the first data connection is designed to assess the quality of the first data connection in that, in light of the codings used, it makes a statement as to whether the first data connection is becoming unstable. The at least one first data network node may, for example, be set up by program technology for carrying out the aforementioned assessment. If the at least one first data network node includes a radio interface, such as a plug-in card for a personal computer such as a PCI card or a PC card for a laptop computer or a USB radio adapter, its (their) operating program may be correspondingly set up. Radio interfaces that are already present, having an operating software in a programmable memory may be updated correspondingly.
Furthermore, the at least one first data network node may be designed so that, in response to a statement that the first data connection is becoming unstable, it initiates the search for the second data connection.
Moreover, the at least one first data network node may be designed to initiate the search by sending a first message to second data network nodes.
The message preferably includes the number of transmissions, the address of a transmitter and the address of a receiver of data of the first data connection.
In particular, the message has a different coding than the coding(s) used in the first data connection.
Typically, the different coding is a higher coding than the coding(s) used in the first data connection.
The device is preferably designed to switch over to the second data connection if the first data connection is disconnected.
Alternatively, or even additionally, the device may be designed to switch over to the second data connection if a prespecified coding is used for the first data connection.
Finally, the data network preferably includes a radio data network and at least two of the data network nodes are radio terminals. This means that a data connection in such a data network is able to include a radio transmission link whose transmission quality depends, among other things, on the motion of at least one of the terminals participating in the radio transmission link.
Further advantages and possible applications of the present invention may be derived from the following description in conjunction with the exemplary embodiments illustrated in the drawing, in which:
In the following, identical and functionally identical elements and method steps may be provided with the same reference symbols or designations.
The exemplary embodiments of the method according to the present invention, described below, for the dynamic adaptation of multihop data connections in data networks having a plurality of data network nodes are, in principle, independent of a predictive method for the behavior of a data connection. In all the exemplary embodiments, the assumption is made that a data network node, which is part of a data connection, monitors the quality of the data connection with respect to the coding method(s) used according to the present invention. This data network node divides the data connection into two parts: an input side part, which is the part of the data connection on the side of the source between either the source or a predecessor data network node and the data network node, and an output side part, which is the part of the data connection between the data network node and either the destination or a subsequent data network node in the direction towards the destination. Therefore, it is essentially able to recognize the three following different situations with respect to the input-side and the output-side part of the data connection:
1. The quality of the output-side data connection part deteriorates, and, accordingly, more stable coding methods are selected for the partial data connection. The conditions of the input-side data connection part remain constant. From this, the data network node is able to conclude that the destination or a data network node that belongs to the output-side part of the data connection is distancing itself from it.
2. Both on the input-side part and the output-side part of the data connection, the transmission quality is deteriorating, and therefore more stable coding methods are selected. This means that the data network node itself is moving.
3. The quality of the input-side part of the data connection is deteriorating, and a more stable coding method is selected. However, the quality of the output-side part of the data connection remains the same. From this, it may be concluded that the source or a data network node of the input-side part of the data connection is moving away from the data network node.
All the above-named situations are now able to be recognized and evaluated, as will be shown below in light of the explanation of various exemplary embodiments.
The first exemplary embodiment of the method according to the present invention is based on a well-timed data connection adaptation. This will also be designated below as early route rearrangement (ERRA). An appropriate ERRA protocol relates, in particular, to a method that changes the route or the data connection if it is predicted that an interruption of an existing data connection threatens. Therefore, the ERRA method is started before a data connection is disconnected, for instance, if a data network node of the data connection moves out of the radio range of its adjacent data network nodes. In addition, the ERRA protocol uses a special coding method for sending connection change messages, which are designated as ERRA_Request (ERRA_REQ), for short.
By way of example,
[See footnote]1
1 Translator's Note: From here on, 10 pages of the German document are marked RECTIFIED SHEET (RULE 91) ISA/EP. In this translation, this ends approximately 2½ pages before the claims. The reason for the marking is explained in the accompanying letter of Deutsche Telekom to the German Patent and Trademark Office.
At the beginning, a routing protocol such as, for instance, AODV, has set up a multihop data connection from source terminal 0 to destination terminal 14 (in
To do this, data network node 5 emits an ERRA_REQ message (denoted by {circle around (3)} in
Each data network node or each radio terminal that receives this message and knows a route to destination terminal 14, or which may even be the destination terminal itself, replies with an ERRA_REP message (in
The ERRA method provides that the ERRA_REQ message is more sensitive, i.e., is coded higher than the data packets transmitted between the two data network nodes 5 and 10. This prevents data network node 10 from being able to receive the ERRA_REQ message on a direct path, that is, by its old predecessor, and from replying using its current route, which, to be sure, is supposed to be changed in the future. This method leads to a second data connection which uses a higher coding, and which accordingly has a better channel quality than the first data connection (denoted in
As shown in
As an alternative to sending the ERRA_REQ message with a higher coding than the coding that is used between data network nodes 5 and 10 of the first data connection, data network node 5 may discard a direct reply from data network node 10, since it knows that data network node 10 is replying using the route that is to be replaced, or rather the first data connection.
After data network node 5 has received the ERRA_REP message, having an alternative route proposal or rather a second data connection, it is able to decide whether it wishes to use the alternative route or the second data connection, or whether the old route or the first data connection should continue to be used. The basis of this decision may, for example, be the number of data network nodes which participate in the second data connection or alternative route. If the new route or the second data connection includes more data network nodes or source terminals than the old route or the first data connection, data network node 5 or the initiator of the search for an alternative route may communicate this to source terminal 0, using a so-called ERRA_INFO message. Based on this message, source terminal 0 is able to decide whether the new route or second data connection should be accepted or whether a search should be made for an additional new route or third data connection as an alternative to the first data connection. As shown in
The exemplary embodiment of the method according to the present invention, clarified in light of
Again, it is assumed that, in a second data network having a source terminal 0 and a destination terminal 7 as well as data network nodes, at the beginning a first data connection exists between source terminal 0 and destination terminal 7, (denoted in
Data network node 3 moves away from its adjacent data network nodes 2 and 4, so that, in this region, the first data connection deteriorates, or, stated more precisely, its quality deteriorates (designated in
Now, the ERU method assumes that almost every data network node permanently checks its connection to adjacent data network nodes and constantly has a current list in which all its adjacent data network nodes are entered, that are located within radio reach. This table is also designated as a neighborhood table. If the functionality of the permanent monitoring of the connections to the adjacent data network nodes is not available, a data network node is able to search actively for adjacent data network nodes, using so-called ERU_HELLO messages.
The initiator of the ERU method, data network node 2, is now able to add data of its adjacent data network nodes, designated also as PATCH_INFO, to a normal data packet (designated as {circle around (3)} in
With the first transmission, the BHC is set to 1. Data network node 3 has meanwhile been informed by its own prediction unit about the quality of the data connection to adjacent data network nodes via the change in the qualities of the data connections input to it and output from it. Data network node 3 thereby detects that the data connections threaten to disconnect, increases the BHC by 1 and forwards the PATCH_INFO message, together with the data packet received from data network node 2, to data network node 4.
Data network node 4 recognizes the PATCH-INFO message and furthermore recognizes that only its incoming data connection, that is, the data connection between it and data network node 3, is changing. It may conclude from this that its position is remaining stable, but that data network node 3 is moving away. It separates the PATCH-INFO message from the received data packet, and sends an ERU_REQ message to all its adjacent data network nodes (denoted by {circle around (4)} in
In the situation shown in
If the initiating data network node 2 receives the ERU_REP message with an alternative route, it may decide whether it wishes to use it. If the DHC and ARHC are different, that is, the number of hops is changing, then initiating data network node 2 must inform source terminal 0 about it using a message ERU_INFO.
However, there is an exception to this. In the ERU_REQ message, the address of source terminal 0 is also included. If this address is known to a data network node, that is, if source terminal 0 is directly adjacent to this data network node, this data network node may reply directly to source terminal 0, which informs the initiating data network node by an ERU_INFO message. In light of the DHC, source terminal 0 is then able to recognize whether and how the number of hops has changed. Using the ERU method, the alternative data connection maximally reaches the length of the breaking point of the first data connection plus 2 hops. The alternative route is denoted by {circle around (6)} in
The third exemplary embodiment of the method according to the present invention shown in
Data network node 3 reacts to this by sending an OPERU_REQ message (denoted by {circle around (3)} in
At this point, several situations may arise:
1. At least one of the addressed data network nodes 2 and 4 replies with an OPERU_REP message to the OPERU_REQ message of moving data network node 3. In this reply message, the corresponding data network node 2 or 4 confirms that it knows the predecessor data network node and the successor data network node or destination terminal 5. An alternative route or a second data connection is thereby found as replacement for the first data connection.
2. At least two data network nodes reply to the OPERU_REQ message. One data network node replies that it knows the predecessor data network node, and one data network node replies that it knows the successor data network node or the destination terminal. At this point, an additional case differentiation may be carried out:
3. No data network node replies to the OPERU_REQ message, or data network nodes reply with adjacent data network nodes of the same side of the breaking point. In that case, the OPERU method fails.
If the predecessor data network node receives an OPERU_REP message with an alternative route, it is able to accept the route or continue to use the previous route. If the predecessor data network node accepts the new route, and if the number of hops of the data connection between source terminal and destination terminal thereby changes, the predecessor data node has to inform source terminal 0 about the new situation, using an OPERU_INFO message. Upon receiving the OPERU_INFO message, on its part, source terminal 0 decides whether it accepts the changes in the route, or whether, instead, it will initiate a new route search.
In summary, it should be mentioned that, of all the methods explained above, the OPERU method requires the least signaling. Therefore, the methods may also be used jointly, in a staggered manner. In this case, there is the opportunity of using the OPERU method as the first method. Within the OPERU method, the predecessor data network node of a breaking point also notices that, using the OPERU method, a second data connection or an alternative route is being searched for. The predecessor data network node then sets a timer OPERU_MAX_DURATION. If it receives no reply within the time period specified by the timer, it then triggers the ERU method. If this method also does not lead to success, the ERRA method is started as the last local possibility, since the ERRA method is able to make do even without a data connection.
If all local methods for searching for a new data connection fail, the source terminal is so informed. The latter will then begin, as provided in the standard method, a new route search, or it will notify the requesting service that, at the moment, no route to the destination terminal is able to be found.
Thus, using the present invention, in data networks having data connections that have a greatly fluctuating transmission quality, it is possible to switch to alternative data connections quickly and without great effort. Because of that, in spite of the changing quality of data connections, one may achieve a high data throughput and a low delay in data transmission, accompanied by a small signalizing expenditure. The present invention is especially suitable for use in multi-hop radio communications networks having mobile radio terminals as relay stations or data network nodes. In such a data network, the quality of an existing data connection can change very rapidly, so that, without using the present invention, only a small data throughput and, above all, great delays in transmission will occur. On the other hand, using the present invention, it is possible to achieve a relatively high data transmission rate and low delay times, especially in the transmission of data packets, with only little technical expenditure. This works out especially advantageously for end-to-end connections via which, for example, multimedia data are being transmitted.
| Number | Date | Country | Kind |
|---|---|---|---|
| 103 24 470.0 | May 2003 | DE | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/DE04/01037 | 5/17/2004 | WO | 6/7/2006 |
