Title of Invention	ARRANGEMENT AND METHOD FOR DELAYED REALLOCATION OF ADDRESSES
Abstract	The present invention relates to an arrangement, in a communications network, for allocating addresses to users or clients connecting to the network. It comprises at least one address holding means holding a plurality of addresses and allocating control means for, from said address holding means, allocating addresses to the users or clients, and for detecting when an address is released or when a user or a client allocated an address disconnects. It further comprises, for each pool of addresses, delay means for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that a minimum time period before reuse of each address can be guaranteed.

Title of Invention

ARRANGEMENT AND METHOD FOR DELAYED REALLOCATION OF ADDRESSES

Abstract

The present invention relates to an arrangement, in a communications network, for allocating addresses to users or clients connecting to the network. It comprises at least one address holding means holding a plurality of addresses and allocating control means for, from said address holding means, allocating addresses to the users or clients, and for detecting when an address is released or when a user or a client allocated an address disconnects. It further comprises, for each pool of addresses, delay means for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that a minimum time period before reuse of each address can be guaranteed.

Full Text	FIELD OF THE INVENTION The present invention relates to an arrangement for, in a communications network, allocating addresses to clients or users connecting to the network. The arrangement comprises at least one address holding means holding a pool, or a plurality, of addresses, and allocating control means for, from said pool, allocating addresses to users or clients, and detecting when an address is released or when a client or a user allocated an address disconnects or is disconnected. The invention also relates to a node in a mobile communications system which comprises an arrangement for allocating addresses to users or clients connecting to the network. The node comprises holding means for holding a plurality of addresses for allocation to users or clients. Still further the invention relates to a method for, in a communications network, providing allocation of addresses to users or clients. STATE OF THE ART In networks implementing dynamic addressing, allocation of addresses has to be used to provide users or clients connecting to the network with temporary addresses. Such addresses are allocated from a server in the network or from a peer. Several protocols have been standardized for handling such address allocation. One standardized protocol is the Dynamic Host Configuration Protocol (DHCP), cf. RFC (Request For Comments) 2131 3315. Another such protocol is Remote Authentication Dial in User Service (RADIUS), standardized in RFC 2865. However, in mobile communications system such as GPRS (GSM Packet Radio Service), CDMA 2000 (Code Division Multiple Access) etc. connections to the network occur frequently and the connections are typically only maintained for short periods. IPv4 or IPv6 (Internet Protocol version 4 and 6 respectively) addresses are used which uniquely identify the user and the addresses may be kept in multiple nodes, such as nodes in authentication systems, in billing systems, and application gateways and application servers, throughout the network. Among others due to the fact that the addresses are kept at a plurality of locations a quick assignment and release of addresses get problematic when the same addresses are reused. Currently addresses are mostly assigned in a circular fashion and a released address will only be reused when all other addresses have been assigned to users. However, with such a solution it can not be guaranteed that an address will not be reused immediately following upon release thereof. The number of addresses that are required is high, and to prevent addresses from being reused too quickly, still larger pools of addresses are required. It is also a problem that the allocation systems used today for allocation of the IPv4 or IPv6 addresses, cannot guarantee a delay before reuse of an address, which means that in fact addresses will be reused frequently or reused more or less immediately upon release, which makes it easier to predict end user IP addresses. Thus there is a problem, in such systems, that end user IP addresses e.g. addresses of mobiles, laptops etc. easily can be predicted and used for illegal purposes, which is serious. SUMMARY OF THE INVENTION It is therefore an object of the invention to provide an arrangement as initially referred to, through which prediction of allocated addresses is made more difficult. It is still further an object of the invention to provide an arrangement through which address allocation in dynamical addressing systems can be provided for in a safer and more trustworthy manner than hitherto. Still further it is an object to provide an arrangement through which allocation of addresses can be provided for in a simple and a flexible manner. Another object is to provide an arrangement through which address allocation can be provided for in a fast and easy manner while still guaranteeing that one and the same address is not reused immediately upon release. Furthermore it is an object to provide an arrangement through which the number addresses that need to be kept in a pool for allocation purposes can be reduced while still meeting the prevailing requirements as to security during address allocation. It is also an object to provide an arrangement through which a large number of addresses can be handled and allocated in a safe manner. It is also an object of the invention to provide a node in a communications network through which one or more of the above mentioned objects can be fulfilled. Further yet it is an object of the invention to suggest a method for allocating addresses in a safe and reliable manner, and at the same time in a fast manner and preferably without requiring any increase in the number of addresses required in a pool as compared to known systems. It is also an object of the invention to suggest a method through which one or more of the above mentioned objects can be fulfilled. Therefore an arrangement as initially referred to is provided which comprises, for each pool of addresses, or address holding means holding a pool of addresses, delay means for allowing control of the minimum time period to lapse before an address having been allocated to a user is again reused, i.e. again allocated to a user, such that a minimum time period before reuse of each address can be guaranteed. Said delay means particularly comprises a limited, given number of timers and most particularly it comprises one timer. In a preferred implementation the address holding means holds addresses in at least three different states depending on the use status of the addresses. Most particularly the address holding means holds addresses in four different states. A first state comprises a state for holding addresses which are available for allocation to a user, a second state is a state for holding addresses occupied by users, a third state is a state for holding addresses released by a user, but not yet available for allocation, and finally a fourth state is a state for holding addresses in a waiting state. Particularly said one timer is used for activation of a transfer of all addresses of the, or a, pool, which addresses are held in said third, also denoted released, state, to said fourth, also denoted waiting, state upon expiry a first time, and for initiating transfer of all addresses held in said fourth, waiting, state to said first, available, state upon expiry of the timer the subsequent time. The allocation of addresses in said first available state to users can be done in different manners. In one implementation the lowest address held in said first state is selected or allocated to a user requesting an address. In an alternative embodiment an arbitrary address held in said first state is selected and allocated to a user requesting an address. Any appropriate algorithm or selection method can be used. Particularly an address is kept unused during a time period which is equal to or longer than a controllably selected delay time and shorter than or equal to twice the set delay time. In one implementation the delay time is (re)configurable whereas in other embodiments it is fixed. In one particular embodiment one single delay means or timer is used for all addresses handled by the arrangement (if there is one or more pools), or one for each pool if an arrangement comprises more than one pool. The arrangement is particularly implemented in or comprised by or associated (in communication) with a node in a mobile communication system (e.g. supporting communication of packet data) . The mobile system may for example be a GPRS or a UMTS system or CDMA 2000 etc. In one implementation the node in or in association/ communication with which the arrangement is implemented comprises a GGSN (Gateway GPRS Support Node) or a CGSN (Combined GPRS Support Node). In another implementation the arrangement is comprised by or associated or in communication with a server, e.g. a Radius- AAA server or a DHCP server. Particularly the addresses in the pool are IPv4 or IPv6 addresses. The concept can, however, of course also be implemented to other addresses as well as the arrangement can be implemented in or in communication with other nodes or servers or peers in general. Therefore also a node in a mobile communications system is suggested, the node being of a kind as initially referred to, comprising or communicating with an arrangement which comprises delay means for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that lapse of a minimum time period before reuse of each address can be guaranteed. Particularly said delay means comprises one single timer. The holding means preferably holds addresses in four different states, which particularly comprises a first state for holding addresses available for allocation to a user, a second state for holding addresses occupied by users, a third state for holding addresses released by users but not yet available for allocation, and a fourth state for holding addresses in a waiting state. Particularly said one (single) timer is used for activation of transfer of all addresses of the/a pool in which addresses are held in said third, released, state to said fourth waiting state upon a first expiry, and for activation of a transfer of all addresses from said fourth, waiting, state to said first available, state upon the subsequent expiry. The delay time may particularly be configurable. Alternatively it is fixed. In one embodiment the node comprises a packet data support node in a communications system supporting communication of packet data. Particularly it comprises a UMTS/GPRS GGSN or a CGSN or a node with a similar functionality. More generally it may comprise any peer node in a network with a functionality of allocating addresses to users. In one particular implementation it comprises a node with a server, or a server, such as for example a DHCP or a Radius AAA-server, allocating addresses in real time. The invention, in order to solve one or more of the objects referred to above, also provides a method, for, in a communications network, providing for allocation of addresses to users or clients. The method comprises the steps of; providing addresses in address holding means comprising a pool in or associated with a node or a server in the network; keeping the addresses in four different states depending on use status of the respective addresses; transferring all addresses kept in a third state in which the addresses are released from being allocated to a user, upon expiry of a timer, to a fourth, waiting, state; transferring all the addresses in the fourth, waiting, state upon expiry of said timer the subsequent time to a first, available, state, in which said addresses are available for allocation to users, such that, for each respective address, a minimum delay time before reuse of said address, can be guaranteed. The method particularly comprises the steps of, when a user or a client requests an address, allocating the lowest address in the first, available, state, in which addresses are held and for which at least the guaranteed minimum delay time has lapsed. Particularly the method comprises allocation of addresses substantially in real time. The addresses are particularly IPv4 or IPv6 addresses. The network particularly comprises a mobile communications network. The method may be implemented in nodes in a communication system supporting communication of packet data, such as network nodes, for example GGSNs or CGSNs or in servers, such as Radius servers or DHCP servers. Particularly the method comprises the step of; configuring the delay time, i.e. the timer expiration time. BRIEF DESCRIPTION OF THE DRAWINGS The invention will in the following be further described, in a non-limiting manner, and with reference to the accompanying drawings, in which: Fig. 1 conceptually illustrates an arrangement according to the inventive concept, Fig. 2 schematically illustrates a block diagram wherein the arrangement according to the invention is implemented in an office environment, e.g. implemented in (for) a DHCP server, Fig. 3 is a schematical block diagram illustrating an implementation of the inventive concept wherein the arrangement is provided or associated with a Radius server connected to a GGSN, Fig. 4A is a schematical block diagram illustrating an arrangement according to the inventive concept associated with a GGSN, Fig. 4B is a schematical block diagram of an implementation in which the arrangement is provided in a CGSN, Fig. 5 schematically illustrates, with reference to a time axis, the variation in actual delay time, and Fig. 6 is a flow diagram illustrating one embodiment of the inventive concept. DETAILED DESCRIPTION OF THE INVENTION Fig. 1 schematically illustrates an allocating arrangement 100 according to the inventive concept with address holding means 10 comprising an address pool in which addresses are kept in four different states. The allocating arrangement 100 may, as referred to earlier in the application, be provided in a network node, such as for example in a GGSN or a CGSN, it may be associated with or in communication with such node or any other node handling address allocation, or it may be provided in server means handling or allocating addresses, or it may be in communication with server means allocating or handling addresses. In Fig. 1 the four different states are illustrated, namely a first or available state, 1, a second or occupied state, 2, a third or released state, 3, and a fourth or waiting state, 4. Software is provided for keeping, at least, information about the different addresses in the different states, and as compared to a known systems, an additional parameter has to be configured, namely a minimum time period before individual addresses should be allowed to be reused. According to the invention the released state and the waiting state are separated in order to be able to avoid the use a timer, or starting a timer, for every separate address being handled. Generally, for example in GGSNs or other nodes performing the same functionality as far as addressing handling is confirmed, even hundreds of thousands of addresses may be used. Starting and handling one timer per address would therefore not be feasable, both due to limitations as far as storing capacity is concerned as well as processor utilization is concerned. If, thus, a combined released and waiting state were to be used, this would mean that several times more storing capacity would be required per address. Thus, according to the inventive concept there is one single timer for each allocating arrangement or one timer per address holding means. It is in Fig. 1 supposed that addresses x0, X1, x3, x4, x5, Xi, X17 are in the first, available state meaning that they are all available for allocation to a network user. In order to explain the inventive concept, it is here supposed that a user requests an address and that, in any appropriate manner, an addresses is selected by the allocating arrangement, i.e. according to any criteria, arbitrarily or according to some algorithm. The inventive concept is not limited to any specific way of a selecting addresses for allocation from the available state 1. Thus, in this case it is supposed that address Xi is allocated. This handling is not illustrated in any figure, but merely the handling in the allocating arrangement since the address allocation itself is known and what is important for the inventive concept is how the addresses are handled in the allocating arrangement. Thus, since an address xi is selected and allocated, it is put in the second, occupied state 2, the addresses in which state are considered to be utilized in for example the GPRS network, in the GGSN and all connected, external networks. An address enters the second, occupied, state 2 at any time when it is allocated to a user. In this case it is supposed that addresses x2, x6, x7, x8, x11, xi for the moment are handled in the second, occupied state 2. It is subsequently supposed that the user to whom address xi has been allocated disconnects or is disconnected, xi is then released and transferred from the second, occupied, state 2 to the third, released state 3. The released state 3 is a state which is used to hold an address before it is transferred to the waiting state 4. Addresses are entered into the released state 2 at any time when a user disconnects or is disconnected. An address in the released state 2 is considered to be released, for example, in the GGSN, the GPRS network and all connected external networks. In the released state 2 the address will not be allocated. In this particular embodiment it is supposed that addresses x9, x10, x14, xi are in the third, released state. It is then supposed that the one or single timer expires. Expiry of the timer (as far as addresses in the third state are concerned) then initiates a substantially simultaneous transfer of all addresses in the third, released state, to the fourth, waiting state, i.e. addresses x9, x10, x11, Xi are all transferred at one and the same time. An address kept in the waiting state is considered to be released in the GGSN, the GPRS network and all connected external networks. Thus, when the timer expires once more, addresses x9, x10, x14, xi are all substantially simultaneously transferred to the first, available, state 1. It should be clear that this figure is shown as referring to different moments of time for the respective states and it is intended to illustrate the procedure when an address (xi) is transferred from one state to another. If for example two addresses are considered, address al used by mobile user A and address b1 used by a mobile user B, the procedure might e.g. be as follows: The user A with address al disconnects. The address al is then substantially immediately transferred to the released state. It is subsequently supposed that the single timer expires and address al is then transferred to the waiting state. Then user B with address bl disconnects and address bl enters the released state. Then the single timer expires again. Address bl then enters the waiting state whereas when the timer expires al enters the available state, i.e. it is the second time that the timer expires as far as address a1 is concerned, whereas it is the first time that it expires as far as address b1 is concerned. In both cases, of course, also all other addresses in the released state, enter the available state at the same time, whereas all the other addresses in the waiting state at the same time enter the available state upon expiry of the timer. The delay is the minimum time period that should lapse before an address is reused. The length of the delay could be configurable or fixed. Generally a long delay has the impact that addresses will be held longer and more addresses will be needed. If for example 100 mobile stations are connecting and requesting an address every second and 100 mobile stations are disconnecting and releasing an address every second and with, for example, a delay of 60 seconds before reuse, 60x100=6000 addresses will be requested before the timer expires and released state addresses are put in waiting state. That is however the best case, the actual moment in time at which an address is released has to be considered. If the single timer expires at 00:00, then an address released at 00:01 will actually be held for 59 seconds and then put into the waiting state for 60 seconds. To calculate the worst case, the delay will be doubled; then 60x2x100=12.000 addresses will be required. As referred to above, an arrangement according to the present invention can be implemented in a number of different ways. Fig. 2 illustrates an address allocating arrangement 1001 as implemented in an office network environment. Any computer, e.g. iMAC 200A, workstation 200B, server 200C, mobile user station 200D, and Laptop 200E are supposed to send a DHCP protocol defined request for an address. The request will arrive at the DHCP server which here uses the address allocation procedure as described above, the request will be answered and then the IP address is provided in the response to the request. It should be clear that the implementation of the inventive concept is not as advantageous in such an environment as it would be in an environment where there are many short connections, for example in wireless networks. However, for example for tracking purposes, addresses may be required to be held for a particular amount of time and therefore the inventive concept may also with advantage be implemented in such systems. Fig. 3 schematically illustrates the implementation of the inventive concept in a mobile communications system. A mobile station, MS1 210A, which here comprises a handheld computer or MS2 210B, which here comprises a mobile telephone, may connect to the radio base station 93, which communicates with a switching arrangement 92 which in turn sets up a connection between the mobile station MS1 or MS2 and, here, SGSN 91. When this has been done, in any conventional manner, it is supposed that the mobile station is attached but still not connected. It is supposed that a mobile station MS1 or MS2 initiates a connection for being connected to the network for data transfer, e.g. Internet browsing, sending of e-mail etc. The GGSN 90 (Gateway GPRS Support Node) will then be contacted by SGSN 91 to request access for example to Internet and optionally an IP address (IPv4 or IPv6) from the Radius server 1102. GGSN 90 then sends a request to the Radius server to get an IP address and the Radius server 1102 then launches the process of allocating an IP address using allocating arrangement 1002 according to the inventive concept including the guaranteed delay functionality. The Radius server 1102 thus, in this implementation, uses the inventive address allocation procedure to provide an address to the mobile station. The Radius server 1102 will then fetch an address in the time delay address allocating arrangement 1002 comprising an address pool 102 (address holding means). The fetched address is forwarded via GGSN 90, SGSN 91, switch 92 and base station 93 to the mobile station, MS1 or MS2. The allocated address may advantageously be shared/used in multimedia and charging nodes, (e.g. a multimedia server 111) in an optional, advantageous embodiment. In such embodiments, the multimedia and/or charging systems may be informed of the allocated IP address. Since such systems may hold on to the address after an address is released, the time delay guaranteed allocation procedure according to the inventive concept can with advantage be implemented. At a subsequent stage it is supposed that the mobile station disconnects from Internet or stops data transfer. Thus, the MS, having finished using data e.g. Internet or e-mail, forwards a request to disconnect, which will reach SGSN 91. SGSN 91 then requests GGSN 90 to release the IP address it previously had allocated. GGSN 90 then forwards a request to the Radius server 1102 to release the IP address. The Radius server informs the address allocating arrangement 1002 and then the procedure will be as described with the reference to Fig. 1 above. In the particularly illustrated embodiment it is supposed that the address allocating arrangement 1002 is associated with and in communication with the Radius server 1102. In an alternative implementation the address allocating arrangement could have been arranged in a node comprising the Radius server and the address allocation arrangement (and possibly other elements). 15 Fig. 4A shows an alternative implementation of the inventive concept in which the address allocating arrangement 1003 is associated with or communicates with GGSN 903 instead of with radio server 1103. Also in this embodiment it should be clear that the address allocating arrangement 1003 alternatively could have been provided in the GGSN node 93. In this embodiment also the allocation control means 1053 are illustrated; they could of course also be included in the other embodiments, but they can function in any known or appropriate manner, performing the actual allocation once an address has been provided according to the inventive concept and establishing when a user disconnects etc. The embodiment of Fig. 4A will in the following be explained. It is supposed that a mobile station MS1 220A or MS2 220B attaches to the network by first connecting to the base station 933. The base station 933 communicates with the switching arrangement 923, which in turns sets up a connection from the mobile station (MS1 or MS2) with the SGSN 913. As in the preceding embodiment, the mobile station at this stage is considered to be attached but not connected. In a subsequent step it is supposed that the mobile station MS1/MS2 initiates a connection for data, i.e. browsing the Internet, sending of e-mail etc. GGSN 903 will then be contacted by SGSN 913 to request for Internet access and optionally an IP address (IPv4 or IPv6). GGSN 903 will then be requested to provide an IP address and initiates the process of allocating an IP address according to the inventive concept by GGSN 903 requesting allocation of an address form the address allocating arrangement 1003, as described with reference to for example Fig. 1. The address will be forwarded via SGSN 913, switching arrangement 923, base station 933 to mobile station MS1 or MS2. As in the preceding embodiment the address allocated may be shared or used in multimedia and/or charging nodes. Subsequently, when the mobile station disconnects e.g. from Internet or stops data transfer, it will forward a request to disconnect, which reaches the SGSN 913 (it should be clear that the MS may also be disconnected by the network etc. It does of course not matter whether the disconnection is initiated from the MS side or from the network side, the important thing being that it is disconnected) . However, it is supposed that SGSN 913 requests that GGSN releases the previously allocated IP address and GGSN then sends a release request to the address allocating arrangement 1003 as discussed above, and the procedure will continue as discussed for example with reference to Fig. 1. Fig. 4B shows an implementation similar to that of Fig. 4A, but with the difference that the SGSN and the GGSN are combined in a GGSN (combined GPRS support node) including both the functionality of an SGSN and a GGSN. In this case it is also illustrated that the address allocating arrangement IOO4 with address pool 104 and the allocation control means 1054 is provided in the CGSN 904. Of course, the address allocating arrangement 1004 might alternatively be provided externally of the SGSN, cf. Fig. 4A, and it is here supposed that the GGSN 904 communicates with one or more servers, here merely schematically illustrated a server 1104, it being of no importance of the functioning of the inventive concept. In other respects the functioning is similar to that described above with reference to Figs. 1-4A. (In Figs. 2,3,4A the allocating arrangement might of course also be provided in the node itself.) Fig. 5 shows a relation between delay and number of addresses required or used. It is illustrated with reference to a time scale and it is supposed that there is one single timer T1. Thus, it is supposed that at 00:00:00 T1 expires. At t=00:00:00 it is supposed that address 1.1.1.1. is released. It then more or less instantaneously enters the state released. Somewhat later, t=00:00:59 it is supposed that address 2.2.2.2. is released and more or less instantaneously enters the state released. Then, when t=00:01:00, it is supposed that Tl expires (with respect to these addresses, the first time). Addresses 1.1.1.1. and 2.2.2.2. which both are in state released, then simultaneously are transferred to the waiting state. After lapse of a further expiration period for Tl, when t=00:02:00, Tl expires for the second time for the respective addresses. This initiates transfer of addresses 1.1.1.1. and 2.2.2.2. in the waiting state to the available state. Subsequently Tl expires at t=00:03:00, which however does not affect the addresses in question. This shows the relation between delay and the number of addresses used. In the "worst" case, when an address is released at the relative start of a cycle, the actual time before the address enters the available state is almost twice the delay time. Address 1.1.1.1. here illustrates the "worst" case when an address is released substantially at that the beginning of a cycle of the timer, whereas address 2.2.2.2. shows the "best" case where an address is released substantially at the end of the cycle of the timer. It should be clear that this figure is extremely simplified and it is merely included to illustrate the relationships and "worst" and "best" case respectively. Fig. 6 is a flow diagram schematically illustrating the procedure when a mobile station requests allocation of an address, and how said address is handled in the address allocating arrangement. Thus, it is supposed that MS request an address, 101. In this embodiment, it is supposed that e.g. a GGSN requests the address from the address allocating arrangement, 102. It should be clear that it does not have to be a GGSN, it might just as well be a CGSN or a server such as a Radius server or a DHCP etc, or more generally any peer/node handling allocation of addresses. The address allocating arrangement then selects an address, for example arbitrarily, among the addresses held in the first, available state, 103. Once selected, the address is forwarded, in this case, via GGSN to MS, 104A (via SGSN etc.). In the address allocating arrangement the selected address is transferred to state occupied, i.e. the second state, 104B. This is supposed to take place substantially simultaneously. It is also, supposed that MS remains connected for some time, 105A. In the meanwhile, the selected address is kept in state occupied, 104. However, after some time it is supposed that MS disconnects, 106A. The selected, previously allocated address, is then released and transferred to state released. Subsequently, when the single timer of the address allocating arrangement expires, the selected address is, together with all other addresses in state released, transferred to state waiting, 106B. Subsequently, when timer T1 expires the next time, the selected address is transferred to state available together with all other addresses in the waiting state, 106C. It should be clear that the invention of course not is limited to the specific, illustrated embodiments but that it can be varied in a number of ways within the scope of the appended claims. It is an advantage of the invention that addresses can be allocated substantially in real time at a high rate, and that for each address a delay time can be guaranteed during which the address is not reallocated. This may be particularly advantageous for IPv6 address allocation, since there will be even more addresses allocations when there is an almost unlimited amount of IPv6 addresses and therefore address allocation rates will presumably increase. It should also be clear that the inventive concept can be implemented in or by every node handling address allocation. It should also be clear that address allocation (selection) itself can be performed using any algorithm among the addresses in state available. Also in other aspects numerous variations are possible. It is an advantage of the invention that a delay time can be guaranteed for every address in any traffic scenario. It is also an advantage that fewer addresses will be required if circular address allocation schemes are implemented. WE CLAIM: 1. An arrangement, in a communications network, for allocating addresses to users or clients connecting to the network, comprising at least one address holding means holding a plurality of addresses and allocating control means for, from said address holding means, allocating addresses to the users or clients, and for detecting when an address is released or when a user or a client allocated an address disconnects, characterized in that the address holding means holds addresses in four different states depending on the use status of the addresses, said states comprising a first state for holding addresses available for allocation to a user, a second state for holding addresses occupied by users, a third state for holding addresses released by users, but not yet available for allocation, and a fourth state for holding addresses in a waiting state, and in that it comprises, for each address holding means with a pool of addresses, delay means comprising a single timer for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that a minimum time period before reuse of each address can be guaranteed, said one timer being used for activation of a transfer of all addresses held in said third (released) state to said fourth, waiting, state upon expiry a first time and for activation of a transfer of all addresses held in said fourth, waiting, state to said first, available, state upon expiry the subsequent time of the timer. 2. An arrangement according to any one of the preceding claims, wherein that the lowest address held in said first, available, state is selected for allocation to a user requesting an address. 3. An arrangement according to claim 1, wherein that an arbitrary address held in said first, available, state is selected and allocated to a user requesting an address. 4. An arrangement according to any one of the preceding claims, wherein that an address is kept unused during a time period which is equal to or longer than an controllably selected delay time and shorter than or equal to twice the set delay time. 5. An arrangement according to any one of the preceding claims, wherein that the delay time is configurable. 6. An arrangement according to any one of the preceding claims, wherein that one single delay means or timer is used for all addresses handled by the arrangement. 7. An arrangement according to any one of the preceding claims, wherein that the arrangement is implemented in, or comprised by, or associated with a node in a mobile communication system supporting communication of packet data. 8. An arrangement according to claim 7, wherein that the mobile system is a GPRS or a UMTS system. 9. An arrangement according to claim 8, wherein that the node comprises a GGSN or a CGSN. 10. An arrangement according to claim 7 or 8, wherein that the arrangement is comprised in or associated with a server, e.g. a Radius-AAA server or a DHCP server. 11. An arrangement according to any one of the preceding claims, wherein that the addresses are IPv4 or IPv6 addresses. 12. A node in a mobile communications system, comprising or being associated with an arrangement for allocating addresses to users/subscribers connecting to the network, comprising address holding means with a plurality or a pool of addresses for allocation to users or subscribers, wherein that it comprises delay means comprising one single timer, for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that lapse of a minimum time period before reuse of each address can be guaranteed, whereby the address holding means holds addresses in four different states, said states comprise a first state for holding addresses available for allocation to users, a second state for holding addresses occupied by users, a third state for holding addresses released by users, but not yet available for allocation, and a fourth state for holding addresses in a waiting state and said one timer is used for, at a first expiry, activation of a transfer of all addresses held in said third, released, state to said fourth, waiting, state and for, at the subsequent expiry, activation of a transfer of all addresses in the pool held in said fourth, waiting, state to said first, available, state. 13. A node according to claim 12, wherein that the delay time is configurable. 14. A node according to claim 12, wherein that the delay time of the timer is fixed. 15. A node according to any one of claims 12-14, wherein that it comprises a packet data support node in a communications system supporting communication of packet data. 16. A node according to claim 15, wherein that it comprises a UMTS/GPRS GGSN or a CGSN or a node with a similar functionality. 17. A node according to any one of claims 12-16, wherein that it comprises or in associated with a node with a server, e.g. a DHCP or a Radius AAA-server, allocating addresses in real time. 18. A method for, in a communications network, providing allocation of addresses to users/clients, connecting to the network, comprising at least one address holding means holding a plurality of addresses and allocating control means for, from said address holding means, allocating addresses to the users or clients, and for detecting when an address is released or when a user or a client allocated an address disconnects, where in that it comprises the steps of: providing a plurality of addresses in address holding means comprised by or associated with a node or a server in the network, keeping the addresses in four different states depending on use status of the respective addresses, transferring all addresses kept in a third state in which the addresses are released from being allocated to a user upon expiry of a timer to a fourth, waiting, state, transferring all the addresses in the fourth waiting state upon expiry of said timer the subsequent time to a first, available, state, in which said addresses are available for allocation to users, holding addresses occupied by users in a second state, such that, for each respective address, a minimum delay time before reuse of said address, can be guaranteed. 19. A method according to claim 18, wherein that it comprises the steps of, when a user or client requests an address: allocating the lowest address in the first, available, state, in which addresses are held and for which at least the guaranteed minimum delay time has lapsed. 20. A method according to claim 18 or 19, wherein that addresses are allocated substantially in real time. 21. A method according to any one of claims 18-20, wherein that the addresses are IPv4 or IPv6 addresses and that the network is a mobile communications network. 22. A method according to any one of claims 18-21, wherein that it is implemented in a node in a communication system supporting communication of packet data, e.g. a GGSN or a CGSN or a server, e.g. a Radius server or a DHCP server. 23. A method according to any one of claims 18-22, wherein that it comprises the steps of: configuring the delay time, i.e. the timer expiration time. ABSTRACT ARRANGEMENT AND METHOD FOR DELAYED RE-ALLOCATION OF ADDRESSES The present invention relates to an arrangement, in a communications network, for allocating addresses to users or clients connecting to the network. It comprises at least one address holding means holding a plurality of addresses and allocating control means for, from said address holding means, allocating addresses to the users or clients, and for detecting when an address is released or when a user or a client allocated an address disconnects. It further comprises, for each pool of addresses, delay means for allowing control of the minimum time period to lapse before an address having been allocated to a user is reused, i.e. again allocated to a user, such that a minimum time period before reuse of each address can be guaranteed.

Full Text

FIELD OF THE INVENTION
The present invention relates to an arrangement for, in a
communications network, allocating addresses to clients or
users connecting to the network. The arrangement comprises at
least one address holding means holding a pool, or a
plurality, of addresses, and allocating control means for,
from said pool, allocating addresses to users or clients, and
detecting when an address is released or when a client or a
user allocated an address disconnects or is disconnected. The
invention also relates to a node in a mobile communications
system which comprises an arrangement for allocating addresses
to users or clients connecting to the network. The node
comprises holding means for holding a plurality of addresses
for allocation to users or clients.
Still further the invention relates to a method for, in a
communications network, providing allocation of addresses to
users or clients.
STATE OF THE ART
In networks implementing dynamic addressing, allocation of
addresses has to be used to provide users or clients
connecting to the network with temporary addresses. Such
addresses are allocated from a server in the network or from a
peer. Several protocols have been standardized for handling
such address allocation. One standardized protocol is the
Dynamic Host Configuration Protocol (DHCP), cf. RFC (Request
For Comments) 2131 3315. Another such protocol is Remote
Authentication Dial in User Service (RADIUS), standardized in
RFC 2865.

However, in mobile communications system such as GPRS (GSM
Packet Radio Service), CDMA 2000 (Code Division Multiple
Access) etc. connections to the network occur frequently and
the connections are typically only maintained for short
periods. IPv4 or IPv6 (Internet Protocol version 4 and 6
respectively) addresses are used which uniquely identify the
user and the addresses may be kept in multiple nodes, such as
nodes in authentication systems, in billing systems, and
application gateways and application servers, throughout the
network. Among others due to the fact that the addresses are
kept at a plurality of locations a quick assignment and
release of addresses get problematic when the same addresses
are reused.
Currently addresses are mostly assigned in a circular fashion
and a released address will only be reused when all other
addresses have been assigned to users. However, with such a
solution it can not be guaranteed that an address will not be
reused immediately following upon release thereof. The number
of addresses that are required is high, and to prevent
addresses from being reused too quickly, still larger pools of
addresses are required. It is also a problem that the
allocation systems used today for allocation of the IPv4 or
IPv6 addresses, cannot guarantee a delay before reuse of an
address, which means that in fact addresses will be reused
frequently or reused more or less immediately upon release,
which makes it easier to predict end user IP addresses. Thus
there is a problem, in such systems, that end user IP
addresses e.g. addresses of mobiles, laptops etc. easily can
be predicted and used for illegal purposes, which is serious.
SUMMARY OF THE INVENTION

It is therefore an object of the invention to provide an
arrangement as initially referred to, through which prediction
of allocated addresses is made more difficult. It is still
further an object of the invention to provide an arrangement
through which address allocation in dynamical addressing
systems can be provided for in a safer and more trustworthy
manner than hitherto. Still further it is an object to provide
an arrangement through which allocation of addresses can be
provided for in a simple and a flexible manner. Another object
is to provide an arrangement through which address allocation
can be provided for in a fast and easy manner while still
guaranteeing that one and the same address is not reused
immediately upon release. Furthermore it is an object to
provide an arrangement through which the number addresses that
need to be kept in a pool for allocation purposes can be
reduced while still meeting the prevailing requirements as to
security during address allocation. It is also an object to
provide an arrangement through which a large number of
addresses can be handled and allocated in a safe manner.
It is also an object of the invention to provide a node in a
communications network through which one or more of the above
mentioned objects can be fulfilled.
Further yet it is an object of the invention to suggest a
method for allocating addresses in a safe and reliable manner,
and at the same time in a fast manner and preferably without
requiring any increase in the number of addresses required in
a pool as compared to known systems. It is also an object of
the invention to suggest a method through which one or more of
the above mentioned objects can be fulfilled.

Therefore an arrangement as initially referred to is provided
which comprises, for each pool of addresses, or address
holding means holding a pool of addresses, delay means for
allowing control of the minimum time period to lapse before an
address having been allocated to a user is again reused, i.e.
again allocated to a user, such that a minimum time period
before reuse of each address can be guaranteed. Said delay
means particularly comprises a limited, given number of timers
and most particularly it comprises one timer.
In a preferred implementation the address holding means holds
addresses in at least three different states depending on the
use status of the addresses. Most particularly the address
holding means holds addresses in four different states. A
first state comprises a state for holding addresses which are
available for allocation to a user, a second state is a state
for holding addresses occupied by users, a third state is a
state for holding addresses released by a user, but not yet
available for allocation, and finally a fourth state is a
state for holding addresses in a waiting state. Particularly
said one timer is used for activation of a transfer of all
addresses of the, or a, pool, which addresses are held in said
third, also denoted released, state, to said fourth, also
denoted waiting, state upon expiry a first time, and for
initiating transfer of all addresses held in said fourth,
waiting, state to said first, available, state upon expiry of
the timer the subsequent time. The allocation of addresses in
said first available state to users can be done in different
manners. In one implementation the lowest address held in said
first state is selected or allocated to a user requesting an
address. In an alternative embodiment an arbitrary address
held in said first state is selected and allocated to a user

requesting an address. Any appropriate algorithm or selection
method can be used.
Particularly an address is kept unused during a time period
which is equal to or longer than a controllably selected delay
time and shorter than or equal to twice the set delay time. In
one implementation the delay time is (re)configurable whereas
in other embodiments it is fixed. In one particular embodiment
one single delay means or timer is used for all addresses
handled by the arrangement (if there is one or more pools), or
one for each pool if an arrangement comprises more than one
pool.
The arrangement is particularly implemented in or comprised by
or associated (in communication) with a node in a mobile
communication system (e.g. supporting communication of packet
data) . The mobile system may for example be a GPRS or a UMTS
system or CDMA 2000 etc.
In one implementation the node in or in association/
communication with which the arrangement is implemented
comprises a GGSN (Gateway GPRS Support Node) or a CGSN
(Combined GPRS Support Node).
In another implementation the arrangement is comprised by or
associated or in communication with a server, e.g. a Radius-
AAA server or a DHCP server. Particularly the addresses in the
pool are IPv4 or IPv6 addresses. The concept can, however, of
course also be implemented to other addresses as well as the
arrangement can be implemented in or in communication with
other nodes or servers or peers in general.

Therefore also a node in a mobile communications system is
suggested, the node being of a kind as initially referred to,
comprising or communicating with an arrangement which
comprises delay means for allowing control of the minimum time
period to lapse before an address having been allocated to a
user is reused, i.e. again allocated to a user, such that
lapse of a minimum time period before reuse of each address
can be guaranteed. Particularly said delay means comprises one
single timer. The holding means preferably holds addresses in
four different states, which particularly comprises a first
state for holding addresses available for allocation to a
user, a second state for holding addresses occupied by users,
a third state for holding addresses released by users but not
yet available for allocation, and a fourth state for holding
addresses in a waiting state. Particularly said one (single)
timer is used for activation of transfer of all addresses of
the/a pool in which addresses are held in said third,
released, state to said fourth waiting state upon a first
expiry, and for activation of a transfer of all addresses from
said fourth, waiting, state to said first available, state
upon the subsequent expiry. The delay time may particularly be
configurable. Alternatively it is fixed.
In one embodiment the node comprises a packet data support
node in a communications system supporting communication of
packet data. Particularly it comprises a UMTS/GPRS GGSN or a
CGSN or a node with a similar functionality. More generally it
may comprise any peer node in a network with a functionality
of allocating addresses to users. In one particular
implementation it comprises a node with a server, or a server,
such as for example a DHCP or a Radius AAA-server, allocating
addresses in real time.

The invention, in order to solve one or more of the objects
referred to above, also provides a method, for, in a
communications network, providing for allocation of addresses
to users or clients. The method comprises the steps of;
providing addresses in address holding means comprising a pool
in or associated with a node or a server in the network;
keeping the addresses in four different states depending on
use status of the respective addresses; transferring all
addresses kept in a third state in which the addresses are
released from being allocated to a user, upon expiry of a
timer, to a fourth, waiting, state; transferring all the
addresses in the fourth, waiting, state upon expiry of said
timer the subsequent time to a first, available, state, in
which said addresses are available for allocation to users,
such that, for each respective address, a minimum delay time
before reuse of said address, can be guaranteed.
The method particularly comprises the steps of, when a user or
a client requests an address, allocating the lowest address in
the first, available, state, in which addresses are held and
for which at least the guaranteed minimum delay time has
lapsed. Particularly the method comprises allocation of
addresses substantially in real time. The addresses are
particularly IPv4 or IPv6 addresses. The network particularly
comprises a mobile communications network. The method may be
implemented in nodes in a communication system supporting
communication of packet data, such as network nodes, for
example GGSNs or CGSNs or in servers, such as Radius servers
or DHCP servers.
Particularly the method comprises the step of; configuring the
delay time, i.e. the timer expiration time.

BRIEF DESCRIPTION OF THE DRAWINGS
The invention will in the following be further described, in a
non-limiting manner, and with reference to the accompanying
drawings, in which:
Fig. 1 conceptually illustrates an arrangement according to
the inventive concept,
Fig. 2 schematically illustrates a block diagram wherein the
arrangement according to the invention is implemented
in an office environment, e.g. implemented in (for) a
DHCP server,
Fig. 3 is a schematical block diagram illustrating an
implementation of the inventive concept wherein the
arrangement is provided or associated with a Radius
server connected to a GGSN,
Fig. 4A is a schematical block diagram illustrating an
arrangement according to the inventive concept
associated with a GGSN,
Fig. 4B is a schematical block diagram of an implementation
in which the arrangement is provided in a CGSN,
Fig. 5 schematically illustrates, with reference to a time
axis, the variation in actual delay time, and
Fig. 6 is a flow diagram illustrating one embodiment of the
inventive concept.

DETAILED DESCRIPTION OF THE INVENTION
Fig. 1 schematically illustrates an allocating arrangement 100
according to the inventive concept with address holding means
10 comprising an address pool in which addresses are kept in
four different states.
The allocating arrangement 100 may, as referred to earlier in
the application, be provided in a network node, such as for
example in a GGSN or a CGSN, it may be associated with or in
communication with such node or any other node handling
address allocation, or it may be provided in server means
handling or allocating addresses, or it may be in
communication with server means allocating or handling
addresses. In Fig. 1 the four different states are
illustrated, namely a first or available state, 1, a second or
occupied state, 2, a third or released state, 3, and a fourth
or waiting state, 4. Software is provided for keeping, at
least, information about the different addresses in the
different states, and as compared to a known systems, an
additional parameter has to be configured, namely a minimum
time period before individual addresses should be allowed to
be reused. According to the invention the released state and
the waiting state are separated in order to be able to avoid
the use a timer, or starting a timer, for every separate
address being handled. Generally, for example in GGSNs or
other nodes performing the same functionality as far as
addressing handling is confirmed, even hundreds of thousands
of addresses may be used. Starting and handling one timer per
address would therefore not be feasable, both due to
limitations as far as storing capacity is concerned as well as
processor utilization is concerned. If, thus, a combined
released and waiting state were to be used, this would mean

that several times more storing capacity would be required per
address. Thus, according to the inventive concept there is one
single timer for each allocating arrangement or one timer per
address holding means.
It is in Fig. 1 supposed that addresses x0, X1, x3, x4, x5, Xi,
X17 are in the first, available state meaning that they are all
available for allocation to a network user. In order to
explain the inventive concept, it is here supposed that a user
requests an address and that, in any appropriate manner, an
addresses is selected by the allocating arrangement, i.e.
according to any criteria, arbitrarily or according to some
algorithm. The inventive concept is not limited to any
specific way of a selecting addresses for allocation from the
available state 1. Thus, in this case it is supposed that
address Xi is allocated. This handling is not illustrated in
any figure, but merely the handling in the allocating
arrangement since the address allocation itself is known and
what is important for the inventive concept is how the
addresses are handled in the allocating arrangement. Thus,
since an address xi is selected and allocated, it is put in the
second, occupied state 2, the addresses in which state are
considered to be utilized in for example the GPRS network, in
the GGSN and all connected, external networks. An address
enters the second, occupied, state 2 at any time when it is
allocated to a user. In this case it is supposed that
addresses x2, x6, x7, x8, x11, xi for the moment are handled in
the second, occupied state 2. It is subsequently supposed that
the user to whom address xi has been allocated disconnects or
is disconnected, xi is then released and transferred from the
second, occupied, state 2 to the third, released state 3. The
released state 3 is a state which is used to hold an address
before it is transferred to the waiting state 4. Addresses are

entered into the released state 2 at any time when a user
disconnects or is disconnected. An address in the released
state 2 is considered to be released, for example, in the
GGSN, the GPRS network and all connected external networks. In
the released state 2 the address will not be allocated. In
this particular embodiment it is supposed that addresses x9,
x10, x14, xi are in the third, released state. It is then
supposed that the one or single timer expires. Expiry of the
timer (as far as addresses in the third state are concerned)
then initiates a substantially simultaneous transfer of all
addresses in the third, released state, to the fourth, waiting
state, i.e. addresses x9, x10, x11, Xi are all transferred at one
and the same time. An address kept in the waiting state is
considered to be released in the GGSN, the GPRS network and
all connected external networks. Thus, when the timer expires
once more, addresses x9, x10, x14, xi are all substantially
simultaneously transferred to the first, available, state 1.
It should be clear that this figure is shown as referring to
different moments of time for the respective states and it is
intended to illustrate the procedure when an address (xi) is
transferred from one state to another.
If for example two addresses are considered, address al used
by mobile user A and address b1 used by a mobile user B, the
procedure might e.g. be as follows: The user A with address al
disconnects. The address al is then substantially immediately
transferred to the released state. It is subsequently supposed
that the single timer expires and address al is then
transferred to the waiting state. Then user B with address bl
disconnects and address bl enters the released state. Then the
single timer expires again. Address bl then enters the waiting
state whereas when the timer expires al enters the available
state, i.e. it is the second time that the timer expires as

far as address a1 is concerned, whereas it is the first time
that it expires as far as address b1 is concerned. In both
cases, of course, also all other addresses in the released
state, enter the available state at the same time, whereas all
the other addresses in the waiting state at the same time
enter the available state upon expiry of the timer.
The delay is the minimum time period that should lapse before
an address is reused. The length of the delay could be
configurable or fixed. Generally a long delay has the impact
that addresses will be held longer and more addresses will be
needed. If for example 100 mobile stations are connecting and
requesting an address every second and 100 mobile stations are
disconnecting and releasing an address every second and with,
for example, a delay of 60 seconds before reuse, 60x100=6000
addresses will be requested before the timer expires and
released state addresses are put in waiting state. That is
however the best case, the actual moment in time at which an
address is released has to be considered. If the single timer
expires at 00:00, then an address released at 00:01 will
actually be held for 59 seconds and then put into the waiting
state for 60 seconds. To calculate the worst case, the delay
will be doubled; then 60x2x100=12.000 addresses will be
required.
As referred to above, an arrangement according to the present
invention can be implemented in a number of different ways.
Fig. 2 illustrates an address allocating arrangement 1001 as
implemented in an office network environment. Any computer,
e.g. iMAC 200A, workstation 200B, server 200C, mobile user
station 200D, and Laptop 200E are supposed to send a DHCP
protocol defined request for an address. The request will
arrive at the DHCP server which here uses the address

allocation procedure as described above, the request will be
answered and then the IP address is provided in the response
to the request. It should be clear that the implementation of
the inventive concept is not as advantageous in such an
environment as it would be in an environment where there are
many short connections, for example in wireless networks.
However, for example for tracking purposes, addresses may be
required to be held for a particular amount of time and
therefore the inventive concept may also with advantage be
implemented in such systems.
Fig. 3 schematically illustrates the implementation of the
inventive concept in a mobile communications system. A mobile
station, MS1 210A, which here comprises a handheld computer or
MS2 210B, which here comprises a mobile telephone, may connect
to the radio base station 93, which communicates with a
switching arrangement 92 which in turn sets up a connection
between the mobile station MS1 or MS2 and, here, SGSN 91. When
this has been done, in any conventional manner, it is supposed
that the mobile station is attached but still not connected.
It is supposed that a mobile station MS1 or MS2 initiates a
connection for being connected to the network for data
transfer, e.g. Internet browsing, sending of e-mail etc. The
GGSN 90 (Gateway GPRS Support Node) will then be contacted by
SGSN 91 to request access for example to Internet and
optionally an IP address (IPv4 or IPv6) from the Radius server
1102. GGSN 90 then sends a request to the Radius server to get
an IP address and the Radius server 1102 then launches the
process of allocating an IP address using allocating
arrangement 1002 according to the inventive concept including
the guaranteed delay functionality. The Radius server 1102
thus, in this implementation, uses the inventive address

allocation procedure to provide an address to the mobile
station. The Radius server 1102 will then fetch an address in
the time delay address allocating arrangement 1002 comprising
an address pool 102 (address holding means). The fetched
address is forwarded via GGSN 90, SGSN 91, switch 92 and base
station 93 to the mobile station, MS1 or MS2. The allocated
address may advantageously be shared/used in multimedia and
charging nodes, (e.g. a multimedia server 111) in an optional,
advantageous embodiment. In such embodiments, the multimedia
and/or charging systems may be informed of the allocated IP
address. Since such systems may hold on to the address after
an address is released, the time delay guaranteed allocation
procedure according to the inventive concept can with
advantage be implemented.
At a subsequent stage it is supposed that the mobile station
disconnects from Internet or stops data transfer. Thus, the
MS, having finished using data e.g. Internet or e-mail,
forwards a request to disconnect, which will reach SGSN 91.
SGSN 91 then requests GGSN 90 to release the IP address it
previously had allocated. GGSN 90 then forwards a request to
the Radius server 1102 to release the IP address. The Radius
server informs the address allocating arrangement 1002 and then
the procedure will be as described with the reference to Fig.
1 above. In the particularly illustrated embodiment it is
supposed that the address allocating arrangement 1002 is
associated with and in communication with the Radius server
1102. In an alternative implementation the address allocating
arrangement could have been arranged in a node comprising the
Radius server and the address allocation arrangement (and
possibly other elements).
15

Fig. 4A shows an alternative implementation of the inventive
concept in which the address allocating arrangement 1003 is
associated with or communicates with GGSN 903 instead of with
radio server 1103. Also in this embodiment it should be clear
that the address allocating arrangement 1003 alternatively
could have been provided in the GGSN node 93. In this
embodiment also the allocation control means 1053 are
illustrated; they could of course also be included in the
other embodiments, but they can function in any known or
appropriate manner, performing the actual allocation once an
address has been provided according to the inventive concept
and establishing when a user disconnects etc.
The embodiment of Fig. 4A will in the following be explained.
It is supposed that a mobile station MS1 220A or MS2 220B
attaches to the network by first connecting to the base
station 933. The base station 933 communicates with the
switching arrangement 923, which in turns sets up a connection
from the mobile station (MS1 or MS2) with the SGSN 913. As in
the preceding embodiment, the mobile station at this stage is
considered to be attached but not connected.
In a subsequent step it is supposed that the mobile station
MS1/MS2 initiates a connection for data, i.e. browsing the
Internet, sending of e-mail etc. GGSN 903 will then be
contacted by SGSN 913 to request for Internet access and
optionally an IP address (IPv4 or IPv6). GGSN 903 will then be
requested to provide an IP address and initiates the process
of allocating an IP address according to the inventive concept
by GGSN 903 requesting allocation of an address form the
address allocating arrangement 1003, as described with
reference to for example Fig. 1. The address will be forwarded

via SGSN 913, switching arrangement 923, base station 933 to
mobile station MS1 or MS2.
As in the preceding embodiment the address allocated may be
shared or used in multimedia and/or charging nodes.
Subsequently, when the mobile station disconnects e.g. from
Internet or stops data transfer, it will forward a request to
disconnect, which reaches the SGSN 913 (it should be clear that
the MS may also be disconnected by the network etc. It does of
course not matter whether the disconnection is initiated from
the MS side or from the network side, the important thing
being that it is disconnected) . However, it is supposed that
SGSN 913 requests that GGSN releases the previously allocated
IP address and GGSN then sends a release request to the
address allocating arrangement 1003 as discussed above, and the
procedure will continue as discussed for example with
reference to Fig. 1.
Fig. 4B shows an implementation similar to that of Fig. 4A,
but with the difference that the SGSN and the GGSN are
combined in a GGSN (combined GPRS support node) including both
the functionality of an SGSN and a GGSN. In this case it is
also illustrated that the address allocating arrangement IOO4
with address pool 104 and the allocation control means 1054 is
provided in the CGSN 904. Of course, the address allocating
arrangement 1004 might alternatively be provided externally of
the SGSN, cf. Fig. 4A, and it is here supposed that the GGSN
904 communicates with one or more servers, here merely
schematically illustrated a server 1104, it being of no
importance of the functioning of the inventive concept. In
other respects the functioning is similar to that described
above with reference to Figs. 1-4A. (In Figs. 2,3,4A the

allocating arrangement might of course also be provided in the
node itself.)
Fig. 5 shows a relation between delay and number of addresses
required or used. It is illustrated with reference to a time
scale and it is supposed that there is one single timer T1.
Thus, it is supposed that at 00:00:00 T1 expires. At
t=00:00:00 it is supposed that address 1.1.1.1. is released.
It then more or less instantaneously enters the state
released. Somewhat later, t=00:00:59 it is supposed that
address 2.2.2.2. is released and more or less instantaneously
enters the state released. Then, when t=00:01:00, it is
supposed that Tl expires (with respect to these addresses, the
first time). Addresses 1.1.1.1. and 2.2.2.2. which both are in
state released, then simultaneously are transferred to the
waiting state. After lapse of a further expiration period for
Tl, when t=00:02:00, Tl expires for the second time for the
respective addresses. This initiates transfer of addresses
1.1.1.1. and 2.2.2.2. in the waiting state to the available
state. Subsequently Tl expires at t=00:03:00, which however
does not affect the addresses in question.
This shows the relation between delay and the number of
addresses used. In the "worst" case, when an address is
released at the relative start of a cycle, the actual time
before the address enters the available state is almost twice
the delay time. Address 1.1.1.1. here illustrates the "worst"
case when an address is released substantially at that the
beginning of a cycle of the timer, whereas address 2.2.2.2.
shows the "best" case where an address is released
substantially at the end of the cycle of the timer. It should
be clear that this figure is extremely simplified and it is

merely included to illustrate the relationships and "worst"
and "best" case respectively.
Fig. 6 is a flow diagram schematically illustrating the
procedure when a mobile station requests allocation of an
address, and how said address is handled in the address
allocating arrangement. Thus, it is supposed that MS request
an address, 101. In this embodiment, it is supposed that e.g.
a GGSN requests the address from the address allocating
arrangement, 102. It should be clear that it does not have to
be a GGSN, it might just as well be a CGSN or a server such as
a Radius server or a DHCP etc, or more generally any peer/node
handling allocation of addresses. The address allocating
arrangement then selects an address, for example arbitrarily,
among the addresses held in the first, available state, 103.
Once selected, the address is forwarded, in this case, via
GGSN to MS, 104A (via SGSN etc.). In the address allocating
arrangement the selected address is transferred to state
occupied, i.e. the second state, 104B. This is supposed to
take place substantially simultaneously. It is also, supposed
that MS remains connected for some time, 105A. In the
meanwhile, the selected address is kept in state occupied,
104. However, after some time it is supposed that MS
disconnects, 106A. The selected, previously allocated address,
is then released and transferred to state released.
Subsequently, when the single timer of the address allocating
arrangement expires, the selected address is, together with
all other addresses in state released, transferred to state
waiting, 106B. Subsequently, when timer T1 expires the next
time, the selected address is transferred to state available
together with all other addresses in the waiting state, 106C.

It should be clear that the invention of course not is limited
to the specific, illustrated embodiments but that it can be
varied in a number of ways within the scope of the appended
claims. It is an advantage of the invention that addresses can
be allocated substantially in real time at a high rate, and
that for each address a delay time can be guaranteed during
which the address is not reallocated. This may be particularly
advantageous for IPv6 address allocation, since there will be
even more addresses allocations when there is an almost
unlimited amount of IPv6 addresses and therefore address
allocation rates will presumably increase. It should also be
clear that the inventive concept can be implemented in or by
every node handling address allocation. It should also be
clear that address allocation (selection) itself can be
performed using any algorithm among the addresses in state
available. Also in other aspects numerous variations are
possible.
It is an advantage of the invention that a delay time can be
guaranteed for every address in any traffic scenario. It is
also an advantage that fewer addresses will be required if
circular address allocation schemes are implemented.

WE CLAIM:
1. An arrangement, in a communications network, for allocating addresses to users or
clients connecting to the network, comprising at least one address holding means
holding a plurality of addresses and allocating control means for, from said address
holding means, allocating addresses to the users or clients, and for detecting when an
address is released or when a user or a client allocated an address disconnects,
characterized in
that the address holding means holds addresses in four different states depending on
the use status of the addresses, said states comprising a first state for holding
addresses available for allocation to a user, a second state for holding addresses
occupied by users, a third state for holding addresses released by users, but not yet
available for allocation, and a fourth state for holding addresses in a waiting state, and
in that it comprises, for each address holding means with a pool of addresses, delay
means comprising a single timer for allowing control of the minimum time period to
lapse before an address having been allocated to a user is reused, i.e. again allocated
to a user, such that a minimum time period before reuse of each address can be
guaranteed, said one timer being used for activation of a transfer of all addresses held
in said third (released) state to said fourth, waiting, state upon expiry a first time and
for activation of a transfer of all addresses held in said fourth, waiting, state to said
first, available, state upon expiry the subsequent time of the timer.
2. An arrangement according to any one of the preceding claims, wherein that the lowest
address held in said first, available, state is selected for allocation to a user requesting
an address.
3. An arrangement according to claim 1, wherein that an arbitrary address held in said
first, available, state is selected and allocated to a user requesting an address.

4. An arrangement according to any one of the preceding claims, wherein that an address
is kept unused during a time period which is equal to or longer than an controllably
selected delay time and shorter than or equal to twice the set delay time.
5. An arrangement according to any one of the preceding claims, wherein that the delay
time is configurable.
6. An arrangement according to any one of the preceding claims, wherein that one single
delay means or timer is used for all addresses handled by the arrangement.
7. An arrangement according to any one of the preceding claims, wherein that the
arrangement is implemented in, or comprised by, or associated with a node in a
mobile communication system supporting communication of packet data.
8. An arrangement according to claim 7, wherein that the mobile system is a GPRS or a
UMTS system.
9. An arrangement according to claim 8, wherein that the node comprises a GGSN or a
CGSN.
10. An arrangement according to claim 7 or 8, wherein that the arrangement is comprised
in or associated with a server, e.g. a Radius-AAA server or a DHCP server.
11. An arrangement according to any one of the preceding claims, wherein that the
addresses are IPv4 or IPv6 addresses.
12. A node in a mobile communications system, comprising or being associated with an
arrangement for allocating addresses to users/subscribers connecting to the network,
comprising address holding means with a plurality or a pool of addresses for
allocation to users or subscribers, wherein that it comprises delay means comprising

one single timer, for allowing control of the minimum time period to lapse before an
address having been allocated to a user is reused, i.e. again allocated to a user, such
that lapse of a minimum time period before reuse of each address can be guaranteed,
whereby the address holding means holds addresses in four different states, said states
comprise a first state for holding addresses available for allocation to users, a second
state for holding addresses occupied by users, a third state for holding addresses
released by users, but not yet available for allocation, and a fourth state for holding
addresses in a waiting state and said one timer is used for, at a first expiry, activation
of a transfer of all addresses held in said third, released, state to said fourth, waiting,
state and for, at the subsequent expiry, activation of a transfer of all addresses in the
pool held in said fourth, waiting, state to said first, available, state.
13. A node according to claim 12, wherein that the delay time is configurable.
14. A node according to claim 12, wherein that the delay time of the timer is fixed.
15. A node according to any one of claims 12-14, wherein that it comprises a packet data
support node in a communications system supporting communication of packet data.
16. A node according to claim 15, wherein that it comprises a UMTS/GPRS GGSN or a
CGSN or a node with a similar functionality.
17. A node according to any one of claims 12-16, wherein that it comprises or in
associated with a node with a server, e.g. a DHCP or a Radius AAA-server, allocating
addresses in real time.
18. A method for, in a communications network, providing allocation of addresses to
users/clients, connecting to the network, comprising at least one address holding
means holding a plurality of addresses and allocating control means for, from said
address holding means, allocating addresses to the users or clients, and for detecting
when an address is released or when a user or a client allocated an address
disconnects,
where in

that it comprises the steps of:
providing a plurality of addresses in address holding
means comprised by or associated with a node or a server in the network,
keeping the addresses in four different states
depending on use status of the respective addresses,
transferring all addresses kept in a third state in
which the addresses are released from being allocated to a user upon expiry of a
timer to a fourth, waiting, state,
transferring all the addresses in the fourth waiting
state upon expiry of said timer the subsequent time to
a first, available, state, in which said addresses are
available for allocation to users,
holding addresses occupied by users in a second state,
such that, for each respective address, a minimum delay time before reuse of said
address, can be guaranteed.
19. A method according to claim 18, wherein that it comprises the steps of, when a user or
client requests an address:
allocating the lowest address in the first,
available, state, in which addresses are held and for which at least the guaranteed
minimum delay time has lapsed.
20. A method according to claim 18 or 19, wherein that addresses are allocated
substantially in real time.
21. A method according to any one of claims 18-20, wherein that the addresses are IPv4
or IPv6 addresses and that the network is a mobile communications network.
22. A method according to any one of claims 18-21, wherein that it is implemented in a
node in a communication system supporting communication of packet data, e.g. a
GGSN or a CGSN or a server, e.g. a Radius server or a DHCP server.

23. A method according to any one of claims 18-22, wherein
that it comprises the steps of:
configuring the delay time, i.e. the timer expiration time.

ABSTRACT

ARRANGEMENT AND METHOD FOR DELAYED
RE-ALLOCATION OF ADDRESSES
The present invention relates to an arrangement, in a
communications network, for allocating addresses to users or
clients connecting to the network. It comprises at least one
address holding means holding a plurality of addresses and
allocating control means for, from said address holding means,
allocating addresses to the users or clients, and for
detecting when an address is released or when a user or a
client allocated an address disconnects. It further comprises,
for each pool of addresses, delay means for allowing control
of the minimum time period to lapse before an address having
been allocated to a user is reused, i.e. again allocated to a
user, such that a minimum time period before reuse of each
address can be guaranteed.

ARRANGEMENT AND METHOD FOR DELAYED REALLOCATION OF ADDRESSES

Documents:

Inventors:

PCT Conventions: