CN106954242A - A kind of satellite distributed dynamic multi-path method for routing based on network code - Google Patents

A kind of satellite distributed dynamic multi-path method for routing based on network code Download PDF

Info

Publication number
CN106954242A
CN106954242A CN201710104206.2A CN201710104206A CN106954242A CN 106954242 A CN106954242 A CN 106954242A CN 201710104206 A CN201710104206 A CN 201710104206A CN 106954242 A CN106954242 A CN 106954242A
Authority
CN
China
Prior art keywords
node
downstream
packet
probability
neighbor
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.)
Granted
Application number
CN201710104206.2A
Other languages
Chinese (zh)
Other versions
CN106954242B (en
Inventor
赵黎晔
李延东
王韶波
梁巍
马洋
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.)
Aerospace Dongfanghong Satellite Co Ltd
Original Assignee
Aerospace Dongfanghong Satellite Co Ltd
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 Aerospace Dongfanghong Satellite Co Ltd filed Critical Aerospace Dongfanghong Satellite Co Ltd
Priority to CN201710104206.2A priority Critical patent/CN106954242B/en
Publication of CN106954242A publication Critical patent/CN106954242A/en
Application granted granted Critical
Publication of CN106954242B publication Critical patent/CN106954242B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W40/00Communication routing or communication path finding
    • H04W40/02Communication route or path selection, e.g. power-based or shortest path routing
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/14Relay systems
    • H04B7/15Active relay systems
    • H04B7/185Space-based or airborne stations; Stations for satellite systems
    • H04B7/18578Satellite systems for providing broadband data service to individual earth stations
    • H04B7/18584Arrangements for data networking, i.e. for data packet routing, for congestion control
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W40/00Communication routing or communication path finding
    • H04W40/02Communication route or path selection, e.g. power-based or shortest path routing
    • H04W40/12Communication route or path selection, e.g. power-based or shortest path routing based on transmission quality or channel quality
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W40/00Communication routing or communication path finding
    • H04W40/24Connectivity information management, e.g. connectivity discovery or connectivity update
    • H04W40/248Connectivity information update
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W84/00Network topologies
    • H04W84/02Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
    • H04W84/04Large scale networks; Deep hierarchical networks
    • H04W84/06Airborne or Satellite Networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Astronomy & Astrophysics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention discloses a kind of satellite distributed dynamic multi-path method for routing based on network code, first, first, packet is split as equal-sized M packet by source node, and M data are surrounded by identical group number, is forwarded after being encoded using stochastic linear;Recompiled and forwarded with same group of packet originally caching when intermediate node receives a new data packets;Destination node, then receive decoding after m data bag and obtain corresponding information.In node forwarding coding bag, it is first depending on downstream node Delay and distributes priority for it, then it is that its dynamically distributes forwards probability according to its priority and other information when packet is forwarded, packet has probability of failure when being forwarded by a certain downstream node, the present invention can be carried out " compensation is forwarded " according to probability of failure with certain probability, improve the successful delivery rate of packet.The present invention has good dynamic adaptable to network emergency situations, ensures higher packet success delivery rate in the case of relatively low redundancy.

Description

A kind of satellite distributed dynamic multi-path method for routing based on network code
Technical field
The present invention relates to a kind of distributed dynamic multi-path route method, particularly a kind of distribution applied to low rail Information Network Formula dynamic multi-path method for routing, belongs to satellite communication field.
Background technology
As the important component of future space information transmission network, satellite communication network will be closed with various earth-based communications nets It is integrated, ubiquitous global communication service is provided for different users.Wherein, low-track satellite network is due to orbit altitude It is low, with world time delay is small, terminal realizes the advantage such as simple and receives much concern.But low-orbit satellite network topology dynamic change, Inter-satellite link transmission error code is big, causes a large amount of packet losses, it is difficult to provide the information transfer service of high efficient and reliable.
Therefore, the design reliable Routing Protocol of efficient stable is always the study hotspot of low-orbit satellite network.Early stage grinds Study carefully, mainly for satellite change in topology rule, provide effective single-path routing algorithm, but these algorithms can not solve star very well Between packet loss of link (due to packet loss caused by error code or switching) unreliable problem of transmission for bringing.
Relative to single-path routing, multipath route technology has in terms of bandwidth is effectively using, congestion control, transmission reliability Its unique advantage, is the effective means for improving dynamic network transmission reliability.Therefore, some scholars also begin to multipath routeing Applied to satellite network, for example, satellite multi-path routing --- CEMR (Compact Explicit Multi-path Routing), dynamic on-demand multi-path routing algorithm etc..But existing method is all based on concentrated route mode greatly, it is necessary to which node is known Road the whole network information;Meanwhile, although algorithm can reduce the packet loss that congestion is brought, link error code or link switching can not be improved Under the conditions of business correct delivery ratio.
Regarding to the issue above, in recent years, some scholars begin attempt to reduce inter-satellite link packet loss pair using network code The influence of network transmitting, network throughput, equally loaded, lifting transmission reliability are lifted by intermediate node coding, And the dynamic satellite network of topology can be applied to very well.But existing method is mostly to be gulped down by network code between stream to lift network The amount of telling, and protocol complexities are high, practicality is often not enough.
The content of the invention
The technology of the present invention solves problem:Overcome the shortcomings of existing route method, it is proposed that one kind is based on network code Satellite distributed dynamic multi-path method for routing, this method has protocol complexities low, and do not need node to know the whole network information The advantages of can perform, is particularly suitable for dynamic topology and on-board processing low rail information network system limited in one's ability.
The present invention technical solution be:A kind of satellite distributed dynamic multi-path method for routing based on network code, Each node i is received after packet, performs following steps:
(1), judge the destination node d in the packet that receives whether be this node i receive for the first time be sent to the mesh Node d bag, if it is, step (2) is transferred to, if it is not, then being transferred to step (3);
(2) the downstream forward node information table towards destination node d, is generated, into step (3);The downstream forwarding Informational table of nodes is included using destination node d as destination node, and node i is sent to neighbor node j' maximum transmit queue length Qmax(j'), average transmit queue length Qavg(j'), node i is passed via the neighbor node j' shortest paths for reaching destination node d Defeated time delay Dmin(j', d) and from node i each neighbor node j' into downstream forward node list forwarding probability Pj', the j' ∈ CT, CT is meet condition in the neighbor node of node i:Dmin(j',d)<The set of D neighbor node number, the D is default Maximum delay thresholding, and downstream forward node information table is according to Dmin(j', d) ascending order arrangement, and forward node Information table, according to fixed period Δ t, is included in adjustment downstream forward node information table in set CT in real time from after setting up Putting in order in node, downstream forward node table and from node i into downstream forward node list each neighbor node j' Forward probability Pj'
(3), extract the source node identification in packet, destination node to identify, and be compared with the mark of this node, such as The really source node identification in this node identification=packet, then into step (4);If the mesh in this node identification=packet Node identification, then into step (10);Otherwise, into step (5);
(4) packet received, is split as one group of equal-sized M sub- packet Xm, m=1,2 ..., M are described M is integer, and last subdata bag, when packet can not be divided exactly by M, is carried out zero padding to gather together enough length, make N=by M >=1 M, into step (6);
(5), receive Encoded Subdata bag and cache, when the Encoded Subdata frame for receiving at least N' identical group of marks When, give up each Encoded Subdata bag frame head, obtain N' therein sub- packets:Xn, n=1,2 ..., N', the M/3≤ N'≤M, makes N=N' enter step (6);
(6), using random linear network encoding mode by one group of N number of subdata bag Xn, n=1,2 ..., N encoded, Generate the Encoded Subdata bag Y of N number of equal sizen, n=1,2 ... N;
(7), group mark and code coefficient are added to before Encoded Subdata bag as frame head, coding subnumber is formed According to frame Yn' n=1,2 ..., N is simultaneously stored;
(8), according in forward node list downstream from node i to each neighbor node j' forwarding probability Pj', will be N number of Encoded Subdata bag is forwarded to corresponding neighbor node j' respectively;
(9), calculate node i each neighbor node j' into downstream forward node list probability of failure P 'j', and with failure Probability P 'j'Based on, generation compensation forwarding probability P "j', and probability P is forwarded with the compensation "j'By packet Yn' it is retransmitted to neighbours Node j';
(10), receive Encoded Subdata bag that each node sends and store, when receiving the M codings with identical group of mark During subdata bag, original M sub- packets are recovered by Gaussian elimination method, not receiving M has identical group of mark During Encoded Subdata bag, then continue to store to wait.
The downstream forward node information table is generated according to the propagation delay time information table of each node, and is believed with propagation delay time The change of breath table and update the row in the node included in the forward node information table of downstream in set CT and downstream forward node table Sequentially, the propagation delay time information table of each node i safeguards the propagation delay time information table of the node, i according to fixed period Δ t to row ∈ CS, CS are the set of all node numbers in network, and the propagation delay time information table is including with other sections of each in network Point t is as destination node, and node i is sent to each neighbor node j maximum transmit queue length Qmax(j), average transmit queue is long Spend Qavg(j) destination node t shortest path propagation delay time D is reached via each neighbor node j with the nodemin(j, t), j ∈ CL, CL are the set of all neighbor nodes number of node i, CL ∈ CS.
Node i reaches the shortest path propagation delay time of destination node t arrival destination node via each neighbor node j DmiN (j, t), j ∈ CL computational methods are:
Dmin(j, t)=D (e (j, t))+D (e (i, j))+dq,
In formula, when D (e (j, t)) and D (e (i, j)) are respectively the most short link transmission between neighbor node j to destination node t Prolong the chain circuit transmission time delay with each neighbor node j to node i, the chain circuit transmission time delay is according to the t extrapolated by ephemeris Satellite geometry topology shortest path length divided by the light velocity are obtained, dqEach neighbor node j transmit queue time delay is sent to for node i,Q (j) is that each real-time queue lengths of neighbor node j are sent in node i, i.e., the number-of-packet in one queue Amount, LavIt is that average data bag is long, unit bit, C are link bandwidth.
As serial number ss of the downstream neighboring node j' in the forward node table of downstream, s ∈ [1, S], S is that downstream forwards section Neighbor node sum in point table, from node i each neighbor node j' into downstream forward node list forwarding probability Pj'Using Following method is calculated:
(2.1) it is p, to initialize the corresponding normalization forwarding probability of each downstream neighboring node j's, specifically it is calculated as:It is first First, it is that all node j' in the forward node table of downstream produce random number between one 0~1 as its probability pint (s) normalization forwarding Probability p, is then calculated each node j's
(2.2), according to each node j' normalization forwarding Probability pssCalculate node j' relative weighting rj', it is specific to calculate For:
(2.3), according to node j' relative weighting rj'Standardization with node j' quene state calculate node j' is relative Weight is
When downstream node j' transmit queues are full, it is if it standardizes relative weighting
When downstream node j' transmit queues are non-full, i.e. j' ∈ BQ, BQ is non-full queue set, calculates downstream node list In non-full downstream node j' standardization relative weighting be:
When ∑ in above formulak∈BQrkWhen=0, the non-full downstream node of last transmit queue swum in forward table is set It is 1 to standardize relative weighting, and other node standardization relative weightings are 0;
(2.4), it is according to calculate node j' standardization relative weightingCalculate from node i to downstream forward node and arrange Each neighbor node j' forwarding probability P in tablej'
A is the set of all node numbers of the sequence number before j' nodes in the forward node table of node downstream.
According to fixed period Δ t adjustment in real time, from node i, each neighbours into downstream forward node list save step (2) Point j' forwarding probability Pj'Concretely comprise the following steps:
First, average transmit queue threshold value min is setthAnd maxth
Then, average transmit queue length Q is judgedavg(j'), as average transmit queue length Qavg(j') it is less than minthWhen, Make Pj'It is constant;As average transmit queue length Qavg(j') it is more than maxthWhen, make Pj'For 0;As average queue length Qavg(j')∈ [minth,maxth] when, Pj'It is linear to reduce, i.e.,:PThis cycle of j'()=PThe upper a cycles of j'()- Δ p, until PThis cycle of j'()Less than or equal to 0, when PThis cycle of j'()During less than 0, P is madej'=0, the Pj'Variation delta p is:
Δ p=pj'(minth-Qavg(j'))/(maxth-minth)。
Probability of failure P ' described in step (9)j'Computational methods be:
WhenWhen, P 'j'=0;
Work as Qavg(j')∈[minth,maxth] when:
P′j'=1- (1-P) (1-Pa)
In formula, PaFor link average packet loss ratio, P is transmit queue bag drop probability, and P is current average queue length Qavg (j') function:
The compensation forwarding probability is P 'j'~5P 'j'Between random number.
The present invention has the beneficial effect that compared with prior art:
(1), network node of the invention need not obtain full mesh topology information to calculate, safeguard the multipath of fixation, section Point only needs the information that neighbor node is provided, it is possible to situations such as judging failure, the congestion of link, the suitable next road of selection Footpath, reduces protocol complexities and algorithm expense, can be applied to topological dynamic very well and network state information is difficult to precise acquisition Low rail information network.
(2), in packet of the invention forwarding, agreement can forward probability, reduction according to downstream node state dynamically distributes Network congestion, and redundancy compensation is carried out to its retransmission failure probability, packet successful delivery rate is improved, effectively reduction is due between star The packet loss that switching and inter-satellite link error code are brought.
(3), agreement of the invention reduces the correlation of redundancy packet and raw data packets by network code in stream, so as to carry The successful delivery rate of packet under high same redundant degree.
Brief description of the drawings
Fig. 1 implements flow chart for the inventive method;
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in detail.
As shown in figure 1, the present invention provides a kind of satellite distributed dynamic multi-path method for routing based on network code, the party Method it is most crucial be initialization, network code, probability compensation forwarding 3 links.When time interval Δ t starts, i.e., at the moment t0, t1..., tn, in time interval Δ t, satellite network topological novariable, each node updates oneself to all other according to ephemeris The propagation delay time information of node;When packet is transmitted, this method need not in advance be set up and safeguard mulitpath, but according to neighbour Occupy the direct delivery data bag of nodal information.Method can be failed using the timely detecting link of feedback information of neighbor node, region Situations such as congestion, the forwarding strategy for changing downstream node in time does any response without source node, meanwhile, lost to reduce Bag rate, algorithm introduces network code and uses dynamic probability retransmission method in coding bag forwarding strategy.This distribution Repeating process make it that data can be very good to avoid areas of congestion, reduces time delay, algorithm is realized with minimal time delay and packet loss cost To the adaptability of dynamic network.
The satellite distributed dynamic multi-path method for routing based on network code is described in detail below.
Each node i in satellite network, i ∈ CS, CS is the set of the node number in network, according to fixed period Δ T, updates the node to other node t, t ∈ CS, t ≠ i propagation delay time information table, the propagation delay time information table are included with institute There is other nodes t as destination node, t ∈ CS, t ≠ i, node i is sent to neighbor node j maximum transmit queue length Qmax (j), average transmit queue length Qavg(j) the arrival destination node of destination node is reached most via neighbor node j with node i Short path propagation delay time Dmin(j, t), j ∈ CL, CL is the set of all neighbor nodes number of node i, CL ∈ CS.Propagation delay time is believed Cease table as shown in table 1.
The propagation delay time information table of table 1
Each node i according to fixed time interval Δ t, according to ephemeris update present node to all other node away from From renewal propagation delay time information.QmaxAnd Q (j)avg(j) it can be read directly from node i router hardware register.
Node i reaches the shortest path propagation delay time of destination node t arrival destination node via each neighbor node j Dmin(j, t), j ∈ CL computational methods are:
Dmin(j, t)=D (e (j, t))+D (e (i, j))+dq,
In formula, when D (e (j, t)) and D (e (i, j)) are respectively the most short link transmission between neighbor node j to destination node t Prolong the chain circuit transmission time delay with each neighbor node j to node i, the chain circuit transmission time delay is defended according to the t extrapolated by ephemeris Star geometry topology shortest path length divided by the light velocity are obtained, dqEach neighbor node j transmit queue time delay is sent to for node i,Q (j) is that each real-time queue lengths of neighbor node j are sent in node i, i.e., the number-of-packet in one queue Amount, LavIt is that average data bag is long, unit bit, C are link bandwidth.
Node i performs following steps after receiving packet:
(1), judge the destination node d in the packet that receives whether be this node i receive for the first time be sent to the mesh Node d bag, if the bag destination node is this node bag for being sent to the destination node for receiving for the first time, be transferred to step (2), if it is not, then being transferred to step (3).
(2) the destination node d in packet, is extracted, according to the propagation delay time information of the node to destination node d, generation Towards destination node d downstream forward node list, into step (3);As shown in table 2, the downstream forward node information Table is included using destination node d as destination node, and node i is sent to neighbor node j' maximum transmit queue length Qmax(j')、 Average transmit queue length Qavg(j'), node i reaches destination node d shortest path propagation delay time D via neighbor node j'min (j', d) and from node i each neighbor node j' into downstream forward node list forwarding probability Pj', j' the ∈ CT, CT are Condition is met in the neighbor node of node i:Dmin(j',d)<The set of D neighbor node number, the D is default maximum delay Thresholding, D typically takes set { Dmin(j', d), j' ∈ CL } average value, and downstream forward node information table is according to Dmin(j',d) Ascending order arrangement, and forward node information table, according to fixed period Δ t, adjusts downstream in real time from after setting up Putting in order and from node i to downstream in the node that is included in forward node information table in set CT, downstream forward node table Each neighbor node j' forwarding probability P in forward node listj'
The downstream forward node list of table 2
In the table, destination node only includes d, all Dmin(j',d)<D is simultaneously sorted by size.The table in fact constitutes network A subgraph.Later each time interval, table 2 all will carry out data synchronization updating with table 1.
As sequence number ss of the downstream neighboring node j' in the forward node table of downstream, s=[1, S], S is downstream forward node In table neighbor node sum, using following method calculate from node i into downstream forward node list each neighbor node j' Forward probability Pj'
(2.1) it is p, to initialize the corresponding normalization forwarding probability of each downstream neighboring node j's, specifically it is calculated as:It is first First, it is that all node j' in the forward node table of downstream produce random number between one 0~1 as its probability pint (s) normalization forwarding Probability p, is then calculated each node j's
(2.2), according to each node j' normalization forwarding Probability pssCalculate node j' relative weighting rj', it is specific to calculate For:
(2.3), according to node j' relative weighting rj'Standardization with node j' quene state calculate node j' is relative Weight is
When downstream node j' transmit queues are full, it is if it standardizes relative weighting
When downstream node j' transmit queues are non-full, i.e. j' ∈ BQ, BQ is non-full queue set, calculates downstream node list In non-full downstream node j' standardization relative weighting be:
When ∑ in above formulak∈BQrkWhen=0, the non-full downstream node of last transmit queue swum in forward table is set It is 1 to standardize relative weighting, and other node standardization relative weightings are 0;
(2.4), it is according to calculate node j' standardization relative weightingCalculate from node i to downstream forward node and arrange Each neighbor node j' forwarding probability P in tablej'
A is the set of all node numbers of the sequence number before j' nodes in the forward node table of node downstream.
Set up the forwarding probability P from node i each neighbor node j' into downstream forward node listj'Afterwards, according to Fixed period Δ t adjusts the forwarding probability P from node i each neighbor node j' into downstream forward node list in real timej''s Concretely comprise the following steps:
First, average transmit queue threshold value min is setthAnd maxth
Then, average transmit queue length Q is judgedavg(j'), as average transmit queue length Qavg(j') it is less than minthWhen, Make Pj'It is constant;As average transmit queue length Qavg(j') it is more than maxthWhen, make Pj'For 0;As average queue length Qavg(j')∈ [minth,maxth] when, Pj'It is linear to reduce, i.e.,:PThis cycle of j'()=PThe upper a cycles of j'()- Δ p, until PThis cycle of j'()Less than or equal to 0, when PThis cycle of j'()During less than 0, P is madej'=0.
The Pj'Variation delta p is:
Δ p=pj'(minth-Qavg(j'))/(maxth-minth)
(3), extract the source node identification in packet, destination node to identify, and be compared with the mark of this node, such as Really this node identification=source node identification, then into step (4);If this node identification=destination node mark, enters step Suddenly (10);Otherwise, into step (5);
(4) packet received, is split as one group of equal-sized M sub- packets, source data bag X is used asm, M=1,2 ..., M, the M are fixed value, when packet can not be divided exactly by M, and the progress zero padding of last subdata bag is gathered Foot length degree, usual M takes 2 power side, and such as 32 bytes, 64 section words make N=M, into step (6);
(5), receive Encoded Subdata bag and cache, when the Encoded Subdata frame for receiving at least N' identical group of marks When, give up each Encoded Subdata bag frame head, obtain N' therein sub- packets:Xn, n=1,2 ..., N', the M/3≤ N'≤M, makes N=N' enter step (6);
(6), using random linear network encoding mode by one group of N number of subdata bag Xn, n=1,2 ..., N encoded, Generate the Encoded Subdata bag Y of N number of equal sizen, n=1,2 ... N;
Generate the coded data packet Y of N number of equal sizenSpecific coding formula be:
Code coefficient gnm, n=1,2 ..., N, in theory in finite fieldSelected at random with uniform mode with being completely independent Take, i.e., generation random number, the random generating process of different coding coefficient are independent according to being uniformly distributed for code coefficient, and wherein s is represented In finite field Fq scope, this programme, s=8.S typically takes 2 power side.
It is same that the N number of Encoded Subdata bag of the group is compiled using random linear network encoding mode for intermediate node Code, code coefficient is randomly selected, and obtains N number of forwarding Encoded Subdata bag Yn, n=1,2,3 ... N, it is advantageous in that:Further Correlation between reduction coding bag, raising is decoded into power, and the linear re-encoding of coding bag is still the line of corresponding original packet Property combination:
In formula, N1For last coded data packet number.
(7), group mark and code coefficient are added to before subdata bag as frame head, Encoded Subdata bag is formed Yn' n=1,2 ..., N is simultaneously stored;
(8), according in forward node list downstream from node i to each neighbor node j' forwarding probability Pj', will be N number of Encoded Subdata bag is forwarded to corresponding neighbor node j' respectively.
Each at least one downstream node of node, each downstream node one transmit queue of correspondence and node are to purpose Node time delay is different.Algorithm is that multiple downstream nodes distribute forwarding priority, node to destination node time delay (Dmin(j', d)) is got over Small, forwarding priority is higher.Node is dynamically its distribution forwarding probability according to downstream node information.When packet is forwarded, downstream Node is polled successively from high to low according to priority.Packet determines that data are surrounded by one when being forwarded by a certain downstream node Fixed probability of failure.In order to reduce packet loss, algorithm is dynamically carried out " compensation is forwarded " to packet, i.e., turned with certain probability Send out the packet replicated.
This method carries out the probability forwarding of coding bag according to downstream node transmit queue information.When N number of coding bag needs forwarding When, downstream node is polled since first node in the forwarding list of downstream.When neighbor node j' is polled, by N × Pj'Individual coded data packet is sent to neighbor node j';
(9), calculate node i each neighbor node j' into downstream forward node list probability of failure P 'j', and in P 'j' ~5P 'j'Between generation one random number be used as compensation forwarding probability P "j', and probability is forwarded with the compensation
Pj'Repetition forwards the packet to neighbor node j'.
Probability of failure P 'j'Computational methods be:
WhenWhen, P 'j'=0;
Work as Qavg(j')∈[minth,maxth] when:
P′j'=1- (1-P) (1-Pa)
In formula, PaFor link average packet loss ratio, P is transmit queue bag drop probability, and P is current average queue length Qavg (j') function:
Without loss of generality, link average packet loss ratio PaWith the maximum bag drop probability max of transmit queuepCounted in the past by statistics According to acquisition.
Pj'~5Pj' between random number be used as forwarding thresholding, it is assumed that the random number of generation is 2Pj', then turned with the compensation Send out probability 2Pj' by packet Yn' it is retransmitted to neighbor node j', that is to say, that there is 1-2Pj' probability do not forward.
(10), receive Encoded Subdata bag that each node sends and store, when receiving the M codings with identical group of mark During subdata bag, original M sub- packets are recovered by Gaussian elimination method, not receiving M has identical group of mark During Encoded Subdata bag, then continue to store to wait..
Assuming that the packet received is Y1, Y2..., YnIf, the corresponding decoding matrix G=of coding bag coefficient vector [gj1, gj2..., gjm] (j=1,2 ..., m) full rank, then can pass through [X1,X2,...Xm]T=G-1[Y1,Y2,...Ym]TRecover Go out raw data packets.If not being collected into m parts of coding bags or can not decode, continue to store wait.
In summary, the most crucial design philosophy of this method is:
(1) reach and wrap for one, if the bag destination node is this node receive for the first time be sent to the destination node Bag, then node needs to be initialized, including parameter definition, downstream node selection, downstream node sequence.
(2) in network code link, a packet is split as equal-sized M packet, M numbers by source node According to identical group number is surrounded by, forwarded after being encoded using stochastic linear;When intermediate node receives a new data packets with it is former The same group of packet first cached is recompiled and forwarded, and afterwards, empties caching;If destination node, then M data are received Decoding obtains corresponding information after bag.
(3), compensate and forward for probability, in node forwarding coding bag, each downstream node has repeater-transmitter meeting, initially Node is first depending on downstream node Delay and distributes priority for it during change, then when packet is forwarded according to its priority And other information is that its dynamically distributes forwards probability, packet has probability of failure, algorithm when being forwarded by a certain downstream node It can be carried out according to probability of failure with certain probability " compensation is forwarded ", improve the successful delivery rate of packet.So, occur when local During congestion, coding bag retransmission failure probability increase, algorithm can increase compensation forwarding probability, when downstream link failure, node Failure node will be removed from downstream node set and update set.Algorithm has good dynamically adapting to network emergency situations Property, ensure higher packet success delivery rate in the case of relatively low redundancy.
The content not being described in detail in description of the invention belongs to the known technology of those skilled in the art.

Claims (7)

1. a kind of satellite distributed dynamic multi-path method for routing based on network code, it is characterised in that each node i is received After packet, following steps are performed:
(1), judge the destination node d in the packet that receives whether be this node i receive for the first time be sent to the purpose section Point d bag, if it is, step (2) is transferred to, if it is not, then being transferred to step (3);
(2) the downstream forward node information table towards destination node d, is generated, into step (3);The downstream forward node Information table is included using destination node d as destination node, and node i is sent to neighbor node j' maximum transmit queue length Qmax (j'), average transmit queue length Qavg(j') when, node i is transmitted via the neighbor node j' shortest paths for reaching destination node d Prolong Dmin(j', d) and from node i each neighbor node j' into downstream forward node list forwarding probability Pj', the j' ∈ CT, CT for node i neighbor node in meet condition:Dmin(j',d)<The set of D neighbor node number, the D is default Maximum delay thresholding, and downstream forward node information table is according to Dmin(j', d) ascending order arrangement, and forward node are believed Table is ceased from after setting up, according to fixed period Δ t, the section included in real time in adjustment downstream forward node information table in set CT Putting in order in point, downstream forward node table and turn from node i each neighbor node j' into downstream forward node list Send out probability Pj'
(3) source node identification in packet, destination node, is extracted to identify, and be compared with the mark of this node, if this Source node identification in node identification=packet, then into step (4);If the purpose section in this node identification=packet Point identification, then into step (10);Otherwise, into step (5);
(4) packet received, is split as one group of equal-sized M sub- packet Xm, m=1,2 ..., M, the M is whole Number, last subdata bag, when packet can not be divided exactly by M, is carried out zero padding to gather together enough length, makes N=M, enter by M >=1 Enter step (6);
(5), receive Encoded Subdata bag and cache, when receiving the Encoded Subdata frame of the individual identical group of mark of at least N', Give up each Encoded Subdata bag frame head, obtain N' therein sub- packets:Xn, n=1,2 ..., N', the M/3≤N' ≤ M, makes N=N' enter step (6);
(6), using random linear network encoding mode by one group of N number of subdata bag Xn, n=1,2 ..., N encoded, and generates N The Encoded Subdata bag Y of individual equal sizen, n=1,2 ... N;
(7), group mark and code coefficient are added to before Encoded Subdata bag as frame head, Encoded Subdata frame is formed Y′nN=1,2 ..., N is simultaneously stored;
(8), according in forward node list downstream from node i to each neighbor node j' forwarding probability Pj', by N number of coding Subdata bag is forwarded to corresponding neighbor node j' respectively;
(9), calculate node i each neighbor node j' into downstream forward node list probability of failure P 'j', and with probability of failure P′j'Based on, generation compensation forwarding probability P "j', and probability P is forwarded with the compensation "j'By packet Y 'nIt is retransmitted to neighbor node j';
(10), receive Encoded Subdata bag that each node sends and store, when receiving the M coding subnumbers with identical group of mark During according to bag, original M sub- packets are recovered by Gaussian elimination method, the M codings with identical group of mark are not received During subdata bag, then continue to store to wait.
2. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 1, its feature It is that the downstream forward node information table is generated according to the propagation delay time information table of each node, and with propagation delay time information table Change and the arrangement that updates in the node included in the forward node information table of downstream in set CT and downstream forward node table is suitable Sequence, the propagation delay time information table of each node i safeguards the propagation delay time information table of the node, i ∈ according to fixed period Δ t CS, CS are the set of all node numbers in network, and the propagation delay time information table is including with other nodes of each in network t As destination node, node i is sent to each neighbor node j maximum transmit queue length Qmax(j), average transmit queue length Qavg(j) destination node t shortest path propagation delay time D is reached via each neighbor node j with the nodemin(j, t), j ∈ CL, CL is the set of all neighbor nodes number of node i, CL ∈ CS.
3. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 2, its feature It is the shortest path propagation delay time D for the arrival destination node that node i reaches destination node t via each neighbor node jmin (j, t), j ∈ CL computational methods are:
Dmin(j, t)=D (e (j, t))+D (e (i, j))+dq,
In formula, D (e (j, t)) and D (e (i, j)) be respectively most short chain circuit transmission time delay between neighbor node j to destination node t and Each neighbor node j is to the chain circuit transmission time delay of node i, and the chain circuit transmission time delay is according to the t satellite extrapolated by ephemeris Geometry topology shortest path length divided by the light velocity are obtained, dqEach neighbor node j transmit queue time delay is sent to for node i,Q (j) is that each real-time queue lengths of neighbor node j are sent in node i, i.e., the packet in one queue Quantity, LavIt is that average data bag is long, unit bit, C are link bandwidth.
4. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 1, its feature It is that, as serial number ss of the downstream neighboring node j' in the forward node table of downstream, s ∈ [1, S], S is downstream forward node table Middle neighbor node sum, from node i each neighbor node j' into downstream forward node list forwarding probability Pj'Using following Method is calculated:
(2.1) it is p, to initialize the corresponding normalization forwarding probability of each downstream neighboring node j's, specifically it is calculated as:First, it is The random number that all node j' are produced between one 0~1 in the forward node table of downstream is used as its probability pint(s), then Normalization forwarding Probability p is calculated each node j's
p s = p int ( s ) &Sigma; s = 1 S p int ( s )
(2.2), according to each node j' normalization forwarding Probability pssCalculate node j' relative weighting rj', specifically it is calculated as:
(2.3), according to node j' relative weighting rj'With node j' quene state calculate node j' standardization relative weighting For
When downstream node j' transmit queues are full, it is if it standardizes relative weighting
When downstream node j' transmit queues are non-full, i.e. j ' ∈ BQ, BQ is non-full queue set, calculates non-in downstream node list Full downstream node j' standardization relative weighting is:
r ^ j &prime; &Element; B Q = r j &prime; &Sigma; k &Element; B Q r k ,
When ∑ in above formulak∈BQrkWhen=0, the standard of the non-full downstream node of last transmit queue swum in forward table is set It is 1 to change relative weighting, and other node standardization relative weightings are 0;
(2.4), it is according to calculate node j' standardization relative weightingCalculate from node i into downstream forward node list it is every Individual neighbor node j' forwarding probability Pj'
P j &prime; = r ^ j &prime; 1 - &Sigma; k &Element; A r ^ k .
A is the set of all node numbers of the sequence number before j' nodes in the forward node table of node downstream.
5. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 1, its feature Be step (2) according to fixed period Δ t adjustment in real time from node i each neighbor node j' into downstream forward node list Forwarding probability Pj'Concretely comprise the following steps:
First, average transmit queue threshold value min is setthAnd maxth
Then, average transmit queue length Q is judgedavg(j'), as average transmit queue length Qavg(j') it is less than minthWhen, order Pj'It is constant;As average transmit queue length Qavg(j') it is more than maxthWhen, make Pj'For 0;As average queue length Qavg(j')∈ [minth,maxth] when, Pj'It is linear to reduce, i.e.,:PThis cycle of j'()=PThe upper a cycles of j'()- Δ p, until PThis cycle of j'()Less than or equal to 0, when PThis cycle of j'()During less than 0, P is madej'=0, the Pj'Variation delta p is:
Δ p=pj'(minth-Qavg(j'))/(maxth-minth)。
6. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 1, its feature It is probability of failure P ' described in step (9)j'Computational methods be:
WhenWhen, P 'j'=0;
Work as Qavg(j')∈[minth,maxth] when:
P′j'=1- (1-P) (1-Pa)
In formula, PaFor link average packet loss ratio, P is transmit queue bag drop probability, and P is current average queue length Qavg(j') Function:
P = max p Q a v g ( j &prime; ) - min t h max t h - min t h .
7. a kind of satellite distributed dynamic multi-path method for routing based on network code according to claim 1, its feature It is that the compensation forwarding probability is P 'j'~5P 'j'Between random number.
CN201710104206.2A 2017-02-24 2017-02-24 Satellite distributed dynamic multipath routing method based on network coding Active CN106954242B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710104206.2A CN106954242B (en) 2017-02-24 2017-02-24 Satellite distributed dynamic multipath routing method based on network coding

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710104206.2A CN106954242B (en) 2017-02-24 2017-02-24 Satellite distributed dynamic multipath routing method based on network coding

Publications (2)

Publication Number Publication Date
CN106954242A true CN106954242A (en) 2017-07-14
CN106954242B CN106954242B (en) 2020-04-10

Family

ID=59466533

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710104206.2A Active CN106954242B (en) 2017-02-24 2017-02-24 Satellite distributed dynamic multipath routing method based on network coding

Country Status (1)

Country Link
CN (1) CN106954242B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109560860A (en) * 2018-12-25 2019-04-02 长沙天仪空间科技研究院有限公司 A kind of satellite communication method for routing and system
CN111585638A (en) * 2020-04-22 2020-08-25 浙江时空道宇科技有限公司 Inter-satellite network communication method, communication satellite and system
CN112887203A (en) * 2021-01-12 2021-06-01 中国人民解放军31007部队 TDMA wireless network multi-path data transmission method based on network coding

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004028034A1 (en) * 2002-09-23 2004-04-01 Inmarsat Ltd Communication method and apparatus
CN101651599A (en) * 2008-08-12 2010-02-17 中国移动通信集团公司 Multipath wireless routing method and device
CN102143566A (en) * 2011-02-18 2011-08-03 上海大学 Life cycle maximizing method for distributed wireless video sensor network
CN102572717A (en) * 2012-02-20 2012-07-11 南京中通电气有限公司 Multipath routing reliable transmission method based on network coding
CN105764084A (en) * 2016-02-05 2016-07-13 南京邮电大学 Wireless sensor network multipath routing protocol realization method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004028034A1 (en) * 2002-09-23 2004-04-01 Inmarsat Ltd Communication method and apparatus
CN101651599A (en) * 2008-08-12 2010-02-17 中国移动通信集团公司 Multipath wireless routing method and device
CN102143566A (en) * 2011-02-18 2011-08-03 上海大学 Life cycle maximizing method for distributed wireless video sensor network
CN102572717A (en) * 2012-02-20 2012-07-11 南京中通电气有限公司 Multipath routing reliable transmission method based on network coding
CN105764084A (en) * 2016-02-05 2016-07-13 南京邮电大学 Wireless sensor network multipath routing protocol realization method

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109560860A (en) * 2018-12-25 2019-04-02 长沙天仪空间科技研究院有限公司 A kind of satellite communication method for routing and system
CN109560860B (en) * 2018-12-25 2021-02-09 长沙天仪空间科技研究院有限公司 Satellite communication routing method and system
CN111585638A (en) * 2020-04-22 2020-08-25 浙江时空道宇科技有限公司 Inter-satellite network communication method, communication satellite and system
CN112887203A (en) * 2021-01-12 2021-06-01 中国人民解放军31007部队 TDMA wireless network multi-path data transmission method based on network coding
CN112887203B (en) * 2021-01-12 2021-10-08 中国人民解放军31007部队 TDMA wireless network multi-path data transmission method based on network coding

Also Published As

Publication number Publication date
CN106954242B (en) 2020-04-10

Similar Documents

Publication Publication Date Title
Guo et al. Efficient error recovery with network coding in underwater sensor networks
CN104243096B (en) A kind of deep space multifile transmission method based on fountain codes
US9553956B2 (en) Self-adaptive network control transmission method and system based on TCP
CN103905152B (en) Using the effective throughput randomized optimization process of layer-span combined coding in erasure channel
Huang et al. Just FUN: A joint fountain coding and network coding approach to loss-tolerant information spreading
CN103906165B (en) A kind of online opportunistic method for routing perceived based on coding
CN107959551A (en) A kind of reliable data transport in wireless sensor networks method based on network code
CN105450357A (en) Adjustment method of encoding parameters, adjustment device of encoding parameters, processing method of feedback information and processing device of feedback information
CN106954242A (en) A kind of satellite distributed dynamic multi-path method for routing based on network code
CN104079483A (en) Multistage security routing method for delay tolerant network and based on network codes
CN107682434A (en) A kind of underwater sensor network framework and its implementation
Liu et al. TCP performance in wireless access with adaptive modulation and coding
CN107222404A (en) The parallel route retransmission method of redundancy encoding is carried in WOBAN
CN102904954B (en) Reliable fragmented transmission method based on network coding for 6LoWPAN (IPv6 over Low Power Wireless Personal Area Network)
Gómez et al. Reliable communications over wireless mesh networks with inter and intra-flow network coding
CN102098132B (en) Wireless cooperative relay network-based hierarchical random network coding method
CN102025473A (en) Retransmission method based on intergenerational network coding in packet network
CN107222284A (en) A kind of fountain coding anti-eavesdropping transmission method in Internet of Things based on channel feedback
CN105227457B (en) Towards the protocol Data Unit cross-layer optimizing method under DTN dynamic links
CN102316315B (en) Real-time video transmission method
CN115378548B (en) Connectionless-oriented binary superposition determination linear network code transmission method
CN112887203B (en) TDMA wireless network multi-path data transmission method based on network coding
CN106254044A (en) A kind of dynamic linear combination repeating method based on multicast network coding
Wang et al. A reliable routing algorithm with network coding in internet of vehicles
CN109104238A (en) The dimensionally-optimised algorithm of satellite network DTN link Bundle based on Markov decision

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant