CN104052745B - Contention window adjusting method taking 802.11e VoIP application as orientation - Google Patents

Abstract

The invention discloses a contention window adjusting method taking an 802.11e VoIP application as orientation. The method includes the steps that (1) the optimal network access probability TauAPopt of a wireless AP is calculated according to an access MAC layer parameter and a VoIP coding mode; (2) the optimal value CWminAP of the size of the minimum contention window and the optimal value CWmaxAP of the size of the maximum contention window of the AP are calculated according to the TauAPopt; (3) the average collision rate favgj in the jth period at the AP is calculated, and then the network access probability TauSTA of a site STA is adjusted according to the relation of the favgj and the threshold value fopt; (4) the maximum contention window and the minimum contention window of the site are calculated by the AP according to the TauSTA; the maximum contention window and the minimum contention window of the site are broadcasted by the AP by using a periodic Beacon frame. By means of the contention window adjusting method taking the 802.11e VoIP application as orientation, the effective availability of a wireless channel can be improved, the network throughput is improved, and the VoIP capacity is improved as well.

Description

Competition window method of adjustment towards 802.11e VoIP application

Technical field

The present invention relates to a kind of competition window method of adjustment towards 802.11e VoIP application.

Background technology

With developing rapidly of voice (VoIP) technology based on Internet Protocol and WLAN (WLAN) technology, base In the VoIP of WLAN obtained in recent years quickly, be widely applied.IEEE802.11 distributed coordination function (DCF) media access (MAC) agreement is controlled to become the WLAN MAC protocol of current main flow with its simplification and vigorousness, its basic thought is that carrier wave is detectd Listen multiple access (CSMA) and conflict avoidance (CA).IEEE802.11DCF agreement is led using transmitting non real-time nature data flow as it Business to be supported, in order to strengthen the support to real time business such as VoIP, IEEE makes enhancing distributed channel access mechanism (EDCA, Enhanced Distributed Channel Access).EDCA carries out QoS on the basis of DCF mechanism and supports to expand Exhibition, there is provided different Business Streams is carried out with the mechanism of prioritization, improves interior VoIP of WLAN (WLAN) etc. real When business with respect to other Business Streams priority, thus ensureing its qos requirement to a certain extent.But EDCA MAC layer parameter Using fixed value it is impossible to carry out Automatic adjusument according to network condition, systematic function can not be made to realize optimum, much study table Bright, under high load condition, due to there being higher collision rate in network, the performance of EDCA is simultaneously not so good as people's will；Additionally, EDCA Website in WLAN (Stations, STAs) and access node (AP) are not carried out with the differentiation of priority, leads to up-downgoing at AP The imbalance of service traffics, and cause bottle neck effect, limit VoIP capacity.

(1) static parameter causes higher collision rate in high-load network

When working in basic access way, its workflow is CSMA/CA mechanism：Node is when channel idle is detected Between send packet afterwards more than DIFS (distributed inter frame space)；If channel busy, node is waiting channel empty Keep out of the way randomly choosing one time delay (backoff period value) and start back-off timer after not busy DIFS, wait this to delay when keeping out of the way Send packet.Back-off timer successively decreases in time in channel idle；To hang up during channel busy, until idle time of channel exceedes Just recover after DIFS to successively decrease；When back-off timer is kept to 0, node starts to send packet.Backoff period value discrete segment [0, CW] in by being uniformly distributed random selection.CW is also called competition window, and competition window initial value is CW_min, send unsuccessfully every time, competing Strive window all to double, and restart said process contention access channel, until sending success or reaching maximum contention window CW_max.Source node, after sending packet, will wait destination node to send the ACK frame confirming to receive；And destination node just receives After true packet, then to send ACK frame to source node after waiting channel idle SIFS, the packet receiving is confirmed.

CSMA/CA mechanism conflict backoff algorithm avoids conflict next time by expansion competition window, but cannot determine Can successfully avoid, perhaps could be able to send successfully through multiple backoff procedure.A lot of researchs show, in high-load network, After one packet generally will experience multiple conflict, could successfully be sent.This mode be difficult to ensure that VoIP transmission when pair The guarantee of QoS such as postpone, shake.CW in EDCA_minWith CW_maxStatic state setting, lead in high-load network most packets to exist Identical discrete segment [0, CW_max] in selected, increased conflict probability.The increase of conflict wastes channel money significantly Source, makes handling capacity decline it is impossible to realize the optimum of systematic function rapidly.

(2) the bottle neck effect restricting VoIP capacity of AP

In a wireless local area network, access node AP (Access Point) is located at the combination of cable network and wireless network Portion, downstream from cable network to wireless network for the simultaneous transmission and the upstream from wireless network to cable network.VoIP leads to Carry out between wireline station outside generally in wireless site and AP for the words process or wireless site.VoIP is only existed in hypothesis system Data flow, each website establishes a two-way VoIP communication process by AP with external site.When there is N number of website, There is the transmission link competitive channel that N bar is up, and channel need to be completed by AP from the reciprocal voice call process of external site Access.Comparatively, the load that sends in AP is that each website sends N times loading.According to the principle of CSMA/CA mechanism, AP has identical access channel probability with each website, is all the access completing wireless medium by fair play.When VoIP talking link capacity increase, when system reaches saturation, AP due to overload produce stack overflow and affect from wired to The data transfer in wireless direction, produces the loss of VoP, the increase of time delay.The increase impact of these packet loss and time delay The VoIP communication process that all websites are initiated, limits the increase of power system capacity.

To sum up it is necessary to design a kind of new self-adapting regulation method, to protocol parameter self-adaptative adjustment to ensure difference Protocol capabilities under network load condition；Website and AP can be carried out with prioritization simultaneously, weaken the bottle neck effect pair of AP The restriction of VoIP capacity, lifts VoIP capacity.

Content of the invention

The technical problem to be solved is to provide a kind of competition window adjustment towards 802.11e VoIP application Method, should can lift wireless channel effective rate of utilization towards the competition window method of adjustment of 802.11e VoIP application, improve net Network handling capacity, lifts VoIP capacity.

The technical solution of invention is as follows：

A kind of competition window method of adjustment towards 802.11e VoIP application,

1) calculate, according to the MAC layer parameter accessing and VoIP coded system, the optimal network that WAP is AP to access generally Rate

Described MAC layer parameter includes slot time, SIFS, DIFS and transfer rate；When slot time represents gap Between；SIFS (Short Interframe Space) is short frame period；DIFS is distributed coordination frame period (DCF Interframe Space)；

2) basisCalculate minimum, the optimal value of maximum contention window size of AP

3) calculate the average collision rate in j-th cycle at APFurther according toWith threshold value f^optRelation adjustment website (STA) network insertion probability τ_STA：

WhenWhen, reduce the network insertion probability τ of website_STATo adapt to offered load；

WhenWhen, increase the network insertion probability τ of website_STATo adapt to offered load；

4) AP is according to τ_STACalculate maximum, the minimum competition window of website；The Beacon frame broadcast station of AP usage cycles Maximum, minimum competition window.

Step 1) inComputational methods as follows：

Wherein δ is slot time；T_dataAnd T_ackWhen being one VoIP data bag of transmission and the corresponding ACK of transmission respectively Between；【It is limited to the transfer rate of wireless network and the VoIP encoding mechanism of selection】.

Step 2) inWithComputational methods as follows：

Wherein m keeps out of the way number of times for maximum, and value is 3.

Step 3) in,It is taken as current period collision rateWith historic conflict rateEWMA, I.e.：

Wherein, α is smoothing factor, and value is 0.875；Average collision rate for jth -1 cycle；For jth week The collision rate of phase, is determined by following formula：

Wherein, E (data_collisions_j) it is the data conflicted in the jth cycle Bag number, E (data_sent_j) it is the number of data packets sending in the jth cycle.

Step 3) in, the network insertion probability τ of website_STADetermine as the following formula：

Wherein, ε is linear increasing factor, and ε=0.1/n, wherein n are network insertion station number；β for ratio reduce because Son, β value is 5/6；f^optValue be 0.1.The adjustment mode that this linearly increasing property taken advantage of reduces (AIMD) can be in network load When lighter, the efficient resource in use network faster；But when network load is heavier, when collision rate is higher, again can be more anxious Reduce the access probability of website acutely, mitigate network conflict rate.

Step 4) in, the minimum contention window size of websiteWith maximum contention window sizeCalculating public Formula is as follows：

Wherein, m keeps out of the way number of times, value 3 for maximum；The branch hazard probability of website N is network insertion station number.

Beneficial effect：

The competition window method of adjustment towards 802.11e VoIP application of the present invention, by by the competition of WAP Window is set to optimum constant, simultaneously by the mode of WAP foundation network conflict rate Automatic adjusument website competition window The VoIP capacity of lifting 802.11e network.The competition window of WAP is one and (includes slot by MAC layer parameter Time, SIFS, DIFS and transfer rate) and VoIP coded system determine, the constant unrelated with network insertion station number.Website Competition window Automatic adjusument is carried out according to network conflict rate by WAP.If it is big that WAP detects collision rate In given threshold value, will be according to actual conflict rate value, the access calculating adaptation current network load using the mode that the property taken advantage of reduces is general Rate；If the collision rate detecting is less than or equal to given threshold value, will be calculated using linearly increasing mode according to actual conflict rate value Adapt to the access probability of current network load.Finally according to the relation of access probability and competition window, calculate the maximum of website Minimum competition window.The present invention can guarantee that the fairness of wireless network up-downgoing stream, and the bottle neck effect alleviating AP is to VoIP capacity Restriction, can be according to network load situation, Automatic adjusument contention window value, lifted wireless channel effective rate of utilization, improve net Network handling capacity, reduces network delay, improves VoIP capacity.

The competition window self-adapting regulation method that this is used for VoIP application in 802.11e wireless network environment should according to VoIP Feature, the access probability optimizing AP is N times of website access probability (N is the number of access website in network), by reason By derivation, show that the competition window of AP is only related to MAC layer parameter and VoIP coded system for one, unrelated with competition website number Constant expression；Additionally, can be according to network load situation, Automatic adjusument website contention window value.This method of adjustment is lifted AP priority, realizes the fair competition of up-downgoing flow at AP, weakens the restriction to VoIP capacity of the bottle neck effect of AP；

The present invention uses the minimax contention window of periodically Beacon frame real-time update website in 802.11e network Mouthful, website need not be modified, be easy to actual deployment.

Brief description

Fig. 1 is the flow chart of the present invention.

Fig. 2 is test environment topological diagram of the present invention.

Fig. 3 is handling capacity with network access station points variation diagram (G.711,802.11b).

Fig. 4 is handling capacity with different coding mechanism variation diagram (10STAs, 802.11b).

Fig. 5 is the ratio (G.711,10STAs, 802.11b) of up-downgoing stream handling capacity.

Fig. 6 is the VoIP capacity (G.711) of algorithms of different.

Fig. 7 is the CDF scattergram (G.711,50STAs, 802.11g) of packet loss.

Fig. 8 is the CDF scattergram (G.711,50STAs, 802.11g) of network delay.

Fig. 9 is that in hybrid network environment, DATA flows handling capacity with VoIP speech channel number variation diagram (G.711,802.11b).

Figure 10 is VoIP flow network delay figure (G.711,802.11b) in hybrid network environment.

Figure 11 is VoIP flow packet loss figure (G.711,802.11b) in hybrid network environment.

Specific embodiment

Below with reference to the drawings and specific embodiments, the present invention is described in further details：

Embodiment 1：

Referring to Fig. 1, it is the flow chart of the present invention.It is the method realized on AP, and process is as follows：

According to the 802.11 mechanism corresponding MAC layer parameter accessing and corresponding VoIP coded method, calculate AP and access The optimal value of probability, uses variableStorage：

Wherein δ is slot time；T_dataAnd T_ackWhen being one VoIP data bag of transmission and the corresponding ACK of transmission respectively Between, computational methods are the length of DATA or ACK bag divided by physical transmission rate.For VoIP application, the length of DATA bag is The packet size of VoIP.

Based on calculatedAccording to formula

Calculate the optimum contention window parameter of APWithWherein m keeps out of the way number of times, value 3 for maximum.It is based on This numerical value, AP updates AC_VO (VoIP rank data flow) the most respectively in updateCWMin (), updateCWMax () function Little maximum contention window value.

VoIP application needs symmetrical up-downgoing flow, and for the VoIP data stream of N number of website, AP needs transmission Load capacity is N times of each website load capacity.Accordingly, it would be desirable to the access probability of preferential AP, alleviate AP as bottleneck node pair The restriction that VoIP capacity causes.The present invention by the access probability of AP being set to N times of website access probability, through a series of Strictly shift onto, obtain the expression formula of AP optimum access probability.This expression formula be one unrelated with network insertion station number, only The constant being determined by MAC parameter and VoIP coded system.

For website competition window adaptive updates, realize process as follows：

Defined variable DATAFailureCount, DATATotalCount are respectively used to count in each cycle, retransmit DATA data packet number and transmission DATA data packet number, the ratio of the two as current period (j-th cycle) conflict The rate factor, and use variableStorage, that is,：

For reducing the deviation of instantaneous conflict, the average collision rate factor is proposedIt is made dynamically to count within a update cycle Calculate, to reflect the average collision rate in j-th update cycle, its computing formula is as follows：

Wherein, α is collision smoothing factor, and value is 0.875.

Arrive when a cycle time, complete including DATAFailureCount, DATATotalCount variable is at the beginning of interior Beginningization.Compare the deviation of this cycle average collision rate and given collision rate threshold value, the side reducing using the linearly increasing property taken advantage of simultaneously Formula adjusts channel access rate τ, is calculated as follows：

When network conflict rate is more than given threshold value, show that network load is heavier, higher channel conflict causes the wasting of resources. Therefore, now network conflict rate should be reduced by reducing channel access probability.Conversely, when network conflict rate is less than or equal to During given threshold value, illustrate that network load is lighter, channel time slot idleness is higher.Therefore it should improve channel access probability, thus Lifting channel time slot utilization rate, lifts handling capacity.

Based on calculated website access probability, using formula：

Wherein, m keeps out of the way number of times, value 3 for maximum；N counts for network access station Mesh.Calculate minimum, the maximum contention window of website respectively

Finally, AP uses Beacon frame, periodically broadcasts site parameter to website (100ms).

Test result on NS2 emulation platform shows, the inventive method with acquiescence EDCA layer mechanism and Y.Gao et al. DCF mechanism and enhancement mechanism BDCF of up-to-date proposition in 2013 is compared, and can effectively lift network throughput, balance up-downgoing flow, time delay, Packet loss, improves VoIP capacity.

Fig. 2 is the network topological diagram that we carry out testing, including an AP, 70 wireless sites and 70 wired sections Point.Wireless site is deployed in AP as the center of circle, and radius is on 500 meters of circle.Descending TCP flow is sent to wirelessly from there being source nodes Receiving node, up TCP flow is sent to cable network reception node from wireless source node.The a width of 100Mbps of band of all wire links, Time delay is 2ms.All wireless sites (including AP) use Destination-Sequenced Distance-Vector Routing (DSDV) Routing Protocol, wireless access mechanism uses the basic access way of 802.11b or 802.11g.Each node Only receive or send a VoIP flow or UDP background stream, during transmission UDP background stream, data package size is 1000bytes, number It is 1Mbps according to transmission rate.The feature of the encoding mechanism used in experiment is as shown in table l.In method, some parameter settings are as follows： α=0.875, m=3, β=5/6.The performance of method of adjustment BEDCA of the present invention and acquiescence EDCA mechanism and BDCF is with offered load Change is as shown in Fig. 3-11.

Table l different coding mechanism characteristic

In conjunction with Fig. 3 to Fig. 5 as can be seen that increase with offered load (network insertion station number increases or VoIP Code rate increases) the inventive method BEDCA, compared with traditional EDCA, BDCF, can obtain close optimum network throughput, energy The fairness of active balance up-downgoing stream handling capacity.Additionally, from Fig. 6 to Fig. 8 as can be seen that BEDCA can active balance network when Prolong, packet loss, lifted VoIP capacity.In conjunction with Fig. 9 to Figure 11 as can be seen that in the network environment that VoIP flow is coexisted with DATA stream, BEDCA can lift VoIP capacity while ensureing DATA stream handling capacity.BEDCA is directed to the feature of VoIP application, by excellent The access probability of first AP is N times (N is the number of the website participating in competition) of website access probability, and self-adaptative adjustment competes website Access probability, effectively reduce network conflict rate it is ensured that in the public affairs of the aspects such as handling capacity, packet loss, time delay between up-downgoing stream Levelling, lift system VoIP capacity.

Claims (6)

1. a kind of towards 802.11e VoIP application competition window method of adjustment it is characterised in that：

1) the optimal network access probability that WAP is AP is calculated according to the MAC layer parameter accessing and VoIP coded system

Described MAC layer parameter includes slot time, SIFS, DIFS and transfer rate；Slot time represents off time； SIFS is short frame period；DIFS is distributed coordination frame period；

2) basisCalculate minimum, the optimal value of maximum contention window size of AP

3) calculate the average collision rate in j-th cycle at APFurther according toWith threshold value f^optRelation adjustment website (STA) Network insertion probability τ_STA：

WhenWhen, reduce the network insertion probability τ of website_STATo adapt to offered load；

WhenWhen, increase the network insertion probability τ of website_STATo adapt to offered load；

4) AP is according to τ_STACalculate maximum, the minimum competition window of website；The Beacon frame broadcast station of AP usage cycles is Greatly, minimum competition window.

2. the competition window method of adjustment towards 802.11e VoIP application according to claim 1 is it is characterised in that walk In rapid 1)Computational methods as follows：

${\mathrm{\τ}}_{\mathrm{AP}}^{\mathrm{opt}}=\sqrt{\frac{\mathrm{\δ}}{T_{\mathrm{data}}+\mathrm{SIFS}+T_{\mathrm{ack}}+\mathrm{DIFS}}}$

Wherein δ is slot time；T_dataAnd T_ackIt is one VoIP data bag of transmission and the time transmitting corresponding ACK respectively.

3. the competition window method of adjustment towards 802.11e VoIP application according to claim 1 is it is characterised in that walk In rapid 2)WithComputational methods as follows：

$\begin{array}{c}{\mathrm{CW}}_{\min }^{\mathrm{AP}}=(\frac{2}{{\mathrm{\τ}}_{\mathrm{AP}}^{\mathrm{opt}}}-1){(1+(1-e^{-{\mathrm{\τ}}_{\mathrm{AP}}^{\mathrm{opt}}})\underset{i=0}{\overset{m-1}{\mathrm{\Σ}}}{\left(2(1-e^{-{\mathrm{\τ}}_{\mathrm{AP}}^{\mathrm{opt}}})\right)}^i)}^{-1}\\ {\mathrm{CW}}_{\max }^{\mathrm{AP}}=2^m*{\mathrm{CW}}_{\min }^{\mathrm{AP}}\end{array}$

Wherein m keeps out of the way number of times for maximum, and value is 3.

4. the competition window method of adjustment towards 802.11e VoIP application according to claim 1 is it is characterised in that walk In rapid 3),It is taken as current period collision rateWith historic conflict rateEWMA, that is,：

$f_{\mathrm{avg}}^j=(1-\mathrm{\α})*f_{\mathrm{curr}}^j+\mathrm{\α}*f_{\mathrm{avg}}^{j-1};$

Wherein, α is smoothing factor, and value is 0.875；Average collision rate for jth -1 cycle；For rushing of jth cycle Prominent rate, is determined by following formula：

Wherein, E (data_collisions_j) it is the number-of-packet conflicted in the jth cycle Mesh, E (data_sent_j) it is the number of data packets sending in the jth cycle.

5. the competition window method of adjustment towards 802.11e VoIP application according to claim 1 is it is characterised in that walk In rapid 3), the network insertion probability τ of website_STADetermine as the following formula：

${\mathrm{\τ}}_{\mathrm{STA}}=\left\{\begin{array}{c}{\mathrm{\β\τ}}_{\mathrm{STA}},f_{\mathrm{avg}}^j>f^{\mathrm{opt}}\\ {\mathrm{\τ}}_{\mathrm{STA}}+\mathrm{\ϵ},f_{\mathrm{avg}}^j\≤f^{\mathrm{opt}}\end{array}\right.,$

Wherein, ε is linear increasing factor, and ε=0.1/n, wherein n are network insertion station number；β reduces the factor for ratio, and β takes It is worth for 5/6；f^optValue be 0.1.

6. the competition window method of adjustment towards 802.11e VoIP application according to any one of claim 1-5, it is special Levy and be, step 4) in, the minimum contention window size of websiteWith maximum contention window sizeComputing formula As follows：

$\begin{array}{c}{\mathrm{CW}}_{\min }^{\mathrm{STA}}=(\frac{2}{{\mathrm{\τ}}_{\mathrm{STA}}}-1){(1+p_{\mathrm{STA}}\underset{i=0}{\overset{m-1}{\mathrm{\Σ}}}{\left(2p_{\mathrm{STA}}\right)}^i)}^{-1}\\ {\mathrm{CW}}_{\max }^{\mathrm{STA}}=2^m*{\mathrm{CW}}_{\min }^{\mathrm{STA}}\end{array}$

Wherein, m keeps out of the way number of times, value 3 for maximum；The branch hazard probability of websiteN is Network insertion station number.