CN101286949A - Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard - Google Patents

Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard Download PDF

Info

Publication number
CN101286949A
CN101286949A CNA2008101145905A CN200810114590A CN101286949A CN 101286949 A CN101286949 A CN 101286949A CN A2008101145905 A CNA2008101145905 A CN A2008101145905A CN 200810114590 A CN200810114590 A CN 200810114590A CN 101286949 A CN101286949 A CN 101286949A
Authority
CN
China
Prior art keywords
service
node
formation
request
business
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CNA2008101145905A
Other languages
Chinese (zh)
Inventor
李旭
申宏刚
石靖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Jiaotong University
Original Assignee
Beijing Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Jiaotong University filed Critical Beijing Jiaotong University
Priority to CNA2008101145905A priority Critical patent/CN101286949A/en
Publication of CN101286949A publication Critical patent/CN101286949A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Mobile Radio Communication Systems (AREA)

Abstract

The invention relates to a wireless Mesh network MAC layer resource scheduling strategy based on standard of IEEE802.16d, which solves the problems of fairness in single service resource scheduling and the quality of service (QoS) during multi-service resource scheduling to certain extent. When only the single service exists in the network, every node establishes two types of queues, namely, a request queue and an authorization queue; round-robin scheduling towards each queue ensures fair resource occupation of each type of service flow, which comprises bandwidth request/authorization mechanism, scheduling mechanism, time slot distribution mechanism and compensation mechanism; under the circumstance of multi-service, different priorities can be set to various different service flows according to the differences of service real-time request so as to distribute bandwidth dynamically and improve the throughput of the system as far as possible with the premise of meeting the quality of service (QoS) of each type of service flow, which comprises classification mechanism, bandwidth reservation mechanism and degrading mechanism.

Description

Wireless Mesh network MAC layer resource scheduling policy based on the IEEE802.16d standard
Technical field
The invention belongs to the Wireless Mesh network communications field, based on the IEEE802.16d standard, the resource dispatching strategy that is proposed at the MAC layer of wireless Mesh netword.
Background technology
The quality of mac-layer protocol directly has influence on the quality of performance index such as network throughput, time delay.How solving a plurality of users and efficiently, reasonably share limited wireless channel resource problem, is the problem that mac-layer protocol will solve.And along with the class of business that Wireless Mesh is supported is more and more, service quality (QoS) difference that different business requires, the different QoS requirements that how to guarantee the different business kind are the problems that the scheduling of MAC layer needs research and solve.
Though in IEEE 802.16-2004 standard the mac-layer protocol of Mesh pattern is stipulated, this mac-layer protocol is not elaborated to bandwidth scheduling, therefore has very big research space.
By designing a kind of resource dispatching strategy, make under the situation of single service source, use a kind of end-to-end dispatching method of justice to carry out scheduling of resource, make the bandwidth resources of using network between the Business Stream liberally.Under multiple services situation, difference according to professional real-time requires can be provided with different priority to various Business Stream, uses a kind of distribution method of dynamic bandwidth, make and under the prerequisite of the QoS that satisfies every kind of Business Stream, improve the throughput of system as far as possible.
Summary of the invention
The objective of the invention is to make Wireless Mesh network under single professional situation, fairness preferably to be arranged, under multi-service circumstances, under the prerequisite of the QoS that satisfies every kind of Business Stream, improve the throughput of system as far as possible.
The technical solution used in the present invention is:
When having only one type professional in the network, in order to guarantee the fairness between every Business Stream, adopt a kind of end-to-end resource regulating method of justice, purpose is to reduce this problem for the resolution system throughput increases with jumping figure.
Fair end-to-end resource regulating method is dispatched the bandwidth between the neighbor node, to reach the occupied bandwidth that makes each end-to-end Business Stream justice.
The enforcement of this dispatching method is divided into four kinds of mechanism: bandwidth request/licensing scheme, scheduling mechanism, time slot allocation mechanism and compensation mechanism.Three prerequisite hypothesis that this method proposes:
1) network topology is fixed.
2) link has enough reliabilities.
3) single wireless radio interface.
Described bandwidth request/licensing scheme particular content comprises: the service data unit that receives from the upper strata (SDU) is sent to the packet scheduler, the grouping that each packet scheduler selection will transmit, be packaged into protocol Data Unit (PDU), be sent to physical layer then.In each packet scheduler, different traffic packets is sent in the different buffer memorys, adopts the scheduling of deficit poll (DRR) scheduling mode.
To give a definition, as long as produce a SDU, send to destination node from source node, then think a Business Stream from the source node to the destination node.Arbitrary nodes X in the path for this Business Stream, carry out following two kinds of situations:
A) when arbitrary nodes X receives this SDU, search its effective stream tabulation, if the record of this Business Stream not in the tabulation then adds this business and flows to its effective tabulation.
B) when arbitrary nodes X does not also receive the SDU of this stream in certain official hour, then arbitrary nodes X is effectively deleted this Business Stream from it the tabulation.
For arbitrary nodes X, safeguarding two Class Queue: request queue and mandate formation.
Request queue: if there is Business Stream to flow to its neighbor node Y from arbitrary nodes X, then arbitrary nodes X is being safeguarded a virtual request queue to its neighbor node Y, comprises following state variable in the request queue:
Req Y Out: arbitrary nodes X sends request by solicited message unit (request IEs) to its neighbor node Y, notifies its byte number that will transmit;
Cnf Y Out: arbitrary nodes X confirms that its neighbor node Y distributes to its byte number;
Blog Y Out: arbitrary nodes X sends the byte number of waiting for to its neighbor node Y, because the bandwidth request message that the data of these waits send is not processed;
If b log Y out = 0 , Think that then request queue is invalid, this formation is deleted from effective tabulation.
Define a variable pending again, this variable calculates suc as formula 1-1.Be threshold value of this variable-definition---pending again Max, when surpassing threshold value, the request that shows requesting node can not be authorized to node and satisfy.
pending = req Y out - cnf Y out - - - ( 1 - 1 )
Authorize formation: if there is Business Stream to flow to arbitrary nodes X from its neighbor node Y, then X is safeguarding a virtual mandate formation to its neighbor node Y, and this formation comprises following state variable:
Req Y In: neighbor node Y sends request by solicited message unit request IEs to arbitrary nodes X, notifies its byte number that will transmit;
Gnt Y In: the byte number that arbitrary nodes X has been authorized to its neighbor node Y;
If req Y in = gnt Y in , Then this formation is deleted from effective tabulation.
Arbitrary nodes X safeguards that to the Business Stream that each flows to oneself is authorized a formation, to provide resource to the Business Stream of sending from its neighbor node.When it is sent to other nodes with this work flow, for this Business Stream is safeguarded a request queue, with the node request resource that will transmit to it.For each formation, all distributed a weighted value φ, the calculating of this weighted value is suc as formula 1-2.
φ i = Σ j ∈ A I i ( j ) | A | - - - ( 1 - 2 )
Wherein, | A| represents to flow into and flow out the summation of the professional fluxion of arbitrary nodes X, and j is effective stream, I i(j) be the indication function, if when j is service under formation i, the value of indicator function is 1, otherwise is 0.Because each Business Stream obtains service so ∑ under a formation of determining iφ i=1.Bandwidth request and licensing scheme are applicable to all effective requests and authorize formation.
Described scheduling mechanism particular content comprises: these formations are serviced by polling mode, and the serviced byte number of formation is Ф in each is taken turns iF RR, F wherein RRBe system parameters, be called and take turns the duration.In distributed scheduling message (MSH-DSCH), each authorizes formation i to have the right to authorize above Ф iF RRByte number give the neighbor node of i, each request queue i request of having the right simultaneously surpasses Ф iF RRByte number give the neighbor node of i.If be less than Ф from the byte number of the neighbor node request of i/authorize iF RR, this formation will be removed from effective tabulation after service.
A kind of mode of poll is adopted in formations all in effective tabulation, till following condition occurring:
1) effectively tabulation is sky;
2) information unit (IE) that does not have remaining time slot to increase other in control frame is given distributed scheduling message MSH-DSCH;
3) the compensating parameter lag value of not serviced formation has surpassed threshold value lag Max
Described time slot allocation mechanism particular content comprises:
1) selects a channel at random;
2) find out first available time slot of first frame;
3) if there is not available time slot, forward next channel to;
4) if all time slots are all searched to be arrived, forward next frame to.
At first, the visit of time slot is that order is a time sequencing, can reduce the time that bandwidth is consulted; Second, the node of double bounce of being separated by is not known the distributed scheduling message MSH-DSCH that the other side sends, they can authorize identical time slot to give the neighbor node that is not identified, and these neighbor nodes are on same channel, and selective channel reduces the probability that this incident takes place at random; At last, the same frame under same channel authorizes many more continuous slots can reduce authorization message unit number among the distributed scheduling message MSH-DSCH.
Described compensation mechanism particular content comprises: a formation not serviced byte number in the scheduling that takes turns is represented with compensating parameter lag.When carrying out the next round scheduling, it can be in these some bandwidth of using with the compensation scheduling of last time more.Compensating parameter lag should be provided with one and go up dividing value, and is not serviced for a long time to prevent this formation, causes the phenomenon of " dying of hunger ".
When polytype Business Stream is arranged in the network, according to the service quality QoS characteristic of every kind of Business Stream, utilize the method for degradation model, adopt a kind of distribution method of dynamic bandwidth, on the basis of the service quality QoS that guarantees every kind of Business Stream, improve the throughput of system as much as possible.
Distribution method of dynamic bandwidth comprises following mechanism: classification mechanism, bandwidth reservation mechanism and degradation mechanism.
Described classification mechanism particular content comprises: the four class business of stipulating in the IEEE802.16d standard are respectively unsolicited grant service (UGS), real-time polling service (rtPS), non-real-time polling service (nrtPS) and Best-Effort service (BE).And the priority of four class business is followed successively by from high to low: unsolicited grant service, real-time polling service, non-real-time polling service, Best-Effort service.According to the difference of service attribute,, with delineation of activities a kind of in the four class business of stipulating in the standard in conjunction with the service quality QoS of every kind of business.Each node maintenance four formations of corresponding four class business, and the grouping that receives from the upper strata enters grader, and grader is according to the information in the packets headers, grouping is delivered in the corresponding queues gone, and waits scheduling.
Described bandwidth reservation mechanism particular content comprises: the unsolicited grant service UGS with limit priority is reserved a part of resource, to guarantee the service quality of the Business Stream that real-time is had relatively high expectations.
Described degradation mechanism particular content comprises: non-real-time polling service nrtPS stream is provided with certain resource scope of application, represents that in this size with occupied bandwidth this scope is [b Nrtps Min, b Rtps Max].Be that to have the utilized bandwidth minimum value be b to non-real-time polling service nrtPS Nrtps Min, the utilized bandwidth maximum is b Nrtps MaxAnd be two variablees of such business setting: degradation step-length-δ and degradation progression-l.Satisfy relation between these four parameters suc as formula shown in the 1-3:
b nrtps max - lδ ≥ b nrtps min - - - ( 1 - 3 )
Occur under the not enough situation of resource when having unsolicited grant service UGS and the real-time polling service rtPS higher than non-real-time polling service nrtPS flow priority, will demote to non-real-time polling service nrtPS stream all in the network, promptly the resource that this business is used is cut down, each value of cutting down is δ, till satisfying high-priority service stream.When non-real-time polling service nrtPS is downgraded to minimum value b Nrtps MinThe time, even high-priority service flows when not being met, also no longer continue non-real-time polling service nrtPS is demoted.
Effect of the present invention is: by scheduling strategy of the present invention, make Wireless Mesh network under the situation of single service source, make the bandwidth resources of using network between the Business Stream liberally.Under multiple services situation, the difference according to professional real-time requires is provided with different priority to various Business Stream, makes and improve the throughput of system as far as possible under the prerequisite of the QoS that satisfies every kind of Business Stream.Therefore the present invention not only has good performance in the wireless Mesh netword of single Business Stream, and preferable performance is also arranged in the wireless Mesh netword of multi-business flow.
Description of drawings
Fig. 1 is based on the flow chart of the Wireless Mesh network MAC layer resource scheduling policy of IEEE802.16d standard;
Fig. 2 is bandwidth request and the mandate formation schematic diagram of safeguarding for its neighbor node in arbitrary nodes X;
Fig. 3 is a time slot allocation mechanism works principle schematic.
Embodiment
The present invention will be further described below in conjunction with accompanying drawing.
As shown in Figure 1, at first the Business Stream in the network is judged, judge that network is single business or multi-service, if during single business network, each node each neighbor node for it when meeting some requirements is being safeguarded two Class Queue: request queue and authorize formation.In order to guarantee the fairness between every Business Stream, adopt a kind of end-to-end resource regulating method of justice, purpose is the fairness that guarantees every Business Stream; If during multibusiness network, QoS characteristic according to every kind of Business Stream, the grouping that the upper strata receives is classified, enter respective queue respectively, each node maintenance four Class Queue: unsolicited grant service formation, real-time polling service formation, non-real-time polling service formation and Best-Effort service formation.Adopt a kind of distribution method of dynamic bandwidth, on the basis of the QoS that guarantees every kind of Business Stream, improve the throughput of system as much as possible.
I) Gong Ping end-to-end resource regulating method is dispatched the bandwidth between the neighbor node, makes each end-to-end Business Stream occupied bandwidth liberally to reach.
Fair end-to-end resource regulating method comprises following mechanism: bandwidth request/licensing scheme, scheduling mechanism, time slot allocation mechanism and compensation mechanism.
1. bandwidth request/licensing scheme
To give a definition, as long as produce a service data unit SDU, send to destination node from source node, then think a Business Stream from the source node to the destination node.Arbitrary nodes X in the path for this Business Stream, carry out following two kinds of situations:
A) when arbitrary nodes X receives this SDU, search its effective stream tabulation, if the record of this Business Stream not in the tabulation then adds this business and flows to its effective tabulation.
B) when arbitrary nodes X does not also receive the SDU of this stream in certain official hour, then X effectively deletes this Business Stream from it the tabulation.
As shown in Figure 2,6 node A, B, C, D, E, X being arranged in the network, 6 Business Streams are arranged, is respectively E → X, E → B, E → D, C → X, C → B, B → C (shown in the empty arrow among Fig. 2).For arbitrary nodes X, for its neighbours are safeguarding 5 formations altogether, be respectively to safeguard that to neighbor node A is authorized a formation G, safeguarding two formations respectively to neighbor node B and C, request queue R and mandate formation G.
Now arbitrary nodes X is described the request queue R that its neighbor node C safeguards, is at first safeguarding following state variable in the request queue in conjunction with Fig. 2:
Req C Out: arbitrary nodes X sends request by request IEs to its neighbor node C, notifies its byte number that will transmit;
Cnf C Out: arbitrary nodes X confirms that its neighbor node C distributes to its byte number;
Blog C Out: arbitrary nodes X sends the byte number of waiting for to its neighbor node C, because the bandwidth request message that the data of these waits send is not processed;
If b log C out = 0 , Think that then request queue is invalid, this request queue is deleted from effective tabulation.
Define a variable pending again, this variable calculates suc as formula 1-1.Be threshold value of this variable-definition---pending again Max, when surpassing threshold value, show that the request of requesting node X can not be satisfied by its neighbor node C.
pending = req C out - cnf C out - - - ( 1 - 1 )
In like manner, arbitrary nodes X also comprises above like variable for the request queue that its neighbor node B safeguards.
Now arbitrary nodes X is described the mandate formation G that its neighbor node B safeguards, is at first safeguarding following state variable in this mandate formation in conjunction with Fig. 2:
Req B In: neighbor node B sends request by request IEs to arbitrary nodes X, notifies its byte number that will transmit;
Gnt B In: neighbor node B to arbitrary nodes X mandate byte number;
If req B in = gnt B in , Then this mandate formation is deleted from effective tabulation.
In like manner, arbitrary nodes X also comprises above like variable for the mandate formation of other neighbor nodes A and C maintenance.
Arbitrary nodes X safeguards that to the Business Stream that each flows to oneself is authorized a formation, to provide resource to the Business Stream of sending from its neighbor node.When it is sent to other nodes with this work flow, for this Business Stream is safeguarded a request queue, with the node request resource that will transmit to it.For each formation, all distributed a weighted value φ, the calculating of this weighted value is suc as formula 1-2.
φ i = Σ j ∈ A I i ( j ) | A | - - - ( 1 - 2 )
| A| represents to flow into and flow out the summation of the professional fluxion of arbitrary nodes X, and j is effective stream, I i(j) be the indication function, if when j is service under formation i, the value of indicator function is 1, otherwise is 0.Because each Business Stream obtains service so ∑ under a formation of determining iФ i=1.Bandwidth request and licensing scheme are applicable to all effective requests and authorize formation.
Owing to flow into and the Business Stream sum that flows out is 10, therefore calculate the denominator in the computing formula of weight of formation from arbitrary nodes X | A|=10.For the request queue of safeguarding for neighbor node C, owing to have only 1 from the Business Stream of X → C, so the molecule of the calculating formula of the weight of this formation is 1.So arbitrary nodes X is 1/10 for the weighted value of the request queue that its neighbor node C safeguards.In like manner, the weighted value of other formations also can draw according to said method: arbitrary nodes X is 3/10 for the mandate formation weighted value that its neighbor node A safeguards; Arbitrary nodes X is 3/10 for the request queue weighted value that its neighbor node B safeguards; Arbitrary nodes X is 2/10 for the weighted value of the mandate formation that its neighbor node C safeguards; Arbitrary nodes X is 1/10 for the weighted value of the mandate formation that its neighbor node B safeguards.
2. scheduling mechanism
Request queue and mandate formation are serviced by polling mode.The serviced byte number of formation is Ф in each is taken turns iF RR, F wherein RRBe system parameters, be called and take turns the duration.In distributed scheduling message MSH-DSCH, each authorizes formation i to have the right to authorize above Ф iF RRByte number give the neighbor node of i, each request queue i request of having the right simultaneously surpasses Ф iF RRByte number give the neighbor node of i.If be less than Ф from the byte number of the neighbor node request of i/authorize iF RR, this formation will be removed from effective tabulation after service.
A kind of mode of poll is adopted in formations all in effective tabulation, till following condition occurring:
1) effectively tabulation is sky;
2) the information unit IE that does not have remaining time slot to increase other in control frame gives distributed scheduling message;
3) there is not the lag value (in compensation mechanism, defining) of serviced formation to surpass threshold value lag Max
Below be the request queue working mechanism false code of safeguarding to neighbor node i:
lag i out ← min { lag i out + φ i F RR , b log i out }
if ( req i out - cnf i out > pending max )
if ( lag i out > lag max ) Stop service
else
req(lag i out)
req i out ← req i out + needed
lag i out ← 0
b log i out ← b log i out - needed
if ( b log i out = 0 ) This formation is deleted from effective tabulation
In the request queue, if the difference of the number of resources of each requested resource number and affirmation is greater than the threshold value pending of pending Max, judge on this basis whether the value of lag has surpassed threshold value, if surpass, it is serviced to illustrate that then this formation does not have for a long time, then stops this formation.If the value of pending does not surpass threshold value, according to polling dispatching, the normal request bandwidth, the quantity of request is proportional to the weight of formation, and request IE is joined in the MSH-DSCH message.If the value of blog equals 0, illustrate that then each requested resource can be satisfied fully, the resource that does not wait in the request queue, so this request queue is deleted from effective tabulation.
Below be the mandate queue work mechanism false code of safeguarding to neighbor node i:
lag i in ← min { lag i in + φ i F RR , req i in - gnt i in }
if ( lag i in > 0 )
lag i in = last lag i in
gnt i in ← gnt i in + granted
lag i in ← lag i in - granted
if ( lag i in > lag max ) , Stop this formation of service
if ( req i in = gnt i in ) This formation is deleted from effective tabulation
Authorize in the formation, in the scheduling of last time, when a formation does not have servicedly because of a variety of causes, so it does not have serviced byte number to represent with variable lag in last scheduled.When this is dispatched, it can be in these some bandwidth of using with the compensation scheduling of last time more, and the value of these bandwidth of using equals should serve last time and the resource that do not obtain serving more.Grant IE adds in the MSH-DSCH message with the authorization message unit.If the value of lag has surpassed threshold value, illustrate that this formation has not obtained having served for a long time.Stop this formation.If the resource of authorizing all equals each requested resource, it is empty then should authorizing formation, then this formation is deleted from effective tabulation.
3. time slot allocation mechanism
1) selects a channel at random;
2) find out first available time slot of first frame;
3) if there is not available time slot, forward next channel to;
4) if all time slots are all searched to be arrived, forward next frame to.
At first, the visit of time slot is that order is a time sequencing, therefore can reduce the time that bandwidth is consulted.The second, the node of the double bounce of being separated by is not known the MSH-DSCH message that the other side sends.Therefore, they may authorize identical time slot to give the neighbor node that is not identified, and these neighbor nodes are on same channel.Selective channel can reduce the probability that this incident takes place at random.At last, the same frame under same channel authorizes many more continuous slots can reduce authorization message unit number in the MSH-DSCH message.
As shown in Figure 3, suppose to have two channels: channel 1 and channel 2, each square frame are represented a time slot, the occupied time slot of square frame representative of band fork, the order that the curve representation time slot of band arrow is searched.At this moment, if certain node will be searched 3 available time slots.At first channel of picked at random supposes to have chosen channel 1, chooses first frame in channel 1, whether first time slot of searching this frame is available, finding first time slot occupied (beating the grid of fork), whether available, down look for successively if continuing to search second time slot.Find that the 3rd time slot and the 4th time slot of first frame can use, keep.With<channel, frame, time slot〉representation represents this two available time slots:<1,1,3 〉,<1,1,4 〉.Then forward second channel to, search first frame of channel 2, search in the time slot method cochannel 1 the same, find that the 4th time slot is idle, be expressed as<2,1,4 〉, but because this time slot and just now the 4th time slot in first frame in the channel 1 be (promptly all in first time slot<1,4 of first frame 〉) of conflicting in time, so this time slot is also unavailable.Then forward second frame of channel 1 to, it is idle finding first time slot, be expressed as<1,2,1, and and preceding two time slots of having searched do not clash in time, therefore select this time slot.So far, this node finds 3 available time slot usefulness<channels, frame, time slot〉be expressed as:<1,1,3 〉,<1,1,4〉and,<1,2,1 〉.
4. compensation mechanism
A formation not serviced byte number in the scheduling that takes turns is represented this variable called after compensating parameter with variable lag.When carrying out the next round scheduling, it can be in these some bandwidth of using with the compensation scheduling of last time more.Variable lag should be provided with one and go up dividing value, and is not serviced for a long time to prevent this formation, causes the phenomenon of " dying of hunger ".
The request queue that arbitrary nodes X is as shown in Figure 2 safeguarded for neighbor node C, compensating parameter a: lag is being safeguarded in this formation C OutThe mandate formation that arbitrary nodes X is safeguarded for neighbor node B, compensating parameter a: lag is being safeguarded in this formation B InIn like manner, also safeguarding this compensating parameter lag in other formations.
Ii) distribution method of dynamic bandwidth comprises following mechanism: classification mechanism, bandwidth reservation mechanism and degradation mechanism.
1. classification mechanism
The four class business of stipulating in the IEEE802.16d standard are respectively: unsolicited grant service UGS, real-time polling service (rtPS), non-real-time polling service (nrtPS) and Best-Effort service (BE).And the priority of four class business is followed successively by from high to low: UGS, rtPS, nrtPS, BE.According to the difference of service attribute,, with delineation of activities a kind of in the four class business of stipulating in the standard in conjunction with the QoS of every kind of business.Each node maintenance four formations of corresponding four class business, and the grouping that receives from the upper strata enters grader, and grader is according to the information in the packets headers, grouping is delivered in the corresponding queues gone, and waits scheduling.
2. bandwidth reservation mechanism
To a part of resource of the professional reservation of the UGS with limit priority, to guarantee the service quality of the Business Stream that real-time is had relatively high expectations.
3. degradation mechanism
The nrtPS Business Stream is provided with certain resource scope of application, represents that in this size this scope is [b with occupied bandwidth Nrtps Min, b Nrtps Max].Be that to have the utilized bandwidth minimum value be b to nrtPS Nrtps Min, the utilized bandwidth maximum is b Nrtps MaxAnd be two variablees of such business setting: degradation step-length-δ and degradation progression-l.Satisfy relation between these four parameters suc as formula shown in the 1-3:
b nrtps max - lδ ≥ b nrtps min - - - ( 1 - 3 )
Below be the false code of degradation mechanism:
Figure A20081011459000182
{
l++;
b nrtPS=b nrtPS-lδ;
if(b needUGS>b UGS) b UGS=b UGS+δ;
else b rtPS=b rtPS+δ;
}
Occur under the not enough situation of resource when having the Business Stream higher (UGS, rtPS) than nrtPS Business Stream priority, will demote to nrtPS Business Streams all in the network, promptly the resource that this business is used is cut down, each value of cutting down is δ, till satisfying high-priority service stream.When the nrtPS business is downgraded to minimum value b Nrtps MinThe time, even high-priority service flows when not being met, also no longer continue nrtPS is demoted.
The present invention simulates on NS2 network simulation platform based on the IEEE802.16d standard.Analog result shows, if when having only one type professional in the network, adopts fair end-to-end dispatching method can well solve the unjustness of Business Stream because of the different services that produce of jumping figure.When the multiclass service resources was arranged in the wireless Mesh netword, the character according to different business stream was provided with priority, and dynamic utilized bandwidth has guaranteed that the QoS of Business Stream has improved the throughput of network simultaneously.

Claims (2)

1. based on the Wireless Mesh network MAC layer resource scheduling policy of IEEE802.16d standard, it is characterized in that: when having only one type professional in the network, adopt a kind of end-to-end resource regulating method of justice, the enforcement of this dispatching method is divided into four kinds of mechanism: bandwidth request/licensing scheme, scheduling mechanism, time slot allocation mechanism and compensation mechanism;
Described bandwidth request/licensing scheme particular content comprises:
The service data unit that receives from the upper strata is sent to the packet scheduler, the grouping that each scheduler selection will transmit, be packaged into protocol Data Unit, be sent to physical layer then, in each packet scheduler, different traffic packets is sent in the different buffer memorys, adopts deficit polling dispatching mode to dispatch;
To give a definition, as long as produce a service data unit, send to destination node from source node, then think a Business Stream from the source node to the destination node, arbitrary node (X) in the path for this Business Stream, carry out following two kinds of situations:
A) when arbitrary node (X) receives this service data unit, search its effective stream tabulation, if the record of this Business Stream not in the tabulation then adds this business and flows to its effective tabulation;
B) when arbitrary node (X) does not also receive the service data unit of this stream in certain official hour, then arbitrary node (X) is effectively deleted this Business Stream from it the tabulation;
For arbitrary node (X), safeguarding two Class Queue: request queue and mandate formation;
Request queue: if there is Business Stream to flow to its neighbor node (Y) from arbitrary node (X), then arbitrary node (X) is being safeguarded a virtual request queue to its neighbor node (Y), comprises following state variable in the request queue:
Req Y Out: arbitrary node (X) sends request by the solicited message unit to its neighbor node (Y), notifies its byte number that will transmit;
Cnf Y Out: arbitrary node (X) confirms that its neighbor node (Y) distributes to its byte number;
Blog Y Out: arbitrary node (X) sends the byte number of waiting for to its neighbor node (Y), because the bandwidth request message that the data of these waits send is not processed;
If b log Y out = 0 , Think that then request queue is invalid, this formation is deleted from effective tabulation;
Define a variable pending again, this variable calculate suc as formula:
pending = req Y out - cnf Y out
The threshold value of this variable is pending Max, when surpassing threshold value, the request that shows requesting node can not be authorized to node and satisfy;
Authorize formation: if there is Business Stream to flow to arbitrary node (X) from its neighbor node (Y), then arbitrary node (X) is being safeguarded a virtual mandate formation to its neighbor node (Y), and this formation comprises following state variable:
Req Y In: its neighbor node (Y) sends request by the solicited message unit to arbitrary node (X), notifies its byte number that will transmit;
Gnt Y In: the byte number that arbitrary node (X) has been authorized to its neighbor node (Y);
If req Y in = gnt Y in , Then this formation is deleted from effective tabulation;
Arbitrary node (X) safeguards that to the Business Stream that each flows to oneself is authorized a formation, to provide resource to the Business Stream of sending from its neighbor node, when it is sent to other nodes with this work flow, for this Business Stream is safeguarded a request queue, with the node request resource that will transmit to it, for each formation, all distributed a weighted value φ, the calculating of this weighted value suc as formula:
φ i = Σ j ∈ A I i ( j ) | A |
Wherein, | A| represents to flow into and flow out the summation of the professional fluxion of arbitrary node (X), and j is effective stream, I i(j) be the indication function, if when j is service under formation i, the value of indicator function is 1, otherwise is 0; Because each Business Stream obtains service so ∑ under a formation of determining iФ i=1, bandwidth request and licensing scheme are applicable to all effective requests and authorize formation;
Described scheduling mechanism particular content comprises:
Request queue and mandate formation are serviced by polling mode, and the serviced byte number of formation is Ф in each is taken turns iF RR, F wherein RRBe system parameters, be called and take turns the duration; In distributed scheduling message, each authorizes formation i to have the right to authorize above Ф iF RRByte number give the neighbor node of i, each request queue i request of having the right simultaneously surpasses Ф iF RRByte number give the neighbor node of i; If be less than Ф from the byte number of the neighbor node request of i/authorize iF RR, this formation will be removed from effective tabulation after service;
A kind of mode of poll is adopted in formations all in effective tabulation, till following condition occurring:
1) effectively tabulation is sky;
2) information unit that does not have remaining time slot to increase other in control frame is given distributed scheduling message;
3) value of the compensating parameter of not serviced formation has surpassed its threshold value;
Described time slot allocation mechanism particular content comprises:
1) selects a channel at random;
2) find out first available time slot of first frame;
3) if there is not available time slot, forward next channel to;
4) if all time slots are all searched to be arrived, forward next frame to;
At first, the access order of time slot is a time sequencing, can reduce the time that bandwidth is consulted; The second, the node of the double bounce of being separated by is not known the distributed scheduling message that the other side sends, and they can authorize identical time slot to give the neighbor node that is not identified, and these neighbor nodes are on same channel, and selective channel reduces the probability that this incident takes place at random; At last, in the same frame under the same channel, authorize many more continuous slots can reduce authorization message unit number in the distributed scheduling message;
Described compensation mechanism particular content comprises:
A formation not serviced byte number in the scheduling that takes turns is represented with compensating parameter, when when carrying out the next round scheduling, it can be in these some bandwidth of using with the compensation scheduling of last time more, compensating parameter should be provided with one and go up dividing value, not serviced for a long time to prevent this formation, cause the phenomenon of " dying of hunger ".
2. based on the Wireless Mesh network MAC layer resource scheduling policy of IEEE802.16d standard, it is characterized in that: when polytype Business Stream is arranged in the network, adopt a kind of dynamic bandwidth allocation method, this method comprises following mechanism: classification mechanism, bandwidth reservation mechanism and degradation mechanism
Described classification mechanism particular content comprises:
The four class business of stipulating in the IEEE802.16d standard are respectively unsolicited grant service, real-time polling service, non-real-time polling service and Best-Effort service, and the priority of four class business is followed successively by from high to low: unsolicited grant service, real-time polling service, non-real-time polling service, Best-Effort service, difference according to service attribute, service quality in conjunction with every kind of business, with delineation of activities a kind of in the four class business of stipulating in the standard, each node maintenance four formations of corresponding four class business, the grouping that receives from the upper strata, enter grader, grader is according to the information in the packets headers, grouping delivered in the corresponding queues go, wait scheduling;
Described bandwidth reservation mechanism particular content comprises:
Unsolicited grant service with limit priority is reserved a part of resource, to guarantee the service quality of the Business Stream that real-time is had relatively high expectations;
Described degradation mechanism particular content comprises:
Non-real-time polling service stream is provided with a resource scope of application, represents that in this size this scope is [b with occupied bandwidth Nrtps Min, b Nrtps Max], promptly to have the utilized bandwidth minimum value be b to this business Nrtps Min, the utilized bandwidth maximum is b Nrtps MaxAnd be that two variable: δ of such business setting are the degradation step-length, l is degradation progression, satisfies relational expression between these four parameters:
b nrtps max - lδ ≥ b nrtps min
When having when the not enough situation of resource occurring than the high unsolicited grant service of non-real-time polling service flow priority and real-time polling service, will demote to non-real-time polling service all in the network, promptly the resource that this business is used is cut down, each value of cutting down is δ, till satisfying high-priority service stream, when non-real-time polling service is downgraded to minimum value b Nrtps MinThe time, even high-priority service flows when not being met, also no longer continue non-real-time polling service is demoted.
CNA2008101145905A 2008-06-06 2008-06-06 Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard Pending CN101286949A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNA2008101145905A CN101286949A (en) 2008-06-06 2008-06-06 Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNA2008101145905A CN101286949A (en) 2008-06-06 2008-06-06 Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard

Publications (1)

Publication Number Publication Date
CN101286949A true CN101286949A (en) 2008-10-15

Family

ID=40058927

Family Applications (1)

Application Number Title Priority Date Filing Date
CNA2008101145905A Pending CN101286949A (en) 2008-06-06 2008-06-06 Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard

Country Status (1)

Country Link
CN (1) CN101286949A (en)

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101808324A (en) * 2010-04-09 2010-08-18 北京交通大学 MAC layer architecture design of wireless Mesh network
CN101478417B (en) * 2009-01-16 2011-01-05 深圳市迈科龙电子有限公司 Method for network information automatic distribution and interaction
CN101958836A (en) * 2010-10-12 2011-01-26 中兴通讯股份有限公司 Queuing resource management method and device in hierarchical service quality
WO2011015080A1 (en) * 2009-08-06 2011-02-10 中兴通讯股份有限公司 Data processing method and device
CN101557346B (en) * 2009-05-14 2011-04-06 中国人民解放军信息工程大学 Polling-based packet queue output scheduling method and packet switching node
CN102075431A (en) * 2011-01-25 2011-05-25 武汉迈威光电技术有限公司 Unshielded QoS flow scheduling strategy simultaneously supporting multiple priorities
CN101657030B (en) * 2008-12-19 2011-11-09 重庆邮电大学 Three-way handshaking method in Mesh network based on IEEE802.16
CN102469602A (en) * 2010-11-19 2012-05-23 普天信息技术研究院有限公司 Method for user multi-service dispatching
CN101729122B (en) * 2008-10-24 2012-07-25 普天信息技术研究院有限公司 Method for transmitting control information by distributed relay system
WO2012116656A1 (en) * 2011-03-02 2012-09-07 华为技术有限公司 Method for establishing neighbourhood relations, and network device
CN102904837A (en) * 2012-08-08 2013-01-30 西安交通大学 Differentiated service survivability method based on virtual service plane
CN105744637A (en) * 2016-04-01 2016-07-06 中国科学院信息工程研究所 Broadcast scheduling method for industrial wireless Mesh network
CN105813165A (en) * 2016-03-03 2016-07-27 北京交通大学 Wireless Mesh network applied MAC layer distributed type scheduling mechanism
CN106506387A (en) * 2015-09-08 2017-03-15 北京国双科技有限公司 The method and apparatus of flow-control
CN110190919A (en) * 2019-05-22 2019-08-30 东南大学 More radio frequency multi-channel wireless network method for channel allocation
CN113808380A (en) * 2021-09-15 2021-12-17 宁波钧飞智能科技有限公司 Multi-peripheral control algorithm and unattended integrated machine device based on priority queue

Cited By (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101729122B (en) * 2008-10-24 2012-07-25 普天信息技术研究院有限公司 Method for transmitting control information by distributed relay system
CN101657030B (en) * 2008-12-19 2011-11-09 重庆邮电大学 Three-way handshaking method in Mesh network based on IEEE802.16
CN101478417B (en) * 2009-01-16 2011-01-05 深圳市迈科龙电子有限公司 Method for network information automatic distribution and interaction
CN101557346B (en) * 2009-05-14 2011-04-06 中国人民解放军信息工程大学 Polling-based packet queue output scheduling method and packet switching node
WO2011015080A1 (en) * 2009-08-06 2011-02-10 中兴通讯股份有限公司 Data processing method and device
CN101990252A (en) * 2009-08-06 2011-03-23 中兴通讯股份有限公司 Data processing method and device
CN101808324A (en) * 2010-04-09 2010-08-18 北京交通大学 MAC layer architecture design of wireless Mesh network
CN101808324B (en) * 2010-04-09 2013-04-24 北京交通大学 MAC layer architecture design of wireless Mesh network
CN101958836B (en) * 2010-10-12 2014-11-05 中兴通讯股份有限公司 Queuing resource management method and device in hierarchical service quality
CN101958836A (en) * 2010-10-12 2011-01-26 中兴通讯股份有限公司 Queuing resource management method and device in hierarchical service quality
CN102469602A (en) * 2010-11-19 2012-05-23 普天信息技术研究院有限公司 Method for user multi-service dispatching
CN102469602B (en) * 2010-11-19 2014-06-04 普天信息技术研究院有限公司 Method for user multi-service dispatching
CN102075431A (en) * 2011-01-25 2011-05-25 武汉迈威光电技术有限公司 Unshielded QoS flow scheduling strategy simultaneously supporting multiple priorities
CN102075431B (en) * 2011-01-25 2013-09-25 武汉迈威光电技术有限公司 Unshielded QoS flow scheduling strategy simultaneously supporting multiple priorities
WO2012116656A1 (en) * 2011-03-02 2012-09-07 华为技术有限公司 Method for establishing neighbourhood relations, and network device
CN102904837A (en) * 2012-08-08 2013-01-30 西安交通大学 Differentiated service survivability method based on virtual service plane
CN102904837B (en) * 2012-08-08 2015-01-07 西安交通大学 Differentiated service survivability method based on virtual service plane
CN106506387A (en) * 2015-09-08 2017-03-15 北京国双科技有限公司 The method and apparatus of flow-control
CN106506387B (en) * 2015-09-08 2019-11-08 北京国双科技有限公司 The method and apparatus of flow control
CN105813165A (en) * 2016-03-03 2016-07-27 北京交通大学 Wireless Mesh network applied MAC layer distributed type scheduling mechanism
CN105744637A (en) * 2016-04-01 2016-07-06 中国科学院信息工程研究所 Broadcast scheduling method for industrial wireless Mesh network
CN110190919A (en) * 2019-05-22 2019-08-30 东南大学 More radio frequency multi-channel wireless network method for channel allocation
CN113808380A (en) * 2021-09-15 2021-12-17 宁波钧飞智能科技有限公司 Multi-peripheral control algorithm and unattended integrated machine device based on priority queue

Similar Documents

Publication Publication Date Title
CN101286949A (en) Wireless Mesh network MAC layer resource scheduling policy based on IEEE802.16d standard
CN111954236B (en) Hierarchical edge calculation unloading method based on priority
Ali et al. A performance study of uplink scheduling algorithms in point-to-multipoint WiMAX networks
Nasser et al. Dynamic multilevel priority packet scheduling scheme for wireless sensor network
Vinay et al. Performance evaluation of end-to-end delay by hybrid scheduling algorithm for QoS in IEEE 802.16 network
CN109618375B (en) UAV ad hoc network time slot scheduling method based on service priority and channel interruption probability
Tsai et al. Introduction to packet scheduling algorithms for communication networks
CN101436987A (en) Method and apparatus for scheduling and configuring QoS service
CN104301933A (en) Method for calculating bandwidth and distributing bandwidth in wireless ad hoc network
Saidu et al. A load-aware weighted round-robin algorithm for IEEE 802.16 networks
CN100399781C (en) Terminal up dispatching method for voice service in wide band wireless cut-in network
CN100459582C (en) Group dispatching and channel distributing method for HSDPA system
CN100369524C (en) CDMA system up-bag dispatching method
CN100466593C (en) Method of implementing integrated queue scheduling for supporting multi service
Ma et al. Hierarchical scheduling framework for QoS service in WiMAX point-to-multi-point networks
CN107659940B (en) Multi-priority multi-channel media access control method based on channel perception
Kaur et al. Core-stateless guaranteed throughput networks
Laias et al. An interactive QoS framework for fixed WiMAX networks
Cheng et al. Fairness-based scheduling algorithm for time division duplex mode IEEE 802.16 broadband wireless access systems
CN102769566A (en) Method and device for configuring multilevel scheduling system, and method and device for changing configuration of multilevel scheduling system
JP2002118576A (en) Adaptive cell scheduling algorithm for radio asynchronous transfer mode(atm) system
Kumar et al. Neural network based Scheduling Algorithm for WiMAX with improved QoS constraints
Nikolova et al. Evaluation of surplus round robin scheduling algorithm
Ghazal et al. A real time adaptive scheduling scheme for multi-service flows in WiMAX networks
Satrya et al. Performance analysis of packet scheduling with QoS in IEEE 802.16 e networks

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Open date: 20081015