GB2433175A

GB2433175A - Wireless network with contended access control

Info

Publication number: GB2433175A
Application number: GB0524870A
Authority: GB
Inventors: Zhong Fan
Original assignee: Toshiba Research Europe Ltd
Current assignee: Toshiba Europe Ltd
Priority date: 2005-12-06
Filing date: 2005-12-06
Publication date: 2007-06-13
Anticipated expiration: 2025-12-06
Also published as: GB2433175B; GB0524870D0

Abstract

Access control apparatus is described for use in a communications network. The network is of a type, such as IEEE802.11e, configured to manage traffic dependent upon the nature of communication. The nature of communication is managed in classes, each class of communication having associated therewith at least one communications management criterion. The apparatus comprises storage means storing at least one communications management criterion for configuring management of communication in the network, for each of a plurality of sub-ranges of a range of number of actively communicating devices in the network, active device measurement means for determining a measure for the active devices in said network, sub-range selection means for determining, on the basis of the measure, the sub-range within which said measure belongs, data retrieval means for retrieving said at least one communications management criterion for said sub-range, and communications configuration means for configuring communications in said network in accordance with said at least one communications management criterion.

Description

I

This relates to control of communications in a wireless communications network and is particularly concerned with a wireless communications network with contended access control.

The IEEE8O2. lie amendment to the IEEE8O2. 11 Standard provides description of technology to enable access to a wireless communications medium to be managed by means of a contention window, based upon differentiation of data communication traffic into classes. Traffic is classified into classes by reference to likely structure and size of the data, and time dependence attached to receipt.

The management of traffic through such a system is thus managed with reference to traffic class. This is achieved by allocating different contention window (CW) sizes and arbitration inter-frame space (AIFS) values to each traffic class. This gives rise to Quality of Service (QoS) differentiation by traffic class. This is known as EDCA and is used in both IEEE8O2.l le and in MBOA MAC for UWB networking arrangements.

The difficulty with this arrangement, as presented in the abovementioned Standards, is that the system designer is given little or no guidance on the appropriate setting of the size of contention windows. It is clear that it would be prudent to set larger contention windows for traffic classes that can tolerate smaller throughput and larger delays.

However, the lack of a method for doing so means that a system cannot offer a guarantee that the expected throughput ratio will be obtained, nor can it achieve the overall maximum system throughput.

This invention aims to address this issue by choosing CW values such that the network achieves maximum system throughput while maintaining throughput ratios among traffic classes.

"Provisioning quality controlled medium access in ultra wideband (UWB) WPANs" (C. Hu et a!.; Tecimical report, Department of ECE, UIUC, 2005) describes an optimization framework for QoS provisioning in UWB WPANs. However, the described arrangement adopts a completely different model and hence obtain different results. Further, the dynamic contention window adaptation rule employed in that arrangement is based on the assumption of CWmin CWmax = CW, which deviates considerably from the 802.11 standard.

"Dynamic Tuning of the Contention Window Minimum (CWmin) for Enhanced Service Differentiation in IEEE 802.11 Wireless Ad-Hoc Networks" (L. Gannoune and S. Robert, 15th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC 04), vol. 1, pp. 311 -317, 2004) proposes an algorithm for dynamic CWmin adaptation based on measured collision rate.

The algorithm of"A TCP-like adaptive contention window scheme for WLAN" (Q.

Pang, S. C. Liew, J. Y. B. Lee, and S.-H. G. Chan, IEEE International Conference on Communications, vol. 6, pp. 3723 -3727, 2004) also concerns adaptive window adjustment rules.

A Markov chain model for 802.11 contention access (distributed coordination function) under the saturation conditions is studied in "Performance analysis of the IEEE 802.11 distributed coordination function" (G. Bianchi; IEEE JSAC, 18(3), 2000). In particular, the model has been shown to be able to provide accurate prediction of performance metrics such as saturation throughput and average delay. Based on this model several enhancements have been made to take into account the effect of QoS differentiation, such as in "Throughput and QoS optimization in IEEE 802.11 WLAN" (J. Zhao et a!.; Proceedings, of 3G Wireless, 2002) and in "Performance analysis of an enhanced IEEE 802.11 distributed coordination function supporting service differentiation" (B. Li and R. Battiti; Proceedings of QOFIS, 2003).

It will be appreciated from the following that the invention avails itself of the teaching of the foregoing, in order to solve problems associated with contention window allocation, either described above or otherwise exhibited.

According to the invention, there is provided a method of managing communication in a communication network, the network being configured to manage traffic dependent upon the nature of communication, and nature of communication being managed in classes, each class of communication having associated therewith at least one communications management criterion, the method comprising defining, within a range of number of actively communicating devices in the network, a plurality of sub-ranges and defining, for each sub-range, at least one communications management criterion for configuring management of communication in the network, determining an estimate for the active devices in said network, determining, on the basis of the estimate, the sub-range within which said estimate belongs, retrieving said at least one communications management criterion for said sub-range; and configuring communications in said network on said at least one communications management criterion.

According to a second aspect of the invention, there is provided access control apparatus for use in a communications network, the network being configured to manage traffic dependent upon the nature of communication, and nature of communication being managed in classes, each class of communication having associated therewith at least one communications management criterion, the apparatus comprising storage means storing at least one communications management criterion for configuring management of communication in the network, for each of a plurality of sub-ranges of a range of number of actively communicating devices in the network, active device estimating means for determining an estimate for the active devices in said network, sub-range selection means for determining, on the basis of the estimate, the sub-range within which said estimate belongs, data retrieval means for retrieving said at least one communications management criterion for said sub-range, and communications configuration means for configuring communications in said network in accordance with said at least one communications management criterion.

Neither the publication by Gannoune et a!. nor that by Pang et a!. deals with the issue of maintaining throughput ratios among different traffic classes and in the meantime maximizing the overall system throughput. Further, both methods are based on heuristics, while the present invention is intended to provide a more optimal CWmin by solving a theoretical model.

The invention will now be exemplified by a specific example of an embodiment thereof, described, by way of example only, with reference to the accompanying drawings, in which: Figure 1 illustrates a wireless network 10 comprising an access point 12 and a plurality of communications devices 14. The access point 12 and communications devices 14 are, in this example, configured in accordance with IEEE Standard 802.lle, but it will be appreciated that the present invention is not limited to such technology and may also be applied to other technologies, whether standard (e.g. MIBOA UWB) or otherwise.

Figure 2 illustrates the access point 12 as being implemented by means of a general purpose computer. In this case, communications facilities are provided by means of hardware, which is in turn configured by means of software. More particularly, the controller comprises a processor 30, in communication with the working memory 32 and a bus 34. A mass storage device (which, in this case, is a magnetic storage device, though other such storage devices would suffice) 36 is provided for long term storage of data and/or programs not in immediate use. A medium access point 38 is connected to an antenna array 40, to provide the controller 12 with access to the wireless communications medium. Further, a broadband modem 42 is provided to allow the communications controller to be connected to hardwired communications networks, such as the internet. This may be used to provide the network established by the access point 12 with a portal to the internet.

It will be appreciated that the access point 12 described above is but one specific embodiment of the invention, and other alternative arrangements will be apparent to the reader.

In conventional manner, the controller further comprises audiovisual output devices 44 and user operable input devices 46.

A schematic functional diagram of the access point 12 is illustrated in figure 3. This schematic functional implementation is achieved on the general purpose computer by storage and execution of suitable software. The software may be stored in the mass storage device 36 and retrieved into working memory 32 as required. To avoid limitations on capacity of the working memory 32, it may be possible to load but a part of the configuring software therein, for execution, while parts of the software not for immediate use are retained in long term storage as required.

It will be appreciated that not all functions and facilities of the access point 12 are illustrated in figure 3, and only those facilities required to demonstrate a specific embodiment of the invention are disclosed. Other facilities will be appreciated by the skilled person as being capable of provision alongside the present described embodiment.

As shown in figure 3, the access point 12 comprises a network traffic controller 16 operable to set parameters for the network which will allow the remainder of the communications devices to conduct wireless communications. These parameters are communicated to the remaining communications devices as network configuration update messages. To do this, the access point 12 comprises a network configuration manager 18 which receives various information concerning the characteristics of traffic in the network and delivers usage information to the network traffic controller 16 to allow the controller to set network communications parameters.

The network configuration manager 18 operates on the basis that the number of traffic classes is M and there are N competing stations in class j, for j 1, 2, ..., M. Operation of the network configuration manager 18 will now be described in conjunction with figure 4 which shows a flow diagram of a process performed by the network configuration manager 18.

By way of background, traffic classes are differentiated by contention windows and have the same inter-frame space (IFS) values. Pi is set as the average conditional collision probability of traffic type i, and t is the probability that the station of traffic type i transmits in a randomly chosen time slot. Then and p are related by the following two equations which can be solved numerically: -2(i-2pj) -(1 -2pj(l4 + 1) + ;.(i -

M

J).=1_(1_T)1) JJ (..)N) (2) i=i.ii where W1 and m are exponential back-off parameters of class i: W is the minimum contention window size CWmin,i, rn is the maximum back-off stage.

The saturation throughput of class i traffic is given by S. Pe+PsT+PcTc' (3) where: L is the packet size, is the duration of an empty slot time, P1 is the probability that there is one (and only one) transmission of a traffic flow of type i on the channel, PidLe is the probability that the slot is idle, P is the probability of a successful transmission in a slot, P is the probability of collision in a slot, T is the average time the channel is sensed busy due to a successful transmission, and T is the average time the channel is sensed busy during a collision.

s is then set as the saturation throughput of a single station oftypej, i.e.: sj = S INS.

Then the overall system throughput is S = E S = Here, based on the above model, the network configuration manager 18 manages joint throughput maximization and service differentiation by adapting contention windows.

Taking class I as a reference class, the ratio r, (j = 1, ..., M) is defined as the per station throughput ratio with respect to the reference class, i.e., rj = sIs, and r1 = 1. Then, the network configuration manager 18 looks to find the set of optimal minimum contention window sizes W, j = 1, ..., M such that it maximizes S. Given the expression for system throUghput (3), the optimal solution oft1 of class ito the above problem is: *-___ 1 -y b -i'[( ATi*)2 + L1 Njr212 * T1= 1-r(1-rj) .,iri, i=2,...,M.

where T* = T/cy.

With optimal t's, equation (1) is used to find the optimal W"s: iv*_ -(1-2p +p(i -where is derived from equation (2).

A simpler alternative would be to use the proportional relation of contention window sizes among traffic classes, i.e. equation (6) set out below. In this way the optimal minimum contention window sizes can be chosen to achieve maximum system throughput while in the meantime maintaining the service differentiation (throughput ratio) among traffic classes.

Ti -Wi 8j--. (6)

However, seeking the optimum solution on each repetition of the problem may cause substantial computational complexity. Therefore, in accordance with a specific embodiment of the invention, a process within the network configuration manager 18 determines information on the basis of which the network traffic controller 16 can manage the network 10 successfully, without going to the computational expense of obtaining an exact measure of the number of active devices in the network.

Class 1 (as defined in IEEE8O2. 11 e) is taken, for the purpose of this example, as a reference class. From equation (4) above, it is known that:

I

1 Nr.)2 + NjrflT From ti, W1 can be calculated and, with equation (6), Wi's of other traffic classes can also be obtained: W1=W1/r. (8) To do this, the number of competing stations N's is calculated, in order for this information to be used in equation (7). Since accurate measurement of this number is often difficult and involves quite high computational complexity, the present embodiment of the invention makes use of the phenomenon that system performance approaches optimal values with the same parameters, when the number of competing stations changes within a certain sub-range. This means that near optimal throughput performance can be achieved without the need for very accurate measurement and estimation of the number of competing stations.

The range of the number of competing stations is partitioned into many sub-ranges and the optimal protocol parameters for each sub-range is calculated based on an analytical model in advance. When the number of competing stations changes, the network configuration manager 18 adjusts the protocol parameters, based on the sub-range in which the number of competing stations is placed, to improve the whole system performance.

The network configuration manager 18 therefore, in step S 1-2, determines an equivalent number of stations U which serves as a working estimate for the actual number of active communications devices 14. U is derived from an extended form of equation (7), compared with the result for a single class of traffic, t = 1 / [N(T*/2)'2I. There are two possible definitions of U: [r M U1 = >i: N1r)2 + Njr, N (9) and a simpler one:

M LT2

i=1 (10) Then U is used in the sub-range-based adaptation protocol. The specific described embodiment of the present invention operates as follows. In step Si -2, the access point 12 measures the number of active stations in each traffic class and calculates U. On an initial execution, the process then, in step Sl-4 selects the sub-range, from a sequence of predetermined sub-ranges, within which the estimate is located. The CWmin value for that sub-range is retrieved in step Si -6 from the analytical model data storage unit 20 and then in step S 1-8 this information is submitted to the network traffic controller 16 for broadcast.

The process is then repeated at suitable intervals, as indicated by step S 1-10. On the second and further occasions of execution of step S1-2, the process checks in step S1-12 to determine if the estimate just obtained is within the currently used sub-range. If this is the case, then the process ends and is again repeated from step SI -2 onwards at a

suitable interval.

Otherwise, the sub-range must be re-selected in step Si -4 and then onwards. When the access point 12 finds that the equivalent number of competing stations goes out of the border of current sub-range, it selects the proper sub-range in step S 1-4 and broadcasts (as initiated in steps S1-6 and S1-8) the optimal CWmin (which is Wi for class i in the present case) of the sub-range to all communications devices 14 in beacon messages.

After receiving the beacon messages, each device 14 adjusts its CWmin and CWmax based on (8). When a station joins the network and associates with the access point, it sets its CWmin based on the CWmin of the system currently in use.

The optimization framework described above can also be used for the purpose of admission control in wireless networks, if it is assumed that, in a wireless LAN or PAN, there are K stations with an optimal contention window set (Wi, ..., WK) that maximizes the system throughput. In addition, the actual throughput s1 of station i satisfies its bandwidth requirement B, i.e., s1 >= B1, i = 1, ..., K. When a new station tries to join the network with bandwidth requirement BK + , the access point calculates the thixughput ratios r1 = B / B1 and computes a new optimal contention window set (W11, W21, ..., WK+ i') using (4) and (5). Then, with (3), it is verified that the resulting new throughput of every station still satisfies their bandwidth requirement, i.e. s' > B, i = 1, ..., K + 1. If the above test is passed, then the new station K + 1 is admitted to the network and all the stations set their contention windows according to the new Ws. Otherwise it is rejected.

Claims

CLAIMS: 1. A method of managing communication in a communication

network, the network being configured to manage traffic dependent upon the nature of communication, and nature of communication being managed in classes, each class of communication having associated therewith at least one communications management criterion, the method comprising: defining, within a range of number of actively communicating devices in the network, a plurality of sub-ranges and defining, for each sub-range, at least one communications management criterion for configuring management of communication in the network, determining a measure of the number of active devices in said network, determining, on the basis of the measure, the sub-range within which said measure belongs, retrieving said at least one communications management criterion for said sub-range; and configuring communications in said network on said at least one communications management criterion.

2. A method in accordance with claim 1 wherein the step of determining a measure comprises determining an estimate.

3. A method in accordance with claim 2 wherein the step of determining an estimate comprises monitoring traffic for each class of communications and collating said communications to determine an estimate.

4. A method in accordance with claim 3 wherein said step of collating comprises determining, for each class, the product of the number of competing stations making communications in that class and the per station throughput ratio for that class.

5. A method in accordance with claim 4 wherein said step of collating comprises determining the sum of the products determined for the classes in the network, the sum being an estimate of the number of stations active in the network.

6. A method in accordance with any preceding claim wherein said step of defining at least one communications management criterion comprises determining a minimum contention window duration for communication for that class.

7. Access control apparatus for use in a communications network, the network being configured to manage traffic dependent upon the nature of communication, and nature of communication being managed in classes, each class of communication having associated therewith at least one communications management criterion, the apparatus comprising storage means storing at least one communications management criterion for configuring management of communication in the network, for each of a plurality of sub-ranges of a range of number of actively communicating devices in the network, active device measurement means for determining measure of the number of the active devices in said network, sub-range selection means for determining, on the basis of the measure, the sub-range within which said measure belongs, data retrieval means for retrieving said at least one communications management criterion for said sub-range, and communications configuration means for configuring communications in said network in accordance with said at least one communications management criterion.

8. Apparatus in accordance with claim 7 and wherein said measurement means comprises active device estimating means, for determining an estimate of the number of active devices in the network.

9. Apparatus in accordance with claim 8 wherein the estimating means is operable to monitor traffic for each class of communications and to collate said communications to determine an estimate.

10. Apparatus in accordance with claim 9 wherein the estimating means is operable to determine, for each class, the product of the number of competing stations making communications in that class and the per station throughput ratio for that class.

11. Apparatus in accordance with claim 10 wherein the estimating means is operable to determine the sum of the products determined for the classes in the network, the sum being an estimate of the number of stations active in the network.

12. Apparatus in accordance with any of claims 8 to 11 wherein said step of defining at least one communications management criterion comprises determining a minimum contention window duration for communication for that class.