KR100927532B1 - Wireless LAN Scheduling Device - Google Patents

Abstract

The scheduling apparatus of a wireless LAN according to the present invention includes an SI table setting unit for setting an SI table including a minimum SI (Service Interval) that can be scheduled and a maximum service interval (MSI) of each station among SIs in the SI table. By including an SI allocator that allocates the SI (approximate SI) to each station, which is the closest to the cycle, it is possible to service the MSI to each station at the closest interval, thereby reducing polling overhead and increasing the TXOP utilization to serve the network. It can increase the number of stations that can be.

WLAN, Scheduling, SI, MSI, TXOP, IEEE802.11

Description

Scheduling APPRATUS FOR WIRELESS LOCAL AREA NETWORKS

The present invention relates to a scheduling apparatus for a WLAN, and more particularly, to an apparatus for enabling more reliable scheduling of a WLAN by improving an SI allocation scheme of a station participating in the WLAN.

Since IEEE 802.11e HCF Controlled Channel Access (HCCA) is a central control method using a polling mechanism, an algorithm for scheduling proper transmission of poll frames and allocation of TXOP is required. In the IEEE 802.11e standard, a reference scheduler is proposed as a basic scheduling algorithm.

The reference scheduler schedules using Mean Data Rate, Nominal MAC Service Data Unit (MSDU) Size, and Maximum Service Interval (MSI) among TSPEC (Traffic Specifications) parameters. Mean Data Rate is the average transmission rate of packets arriving from the upper layer to the MAC layer, Nominal MSDU Size represents the average MSDU size of the traffic stream, and MSI represents the maximum time interval between two consecutive transmissions of a particular traffic stream. . Traffic streams accepted by the HC (Hybrid Coordinator) are scheduled in two steps. The first process is a process of calculating a service interval (SI), and the second process is a process of calculating a transmission opportunity (TXOP) allocated to a traffic stream. The process of calculating the SI is as follows. Select the smallest value among the MSI values of the accepted traffic streams. The largest value among divisors of the beacon interval not exceeding this value is defined as SI. The SI thus determined is used to schedule all stations in the polling list.

Once the SI is determined, TXOP is calculated using Mean Data Rate (ρ _i ), Nominal MSDU Size (L _i ), Physical Transmission Rate (R _i ), and the maximum size of MSDU M (2304 bytes). First, the scheduler calculates the number N _i of the average MSDUs arriving at the MAC layer at Mean Data Rate during SI as shown in Equation 1 below.

Then, as shown in Equation 2, when the Physical Transmission Rate is R _i , the i-th accepted traffic is the larger of the time required to transmit N _i average size MSDUs and the time required to transmit one MSDU of maximum size. Assigned by TXOP _i of the stream. O is the sum of the time for transmitting QoS CF-Poll frame, the time for receiving ACK for data frame, and the IFS (Iterated Function System) time for polling overhead.

The admission control algorithm of the standard document is as follows. When the new traffic stream requests acceptance, the number of MSDUs is calculated using the newly calculated SI as in Equation 1, and then the TXOP of the traffic stream is calculated as in Equation 2. Then, if the inequality of Equation 3 is satisfied, the traffic stream is accepted. Where k is the number of traffic streams accepted to date, k + 1 is the traffic streams requesting the current acceptance, T is the beacon interval, and T _cp is the time used as EDCA.

The reference scheduler is inefficient in terms of polling overhead because it polls all accepted stations with the same SI as described above. Since the SI in the reference scheduler is determined by the smallest of the MSI values of the accepted stations, the polling period can be much smaller than necessary for stations whose MSI is relatively larger than the SI, leading to an increase in the polling overhead rate. . The polling mechanism causes polling overhead due to the transmission of poll frames and inter frame space (IFS). This is because a shorter polling period increases overhead because more poll frames are transmitted.

If the SI of the station becomes smaller than necessary, there is also a problem in that a lot of time is wasted because there is no data to send when the station is polled. The reference scheduler allocates TXOP in proportion to the average number of MSDUs (integer value) generated per SI, and guarantees the minimum time to send one MSDU of maximum size. Therefore, when the SI of the station becomes small, TXOP does not decrease in direct proportion to the SI. If the SI becomes smaller than necessary, a large amount of TXOP is allocated to the average data generation amount, which reduces the efficiency of the TXOP. In addition, as the SI of the station decreases, the average number of MSDUs generated per SI is reduced, and the TXOP allocated per SI is also reduced, thereby reducing the frequency of using the Block ACK, which can reduce overhead. Therefore, the reference scheduler has a large unnecessary polling overhead and has a problem of lowering the efficiency of TXOP. This results in reducing the overall transmission efficiency and the number of acceptable stations.

The present invention was created to improve the above problems, and aims to increase the number of stations that can be serviced in the network by reducing polling overhead of each station and maximizing TXOP utilization.

It is also another object that the scheduling is completed while the station is admission controlled, providing a convenient benefit in stages.

In order to achieve the above object, the present invention provides an SI table setting unit for setting an SI table including a minimum service interval (SI) that can be scheduled; And an SI allocator for allocating an SI (approximate SI) closest to an MSI (Maximum Service Interval) of each station among the SIs in the SI table to each station.

Here, the SI table is a minimum SI that can be scheduled; And SI _k to which a regular increase to the minimum SI is applied, wherein SI _k is

SI _k = m × SI _(k-1)

SI _(k-1) : previous SI _k (SI ₀ is minimum SI)

m: real> 1

k: positive integer

It is desirable to satisfy. At this time, m is preferably 2.

In addition, it is preferable to further include an acceptance determination unit for calculating the transmission opportunity (TXOP) of each station using the assigned SI, and determines whether to accept the participation of the station based on the TXOP. In this case, the acceptance determination unit TXOP (TXOP _i ) of the station is the following equation

here,

SI _i : SI of the station

ρ _i : Mean Data Rate

L _i : Nominal MSDU (MAC Service Data Unit)

R _i : Physical Transmission Rate

O: Polling overhead (sum of QS CF-Poll frame transmission time, ACK receiving time for data frame, Inter Frame Space time)

TXOP _n : TXOP of each station in each polling group

TXOPSUM: Total TXOP for each polling group

BSI: minimum SI

l: the number of stations included in the polling group

It is preferable to accept the station when satisfying, and the acceptance determination unit, if the TXOP of the station does not satisfy the above equation, the SI _i to the following equation

SI _i = SI _(i-1) / m

After resetting by applying the above, if the above process of applying the above equation is satisfied again, the SI _i is accepted as the polling group of the assigned station, and if it is not satisfied until the SI _i is BSI, the acceptance is rejected. desirable.

In addition, when there are a plurality of polling groups satisfying the TXOP of the station, the acceptance determining unit may accept the participation of the station as the polling group having the minimum TXOPSUM.

On the other hand, the SI allocator is smaller than the MSI of each station, it is preferable to assign the SI closest to the MSI to each station.

As described above, the scheduling apparatus of the WLAN according to the present invention reduces the polling overhead and increases the TXOP utilization rate by increasing the number of stations that can be serviced in the network by serving at a cycle closest to the MSI of each station.

In addition, the present invention has defined a scheduling tree in order to simplify admission control and scheduling, which represents information about the SI, polling time, and TXOP of each station included in the polling list of the HC, and is polled within the same SI. Assign stations to the same group. By using the scheduling tree, in each polling group serviced within the same SI, new stations can be serviced without changing the SI and TXOP of other stations already participating in joining or leaving the network. In addition, the scheduling is completed at the same time as the admission control to obtain a simple step advantage.

As a result, reliable scheduling is possible in the WLAN.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically illustrating an apparatus for scheduling a WLAN according to an embodiment of the present invention.

Referring to FIG. 1, a scheduling apparatus of a wireless LAN according to the present embodiment includes an SI table setting unit 10 for setting an SI table including a minimum service interval (SI) that can be scheduled; And an SI allocator 20 for allocating an SI (approximate SI) closest to an MSI (Maximum Service Interval) of each station among the SIs in the SI table.

The SI table setting unit 10 tables the SI to be allocated to each station in advance. In the conventional reference scheduler, one SI (generally minimum SI) is allocated to each station, whereas in the present embodiment, various SIs are allocated to each station, and the reference is assigned to the SI table setting unit 10. It becomes the SI table set by.

The SI table is composed of a minimum SI (BSI, Basic SI) that can be scheduled and an SI _k to which a regular increase is applied to the minimum SI. An example thereof is shown in Table 1 below.

SI table 20ms (Basic SI) 40 ms 80 ms 160 ms

In this embodiment, 20 ms is set as the minimum SI (set in consideration of network conditions, etc.), and an SI table is formed with 40 ms, 80 ms, and 160 ms, which is SI _k to which regular increments are applied. Thus, one SI in total may be assigned to each station. Since a plurality of SIs to be allocated to each station exists, any one SI may be selectively assigned to each station according to a predetermined rule, which is performed through the SI allocator 20.

The SI _k must have a regular increase applied to the minimum SI, because otherwise scheduling and acceptance algorithms can be complicated. In this embodiment, the SI _k is set by Equation 4 below.

SI _k = m × SI _(k-1)

SI _(k-1) : previous SI _k (SI ₀ is minimum SI)

m: real> 1

k: positive integer

That is, the SI table is formed in such a manner that m times the previous SI value, thereby simplifying an acceptance algorithm for joining a station to be described later. It is more preferable that m is 2, because the calculation is simple without giving a large interval to the SI value of each participating station on the network.

The SI allocator 20 appropriately allocates each SI in the SI table set by the SI table setting unit 10 to each station. The allocation of the SI is made based on the MSI (Maximum Service Interval) of each station to which the SI is to be allocated. More specifically, the SI (approximate SI) closest to the MSI is allocated to the stations. do. In general, if the approximate SI is larger than the MSI of the station to which it is assigned, scheduling will not be meaningful, so the approximate SI is preferably determined from a value smaller than the MSI of the station.

In summary, the scheduling apparatus of the WLAN according to the present invention can reduce the polling overhead by serving the cycle closest to the MSI of each station, and increase the number of stations that can be serviced in the network by increasing the TXOP utilization rate. Will be. Data on this will be described later.

FIG. 2 further includes an acceptance determining unit 30 for calculating a transmission opportunity (TXOP) of each station using the allocated SI and determining whether to accept the participation of the station based on the TXOP. The constitution was shown.

Unlike the prior art, since each station is given the most close SI value to its own MSI, when there are a large number of stations to participate, processing regarding acceptance of participation may be problematic. Accordingly, the acceptance determination unit 30 is required to calculate the TXOP using the assigned SI, that is, the approximate SI, and determine whether to accept the participation of each station based on the TXOP. An improvement of the judgment method is also required.

Conventionally, such acceptance has been processed by Equation 3, but it may be difficult to apply Equation 3 in the embodiment of FIG. Therefore, the acceptance determination unit 30 preferably accepts the participation of the station when the TXOP (TXOP _i ) of the station satisfies the following Equations 5a, 5b and 5c.

here,

SI _i : SI of the station

ρ _i : Mean Data Rate

L _i : Nominal MSDU (MAC Service Data Unit)

R _i : Physical Transmission Rate

O: Polling overhead (sum of QS CF-Poll frame transmission time, ACK receiving time for data frame, Inter Frame Space time)

TXOP _n : TXOP of each station in each polling group

TXOPSUM: Total TXOP for each polling group

BSI: minimum SI
M: maximum size of MSDU

l: the number of stations included in the polling group

If the TXOP of the station does not satisfy the equations 5a, 5b and 5c, the acceptance determination unit 30 resets the SI _i by applying the following equation 6, and then the equations 5a, Repeat the process of applying 5b and 5c to satisfy the equations 5a, 5b and 5c, and accept the polling group of the station to which the SI _i is assigned, and rejects the acceptance if the SI _i is not satisfied until the SI _i is BSI. Done. M in Equation 6 is the same as m in Equation 4, thereby simplifying the acceptance algorithm for station participation and defining a scheduling tree corresponding to the SI table in which m is involved.

SI _i = SI _(i-1) / m

On the other hand, as a result of the application of Equations 5a, 5b and 5c, there may be a case where a plurality of polling groups satisfying the TXOP of the station exists. In this case, the acceptance determination unit 30 polls having the minimum TXOPSUM among the polling groups. Allow the group to accept the station's participation.

In summary, the SI table setting unit, the SI allocator, and the acceptance determination unit thereof as described with reference to FIGS. 1 and 2 are provided to define a scheduling tree according to the SI table. The scheduling tree expresses information about the SI, polling time, and TXOP of each station included in the polling list of the HC, and designates the stations polled in the same SI as the same group. By using the scheduling tree, in each polling group serviced within the same SI, a new station can be serviced without changing the SI and TXOP of other stations already participating in joining or leaving the network. In addition, the scheduling is completed at the same time as the admission control to obtain a simple step advantage.

This effect will be confirmed through operation description of the embodiment of FIG. 2.

This embodiment includes a first step of defining an SI table setting and a scheduling tree for admission control and traffic scheduling of a station requesting acceptance; A second step of allocating an SI closest to the MSI of each station from the SI table set in the first step; A third step of calculating and assigning TXOP with the SI assigned in the second step; A scheduling and admission control technique including a fourth step of determining whether to accept a station requesting acceptance with the TXOP allocated in the third step, and FIG. 3 illustrates an admission control and scheduling process in the present invention.

In the present embodiment, SI is allocated differently for each station, unlike the reference scheduler assigned to all stations. The SI of each station is assigned through the admission control algorithm and selected from among predefined values before the station is included in the polling list. Table 1 is an example of an SI table that defines a value that can be assigned to an SI.

Basic SI (BSI) is the minimum SI that can be scheduled and the values of the SI table are defined as a multiple of the previous SI starting from the BSI. By defining the SI values as multiples of the previous SI relative to the BSI (where m = 2 in Equation 4), transmissions between stations polled in the BSI are bursty and simplify scheduling and admission control algorithms.

The scheduler assigns SI to the possible similar values among the smaller values than the MSI of the station in the SI table. For example, if the MSIs of stations A, B, C, and D are 25ms, 45ms, 50ms, and 90ms, respectively, the SIs allocated are 20, 40, 40, and 80ms. 4 shows a comparison of SI using a reference scheduler and an SI table. Referring to FIG. 4, it can be seen that the polling overhead is greatly reduced when the number of polls is greatly reduced in the present invention than the reference scheduler. You can also see that SI is assigned differently in stages according to the MSI of the stations, and the TXOP allocated at once is also larger.

In this embodiment, in order to schedule a new station requesting acceptance, the SI of the station must first be determined, and a polling start time must be determined according to the determined SI. The reason for this is that even though the system has the same SI as shown in FIG. 5, the scheduling is changed according to the polling start time. To determine the polling start time, the sum of the TXOPs of each polling group and the TXOP of the station requesting acceptance should be calculated with the determined SI, and then the acceptable polling start time should be checked. In this embodiment, in order to easily perform such calculations, a scheduling tree as shown in FIG. 6 is constructed, and the HC performs scheduling and admission control by maintaining this tree. In FIG. 6 the lowercase alphabet above each node is the node name and the uppercase alphabet inside indicates the accepted stations. In the scheduling tree, the SIs of the stations belonging to the root node are BSI, and as the level is increased by one level, the SIs of the stations belonging to the node are doubled. That is, in FIG. 7, SIs of stations B and C belonging to nodes b and c are 2BSI, and SIs of stations D, E, F, and G belonging to nodes d, e, f, and g are 4BSI. There may be several stations included in the node, and the level of BSI and tree may be adjusted to suit the network. Stations belonging to the leaf node and its ancestor nodes in the scheduling tree are always polled within the same BSI. We define these stations as polling groups. For example, in Figure 6 A, B, D stations are stations that are always polled within the same BSI. The number displayed under the leaf nodes in the scheduling tree is the number of each polling group and indicates the order in which the polls are polled. That is, when the scheduling tree is constructed as shown in FIG. 6, each of the BSI includes (A, B, D), (A, C, F), (A, B, E), (A, C, G), (A, B, Polling is done in D) order. 7 shows an example of the polling procedure.

In the reference scheduler, when the new station is accepted and added to the polling list or the existing station transmits traffic and exits the polling list, the SI and TXOPs of the stations change. In other words, when the smallest MSI among the MSIs of the stations in the polling list changes, a new SI must be allocated and thus a new TXOP must be allocated to all stations. However, in this embodiment, when a new station joins or exits the network, it does not affect the SI or TXOP values of other stations.

In the present embodiment, the admission control is a process of adding a new station to the scheduling tree through the following three steps.

STEP 1: The SI of the station i requesting acceptance is calculated as in Equation 7, i.e., the SI closest to the MSI is smaller than the MSI.

STEP 2: Obtain the TXOPSUM of the TXOPs of each polling group and the TXOP of the station i requesting acceptance. TXOPSUM has been shown in equations 5a, 5b and 5c. Where l represents the number of stations included in the polling group.

The method of obtaining the TXOP of the station i is the same as the reference scheduler except the SI. The number of MSDUs and TXOP _i arriving at the MAC layer at Mean Data Rate during SI _i at station i are also calculated through Equations 5a, 5b and 5c.

STEP 3: Finally, in the scheduling tree, the node i belongs to is determined. If there is a node that may include station i, the node is accepted and the conditions are as follows. When the SI includes station i among nodes whose SI _i , all polling groups to which the node belongs must satisfy the comparison in Equations 5a, 5b and 5c.

If there is more than one node that can include station i, select a node that includes a polling group with a minimum TXOPSUM. If there is no polling group that station i can join and SI _i is greater than BSI, reduce SI _i by half (divide by m) and repeat from step 2. Station i is rejected if there is no group available to join until SIi becomes BSI. The meaning of comparison in STEP 3 is as follows. Since stations in a polling group are all polled within a BSI, the total number of TXOPs of a new station and the total TXOP of that polling group must be less than or equal to the BSI before a new station is included in a particular polling group.

8 shows a scheduling tree before admission control of a station is performed. When the scheduling tree is constructed as shown in FIG. 8, the number in parentheses next to the uppercase letters of the alphabet indicates the TXOP assigned to the station. At this time, two stations K and L, whose MSI is 85ms and 80ms, respectively, request the acceptance.

First, the SI of the station K becomes 80 ms by the equation (7) and the TXOP is 4 ms by the equations (5a, 5b and 5c). At this time, nodes having an SI of 80ms are d, e, f, and g, and the total TXOPs of polling groups 1, 2, 3, and 4 are 17ms, 17ms, 18ms, and 13ms, respectively. Since only the polling group 4 satisfies the comparison in Equations 5a, 5b and 5c, the node K may be included in is a node. 9 is a reconstructed scheduling tree after station K has been accepted.

Secondly, the SI of station L is 80ms and TXOP is 5ms. In this case, nodes having an SI of 80 ms are d, e, f, and g, and the total TXOPs of polling groups 1, 2, 3, and 4 are 17 ms, 17 ms, 18 ms, and 17 ms, respectively. Since all polling groups do not satisfy the comparison in Equations 5a, 5b and 5c, there are no nodes where station L can be included. Therefore, reduce SI of station L to 40ms and repeat the above process. When SI is 40ms, TX L of station L becomes 3ms. At this time, since the nodes SI and 40ms are b and c, the polling groups 2 and 4 to which node c belongs satisfy the comparison expressions in Equations 5a, 5b, and 5c, and thus, station L is accepted and included in node c. 10 is a scheduling tree after station L has been accepted.

In this embodiment, the new station is accepted through the admission control and the scheduling is completed. This is because the new station is inserted into the scheduling tree through the admission control to configure the scheduling tree and to perform the scheduling of the station.

Next, the performance of this embodiment is compared with a reference scheduler using a network simulator (NS) -2, which is a network simulation program. Simulation assumptions for performance analysis of the present invention are as follows. The physical layer used for the simulation is IEEE 802.11b, and the related parameters are shown in Table 2. The MAC layer uses the IEEE 802.11e HCCA module, and the channel state is assumed to be ideal. RTS / CTS frame is not used and block ACK is used. Each station has one uplink traffic stream to send to the access point.

Slot Time 20 μs SIFS 10 μs PIFS 30 μs Preamble Length 144 bits PLCP Header Length 48 bits PLCP Tx Rate 1 Mbps MAC Header Length 60 Bytes Basic Tx Rate 1 Mbps Data rate 11 Mbps

Simulations were performed with traffic for scenario 1 and scenario 2 in Table 3, respectively. In scenario 1, traffic with the same frame interval as MSI, that is, one frame per MSI, was used. In Scenario 2, on the other hand, traffic with multiple packets per MSI was used.

Type Frame Interval (ms) Mean Frame Size (byte) Mean Data Rate (Kbps) MSI (ms) Delay Bound (ms) Voice 1 Scenario 1 20 120 48 20 40 Scenario 2 20 60 48 20 40 Video 1 Scenario 1 40 1280 256 40 80 Scenario 2 20 1280 512 40 80 Voice 2 Scenario 1 40 120 24 40 80 Scenario 2 20 60 48 40 80 Video 2 Scenario 1 80 1280 128 80 160 Scenario 2 20 1280 512 80 160

The VBR traffic used for the simulation used h.261 video file. The simulations were performed for 100 minutes, and the total transmission rate, HCCA channel occupancy time ratio, and packet loss rate due to transmission delay were measured by increasing the stations with traffic of scenario 1 and scenario 2 shown in Table 3.

In FIG. 11, as the number of stations increases in scenario 1, the total transmission rates of the present invention and the reference scheduler are compared. Until the number of stations is ten, the total transmission rates of the present invention and the reference scheduler are increased equally. At this time, there was no packet loss due to transmission delay in both the present invention and the reference scheduler. Therefore, the total data rate is the same. However, when there are more than 11 stations, the reference scheduler does not increase the transmission rate anymore and only increases in the case of the present invention until there are 20 stations. This is because the reference scheduler can accept up to 10 stations and the present invention can accept up to 20 stations.

In scenario 1, the reason why the present invention is accepted twice as many as the reference scheduler is as follows. Firstly, the traffic of scenario 1 has 20ms, 40ms, and 80ms frame intervals, but the reference scheduler allocates TXOP to all stations every 20ms and guarantees the minimum time to send one MSDU of maximum size. As a result, unnecessary TXOP is allocated as compared with the present invention. Second, the present invention reduces the overhead due to polling because the number of polling is reduced than the reference scheduler.

12 illustrates HCCA channel occupancy of the present invention and the reference scheduler in scenario 1. FIG. As stations increase, the rate of increase of the present invention increases more slowly. When there are 10 stations, the reference scheduler is 95.8% and in the present invention, it is almost 1/2, 49.5%. Since the time remaining without being used as HCCA can be used as a competition section, the transmission efficiency can be further improved in the present invention.

13 and 14 illustrate the total transmission rate and the HCCA channel occupancy rate of the present invention and the reference scheduler as the number of stations increases in scenario 2. In scenario 2, the difference between the number of acceptable stations and the HCCA channel occupancy between the present invention and the reference scheduler is reduced. This is because in scenario 2, since the frame intervals of the traffic are all 20ms, the waste due to excessive TXOP allocation of the reference scheduler is reduced unlike in scenario 1.

It is possible to apply to a WLAN that requires scheduling, and in particular, it is advantageous to apply to IEEE802.11x such as IEEE802.11e.

1 is a block diagram showing a scheduling apparatus of a wireless LAN according to an embodiment of the present invention.

2 is a block diagram according to another embodiment of FIG.

3 is a flowchart illustrating an admission control and traffic scheduling process according to the embodiment of FIG.

4 is an explanatory diagram showing a reference scheduler and an SI (Service Interval) of the present invention;

5 is a structural diagram showing scheduling according to SI and polling time point.

6 is a structural diagram showing a scheduling tree in which stations are served.

7 is an explanatory diagram showing a scheduling sequence of FIG. 4;

8 is a diagram of a scheduling tree structure before acceptance of stations wishing to join the network.

9 is a scheduling tree structure diagram illustrating stations accepted in FIG. 6;

10 is a structural diagram of another example illustrating a scheduling tree after a station is accepted.

FIG. 11 is a graph comparing transmission rates of the present invention and a reference scheduler in simulation scenario 1. FIG.

12 is a graph comparing HCCA channel occupancy ratio of the present invention and a reference scheduler in simulation scenario 1. FIG.

FIG. 13 is a graph comparing transmission rates of the present invention and a reference scheduler in simulation scenario 2. FIG.

14 is a graph comparing HCCA channel occupancy ratio of the present invention and a reference scheduler in simulation scenario 2. FIG.

* Description of the symbols for the main parts of the drawings *

10 ... SI table setting section 20 ... SI allocation section

30 ... Acceptance Judgment Department

Claims (9)

An SI table setting unit for setting an SI table including a minimum service interval (SI) that can be scheduled; And And an SI allocator for allocating an SI (approximate SI) closest to an MSI (Maximum Service Interval) of each station among the SIs in the SI table to each station. The method according to claim 1, The SI table includes a minimum SI that can be scheduled; And Scheduling apparatus of a wireless LAN which comprises a; SI _k regular increments for the minimum SI applied. The method according to claim 2, SI _k is the following equation SI _k = m × SI _(k-1) SI _(k-1) : previous SI _k (SI ₀ is minimum SI) m: real> 1 k: positive integer Wireless LAN scheduling apparatus characterized in that it satisfies. The method according to claim 3, M is 2, the scheduling apparatus of the wireless LAN. The method according to claim 3, A transmission determining unit for calculating transmission opportunity (TXOP) of each station using the allocated SI and determining whether to accept participation of each station based on the TXOP; Device. The method according to claim 5, The acceptance determination unit TXOP (TXOP _i ) of each station is the following equations 5a, 5b and 5c

Equation 5a

[Equation 5b]

[Equation 5c] here,

SI _i : SI of the station ρ _i : Mean Data Rate L _i : Nominal MSDU (MAC Service Data Unit) R _i : Physical Transmission Rate O: Polling overhead (sum of QS CF-Poll frame transmission time, ACK receiving time for data frame, Inter Frame Space time) TXOP _n : TXOP of each station in each polling group TXOPSUM: Total TXOP for each polling group BSI: minimum SI l: the number of stations included in the polling group M: maximum size of MSDU If it satisfies, the scheduling apparatus of the WLAN, characterized in that for accepting each station. The method according to claim 6, When the TXOP of each station does not satisfy the equations 5a, 5b, and 5c, the acceptance determining unit converts the SI _i to the following equation (6). SI _i = SI _(i-1) / m After resetting by applying, repeating the process of applying the equations (5a, 5b and 5c) to satisfy the equations (5a, 5b and 5c) to accept the polling group of the station to which the SI _i is assigned, the SI _{If the i} is not satisfied until the BSI, the scheduling apparatus of the WLAN, characterized in that rejecting the acceptance. The method according to claim 6, And the acceptance determining unit accepts the participation of each station as a polling group having a minimum TXOPSUM when there are a plurality of polling groups satisfying the TXOP of each station. The method according to claim 1, The SI allocator is smaller than the MSI of each station, and allocates the SI closest to the MSI of each station to each station, characterized in that the WLAN.

Images