CN110392394A - MPTCP dispatching method in wireless network based on link-state information - Google Patents
MPTCP dispatching method in wireless network based on link-state information Download PDFInfo
- Publication number
- CN110392394A CN110392394A CN201910682837.1A CN201910682837A CN110392394A CN 110392394 A CN110392394 A CN 110392394A CN 201910682837 A CN201910682837 A CN 201910682837A CN 110392394 A CN110392394 A CN 110392394A
- Authority
- CN
- China
- Prior art keywords
- data packet
- tcp subflow
- dsn
- pri
- tcp
- 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
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
- H04L1/0006—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission format
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/0231—Traffic management, e.g. flow control or congestion control based on communication conditions
- H04W28/0236—Traffic management, e.g. flow control or congestion control based on communication conditions radio quality, e.g. interference, losses or delay
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/0247—Traffic management, e.g. flow control or congestion control based on conditions of the access network or the infrastructure network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/0289—Congestion control
Abstract
The invention discloses the MPTCP dispatching methods in a kind of wireless network based on link-state information, are related to technical field of the computer network.The dispatching method, every TCP subflow is after receiving ack msg packet, according to its position in the sequence of all TCP subflow network state quality of the updating network state of the TCP subflow, further according to the transmitted data packet of the position selection of the TCP subflow in the ranking;As long as every TCP subflow, which receives ack msg packet, updates its network state, the transmission of the data packet of next round is carried out again, without waiting for the scheduling for carrying out next round after all TCP subflows completion data packet transformation task again, the waiting time is substantially reduced, data transmission efficiency is improved.
Description
Technical field
The invention belongs to computer networking technology more particularly to a kind of MPTCP transmission paths good to network link situation
The dispatching method of pre- subpackage.
Background technique
With the development of network and mobile device, more and more network applications and technology, such as live streaming, VR and cloud computing,
Bigger network bandwidth is needed to realize higher performance.More host's technologies enable mobile device to be equipped with multiple network interfaces
Card, such as mobile device can connect multiple WiFi and 4G cellular networks simultaneously, realize multipath parallel transmission (as shown in Figure 1,
Multi-path network topological structure comprising three TCP subflows).Obviously, multipath parallel transmission brings many benefits a: side
Face, when the bandwidth of single-pathway is limited, it can increase flow and handling capacity with aggregate bandwidth;On the other hand, it can be mentioned
The reliability of high data transmission, for example, other paths still can connect transmission data when a paths are interrupted.
MPTCP (MultiPath Transport Control Protocol) multipath TCP transmission data are to send
A plurality of TCP subflow (as shown in Figure 1) is established between end and receiving end, Parallel Scheduling mulitpath carries out data transmission.By interconnecting
Net engineering task force (IETF) standardized MPTCP has good adaptability and scalability, is readily applied to actual net
In network application program.Currently, MPTCP v0.94 version is realized on linux kernel v4.14, it is deployed in
As agency in the routers such as OpenMPTCProuter, 17wifi, in the operating system of smart phone, or even in data
In central site network.
MPTCP extends conventional TCP, to provide the service for supporting transmission connection simultaneously across mulitpath operation.MPTCP is fixed
In justice, every established path is all a TCP subflow, and the starting and termination of each subflow are similar with conventional TCP connection.
MPTCP passes through the transmission of DSN (data sequence number) Lai Guanli data packet, and DSN counts total message segment serial number, and
Serial number in each TCP subflow channel is continuous always.The unordered arrival of data packet be MPTCP must face mainly ask
It inscribes, research before is taught that, the greatest differences of path or link condition, including delay, loss rate and shared buffer
Area's size can all cause serious data packet disorder to reach, and receiving buffer area will frequently block, under handling capacity (rate) will be serious
Drop, in addition it is more even worse than the path single TCP.
Therefore, dispatching algorithm is the key that design MPTCP, it can guarantee that packet sequence reaches as far as possible, when examining
When considering the quality or state in path, how dispatching algorithm decision delivers a packet to corresponding network link.Currently, MPTCP
Default schduling algorithm be known as polling algorithm (Round Robin), each subflow of polling algorithm poll delivers data packet, without
Consider path quality or state;Second alternative scheduler MiniRTT is always by round-trip delay (Round-Trip Time, RTT)
The smallest packet distributes to the subflow with idle window.However, when the small data packet of DSN is sent to the link of state difference,
MiniRTT also results in a large amount of buffer area obstruction.Therefore, when designing dispatching algorithm, DAPS and STMS proposition are smaller by DSN
Data packet be sent to the path of low latency, high bandwidth, send the preferable path of network state for the biggish data packet of DSN,
But when dispatching data packet, the case where receiving end caching blocks is not accounted for, two paths must be complete in same amount of time
Next round scheduling is just carried out at respective transformation task, this results in efficiency of transmission declines;In addition, STMS is one relatively high
The improved predistribution dispatching algorithm of grade, it can select data packet to suitable subflow or path, but STMS needs
After all subflows complete data transmission, next round scheduling can be just carried out, the decline of throughput is caused.
Summary of the invention
For in the prior art, improvement polling algorithm and MiniRTT do not consider that DSN sends data packet directly to subflow and makes
The problem of blocking at receiving end, and predistribution dispatching algorithm need that all subflows is waited to complete data and throughput is caused to decline
The problem of, the present invention provides the MPTCP dispatching method based on link-state information in a kind of wireless network.
The present invention is to solve above-mentioned technical problem by the following technical solutions: based on link in a kind of wireless network
The MPTCP dispatching method of status information, including the following steps:
Step 1: all TCP subflows being sorted according to the quality of network state, are waited for according to every TCP substream allocation is ordered as
Data packet is sent, every TCP subflow sends data packet according to DSN;
Step 2: updating position pri of the TCP subflow i in the sequence of all TCP subflow network state quality;
Step 3: if the network state of TCP subflow i be it is best, send DSN near preceding not sent data packet;If
The network state of TCP subflow i be not it is best, then be transferred to step 4;
Step 4: judging whether to need to level-one TCP subflow reserved data packet more preferable than the TCP subflow i network state;If
Reserved data packet is not needed, and the last TCP subflow in position pri has not sent data packet, then sends last place
In the not sent data packet of the TCP subflow of position pri, the data packet for otherwise selecting DSN bigger is sent;If desired reserved data
Packet, the then DSN of the data packet sent are that the quantity of current reservations data packet and the last TCP subflow in position pri are sent
The sum of the DSN of data packet;
Step 5: step 2-4 is repeated, until all data packets are sent.
Every TCP subflow of MPTCP dispatching method of the invention updates it in institute again after having sent a data packet
There is the position in the sequence of TCP subflow network state quality, according to data packet transmitted by the fine or not reselection of network state, network
The DSN of data packet transmitted by the good TCP subflow of state is forward, and the TCP subflow of network state difference is more than its network state
The TCP subflow reserved data packet of good level-one, the DSN of data packet transmitted by the TCP subflow of network state difference rearward, every TCP
Subflow receives ack msg packet (complete a data packet send), updates the quality after network state all in accordance with its network state
The transmission of data packet next time is carried out again, and the tune for carrying out next round after data packet is sent again is completed without waiting for all TCP subflows
Degree, substantially reduces the waiting time, improves data transmission efficiency;Meanwhile the data packet that the good TCP subflow of network state is sent
Affiliated DSN is forward, and the data packet sent is more, and DSN belonging to the data packet that the TCP subflow of network state difference is sent is leaned on
Afterwards, and the data packet that sends is few, so that all data packets orderly reach as far as possible, alleviates asking of receiving that buffer area frequently blocks
Topic, further improves the efficiency of transmission of data.
Further, in the step 1, data packet to be sent includes reserved data packet, the quantity of reserved data packet with should
The network state of TCP subflow is related, and TCP subflow network state is better, and the quantity of reserved data packet is more.
Further, in the step 2, TCP subflow i is updated in the ack msg packet for receiving the TCP subflow every time should
Position of the TCP subflow i in the sequence of all TCP subflow network state quality.
Further, in the step 4, judge whether the concrete operations for needing reserved data packet are as follows: judgement is last pre-
Stay whether the sum of quantity of data packet is less than three times of the sum of last congestion window;
Wherein, the sum of the quantity of last reserved data packet refers to that the last TCP subflow in position pri is than this
Reserved the sum of the quantity of data packet of the good all TCP subflows of TCP subflow network state;The sum of last congestion window refers to
The sum of the congestion window of all TCP subflows better than the last TCP subflow network state in position pri;
If so, needing reserved data packet;If it is not, not needing reserved data packet then.
Further, in the step 4, the calculation formula of the quantity of reserved data packet is needed are as follows:
Res=(RTTP × BWP+Rud ÷ CWNDpri-1)×(DSNpri-1-DSN0)
Wherein, Res indicates the quantity of current reservations data packet;SRTTpri-1Indicate current place
In the smooth roundtrip time delay of the TCP subflow of position pri-1, SRTTpriIndicate the smooth past of the TCP subflow for being currently at position pri
Return time delay;BWP expression is currently at the ratio between the bandwidth of the TCP subflow and the TCP subflow for being currently at position pri of position pri-1,
I.e.Rud indicates the redundancy size of the data packet of current receiving end;CWNDpri-1Expression is currently at position
The congestion window of the TCP subflow of pir-1;DSNpri-1Indicate the data packet that the TCP subflow for being currently at position pir-1 is sent
DSN;DSN0Indicate the DSN for the data packet that the best TCP subflow of current network state is sent.
Further, the smooth roundtrip time delay SRTT of the TCP subflow in position pripriCalculation formula are as follows:
SRTTpri=(α × SRTTHis, pri)+((1-α)×RTTpri)
Wherein, α indicates impact factor, for indicating to RTTpriAcceptance level, SRTTHis, priIndicate that the last time is in
The smooth roundtrip time delay of the TCP subflow of position pri, RTTpriIndicate the round-trip delay of the TCP subflow in position pri.
Further, the calculation formula of the redundancy size Rud of the data packet of current receiving end are as follows:
Rud=β × Rudhis+(1-β)×ΔRud
Wherein, RudhisIndicate that the backup length in last rx-side buffering area, β indicate impact factor, Δ Rud expression connects
The difference of receiving end buffer area current congestion length and last congestion length.
Further, the specific calculating of the difference Δ Rud of rx-side buffering area current congestion length and last congestion length
Formula are as follows:
Δ Rud=DSN0-DATAACK-Rudhis
Wherein, DSN0Indicating the DSN for the data packet that the best TCP subflow of current network state is sent, DATAACK is
The confirmation message of the data packet of TCP subflow rank in MPTCP scheduler.
Further, in the step 4, if desired reserved data packet, then also need to judge transmitted by TCP subflow i
Whether the DSN of data packet is greater than scheduler dispatches window, if so, the TCP subflow i is waited, sends on TCP subflow i empty
Data packet, then it is transferred to the step 2;
If it is not, then judging whether the DSN of data packet transmitted by TCP subflow i is greater than is currently at position pri+1's
The DSN of data packet transmitted by TCP subflow;If so, the DSN for the data packet that the TCP subflow i is sent is than TCP subflow i's
The DSN of data packet transmitted by the TCP subflow of network state difference level-one, if it is not, the data packet that then the TCP subflow i is sent
DSN is the quantity of current reservations data packet and the sum of the DSN of data packet that the last TCP subflow in position pri is sent.
By the judgement of the DSN and scheduler dispatches window size of the data packet of transmission, it is poor to avoid network state
Data packet data packet transmitted by the preferable TCP subflow of receiving end clogging networks state that TCP subflow is sent.
Beneficial effect
Compared with prior art, dispatching method of the present invention, every TCP subflow are receiving ack msg Bao Hougen
According to its position in the sequence of all TCP subflow network state quality of the updating network state of the TCP subflow, further according to the TCP
The transmitted data packet of the position selection of subflow in the ranking;As long as every TCP subflow, which receives ack msg packet, updates its net
Network state, then carry out the transmission of the data packet of next round, without waiting for all TCP subflows complete after data packet transformation task again into
The scheduling of row next round, substantially reduces the waiting time, improves data transmission efficiency;Meanwhile the TCP subflow that network state is good
DSN belonging to the data packet of transmission is forward, and the data packet sent is more, and the data packet that the TCP subflow of network state difference is sent
Rearward, and the data packet sent is few, so that all data packets orderly reach as far as possible, alleviates and receives buffer area by affiliated DSN
Frequently the problem of obstruction, the efficiency of transmission of data is further improved.
The present invention judges also according to the DSN of transmitted data packet and the size of scheduler dispatches window, effectively prevents
Data packet number transmitted by the preferable TCP subflow of receiving end clogging networks state that the poor TCP subflow of network state is sent
According to packet, avoids and do not consider the problems of that DSN sends data packet directly to subflow and receiving end is caused to block.
Detailed description of the invention
It, below will be to attached drawing needed in embodiment description in order to illustrate more clearly of technical solution of the present invention
It is briefly described, it should be apparent that, the accompanying drawings in the following description is only one embodiment of the present of invention, general for this field
For logical technical staff, without creative efforts, it is also possible to obtain other drawings based on these drawings.
Fig. 1 is the transmission topological diagram of MPTCP in the embodiment of the present invention;
Fig. 2 is the flow chart of dispatching method in the embodiment of the present invention;
Fig. 3 is the sequence of TCP subflow and allocation of packets figure in the embodiment of the present invention;
Fig. 4 is the scheduling graph in the embodiment of the present invention under TCP2 network state situation of change;
In Fig. 4, (a) is the DSN for the data packet that TCP1, TCP2 and TCP3 are sent under initial situation, (b) network-like for TCP2
The DSN of the data packet sent when state becomes preferably (c) is still within the DSN of the data packet of second position transmission, (d) for TCP2
The DSN of the data packet sent when becoming worst for TCP2 network state;
Fig. 5 is the scheduling graph in the embodiment of the present invention in the case of reserved data packet;
In Fig. 5, (a) is the DSN for the data packet that TCP1, TCP2 and TCP3 are sent under initial situation, (b) is not need to reserve
The DSN for the data packet that TCP2 is sent when data packet is (c) DSN of the data packet sent of TCP2 when needing reserved data packet;
Wherein, 1- transmitting terminal, the receiving end 2-, 3-MPTCP send window.
Specific embodiment
With reference to the attached drawing in the embodiment of the present invention, the technical solution in the present invention is clearly and completely described,
Obviously, described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based in the present invention
Embodiment, those of ordinary skill in the art's every other embodiment obtained without creative labor,
It shall fall within the protection scope of the present invention.
MPTCP dispatching method in a kind of wireless network provided by the present invention based on link-state information, including it is following
Several steps:
Step 1: all TCP subflows are sorted in network state sequencing table sort according to the quality of network state, according to
It is ordered as every TCP substream allocation data packet to be sent, every TCP subflow sends data packet according to DSN;
After step 2:TCP subflow i receives ack msg packet, updates the TCP subflow i and sort in all TCP subflow network states
Position pri in table sort;
Step 3: if the network state of TCP subflow i be it is best, send DSN near preceding not sent data packet;If
The network state of TCP subflow i be not it is best, then be transferred to step 4;
Step 4: judging whether to need to lead to level-one TCP subflow reserved data packet more preferable than the TCP subflow i network state
The TCP subflow that network state sequencing table sort obtains the more preferable level-one of network state is crossed, that is, is in the TCP subflow of position pri-1;If
Reserved data packet is not needed, and the last TCP subflow in position pri has not sent data packet, then sends last place
In the not sent data packet of the TCP subflow of position pri, the data packet for otherwise selecting DSN bigger is sent;If desired reserved data
Packet, the then DSN of the data packet sent are that the quantity of current reservations data packet and the last TCP subflow in position pri are sent
The sum of the DSN of data packet;
Step 5: step 2-4 is repeated, until all data packets are sent.
Every TCP subflow of MPTCP dispatching method of the invention updates it in institute again after having sent a data packet
There is the position in the sequence of TCP subflow network state quality, according to data packet transmitted by the fine or not reselection of network state, network
The DSN of data packet transmitted by the good TCP subflow of state is forward, and the TCP subflow of network state difference is more than its network state
The TCP subflow reserved data packet of good level-one, the DSN of data packet transmitted by the TCP subflow of network state difference rearward, every TCP
Subflow receives ack msg packet (complete a data packet send), updates the quality after network state all in accordance with its network state
The transmission of data packet next time is carried out again, and the tune for carrying out next round after data packet is sent again is completed without waiting for all TCP subflows
Degree, substantially reduces the waiting time, improves data transmission efficiency;Meanwhile the data packet that the good TCP subflow of network state is sent
Affiliated DSN is forward, and the data packet sent is more, and DSN belonging to the data packet that the TCP subflow of network state difference is sent is leaned on
Afterwards, and the data packet that sends is few, so that all data packets orderly reach as far as possible, alleviates asking of receiving that buffer area frequently blocks
Topic, further improves the efficiency of transmission of data.
Embodiment
MPTCP is efficiently to transmit number by a plurality of TCP link (or TCP subflow) between a source node and a destination node
According to.As shown in Figure 1, in the present embodiment, including 3 TCP subflows, transmitting terminal and receiving end are to realize MPTCP kernel
The computer of Ubuntu system is added to 6 router gateways between transmitting terminal and receiving end, distinguishes in every TCP subflow
For 2 router gateways, the time delay and bandwidth of every TCP subflow are as shown in table 1 below, and the present embodiment passes through modification MPTCP kernel
In dispatching method so that data orderly reach receiving end as far as possible in transmission process.
The time delay and bandwidth of 13 TCP subflows of table
Subflow title | Time delay | Bandwidth |
TCP1 | 20ms | 4Mbps |
TCP2 | 50-300ms | 1-4Mbps |
TCP3 | 0-100ms | 0-2Mbps |
As shown in Fig. 2, specific dispatching method is as follows:
1) MPTCP connection is established between transmitting terminal and receiving end, while passing through 3 TCP subflow (TCP1/TCP2/
TCP3 data packet (assuming that the window value of 3 TCP subflows is all 10 data packets)) is sent to receiving end, receives ack msg at first
The TCP subflow of packet indicates that its network state is best, is ranked up 3 TCP subflows according to the quality of network state, establishes network
State sequencing table sort, in the present embodiment, network state is from good to bad TCP1, TCP2, TCP3 of being ordered as, data to be sent
Bao Jun has DSN, and data packet to be sent is ranked up according to DSN, is ordered as every TCP substream allocation number according to network state
According to packet, every TCP subflow sends data packet according to DSN.
It as shown in figure 3, TCP1, which is assigned to DSN, is 1~10 data packet, and be the DSN of data packet that TCP1 is reserved is 11
~60, TCP2 are assigned to DSN and are 61~70 data packet, and be the DSN of the reserved data packet of TCP2 are that 71~100, TCP3 divides
It is fitted on the data packet that DSN is 101~110.Since TCP1 is the best subflow of network state, TCP2 takes second place, and TCP3 is worst, so
The quantity of TCP1 reserved data packet be it is most, the quantity of TCP2 reserved data packet is taken second place, and guarantees that 3 TCP subflows are sent
Time required for complete data packet close to.Theoretically, if the network state of 3 TCP subflows does not change, 3
TCP subflow can sequence send the data packet of distribution, i.e. TCP1 sequence has sent the data packet that DSN is 51~60, TCP2 sequence
The data packet that DSN is 91~100 is sent, TCP3 has sent the data packet that DSN is 101~110, and 3 TCP subflows are completed all
The time of the transmission of the data packet of distribution close to, avoid between each subflow it is mutual waiting caused by transmission rate
Slow problem.
2) network state is changeable, and the time delay and bandwidth of every TCP subflow are not necessarily stable always, therefore,
The network state of every TCP subflow is likely to improve or worse.Every TCP subflow is (after completing a data packet transmission)
Transmitting terminal receives ack msg Bao Shijun and updates the TCP subflow again in all TCP subflow network state sequencing table sort
Position.
If 3) network state of certain TCP subflow becomes best, DSN is sent near preceding not sent data packet.
As shown in Fig. 4 (a), TCP1 sends the data packet that DSN is 31~40, and TCP2 sends the data packet that DSN is 81~90,
TCP3 sends the data packet that DSN is 111~120, and network state changes, and TCP2 is sending data of the DSN for 81~90
Network state becomes best when packet, then TCP2 sends DSN near preceding not sent data packet, and as shown in Fig. 4 (b), TCP2 is sent out
Sending DSN is 41~50 data packet.
If 4) network state of TCP2 is not to become best, need to judge whether to need to more than TCP2 network state
The TCP subflow reserved data packet of good level-one, specific judgment formula are as follows:
Wherein, DSNHis, priIndicate the DSN for the data packet that the last TCP subflow in position pri is sent,
DSNHis, pri-1Indicate the DSN, CWND of the data packet that the last TCP subflow in position pri-1 is sentHis, jIndicate last
The congestion window of certain the TCP subflow j on position 0 to pri-1,It indicates last and is in position
Set the sum of the congestion window of all TCP subflows on 0 to position pri-1;DSNHis, pri-DSNHis, pri-1As the last time is in
The TCP subflow of position pri is the quantity of the reserved data packet of last all TCP subflows in position 0 to position pri-1
The sum of;Pri indicates position of the TCP subflow in network state quality sequencing table sort, and pri-1 indicates network state ratio pri more
The position of good level-one, pri+1 indicate the position of network state level-one worse ratio pri.
For example, current network state quality is ordered as TCP1, TCP2, TCP3, TCP4, last network state quality
It is ordered as TCP1, TCP3, TCP2, TCP4, judges whether to need to the TCP2 reserved data of level-one more preferable than TCP3 network state
Packet, then judge whether the sum of quantity of last reserved data packet is less than three times of the sum of last congestion window;
Wherein, the sum of the quantity of last reserved data packet refers to that the last TCP2 in the third place is than TCP2 net
Reserved the sum of the quantity of data packet of the good all TCP subflows (i.e. TCP1 and TCP3) of network state;The sum of last congestion window
Refer to the congestion window of all TCP subflows (i.e. TCP1 and TCP3) better than the last TCP2 network state in the third place
The sum of.
If formula (1) is set up, reserved data packet is needed, if formula (1) is invalid, does not need reserved data
Packet.
If you do not need to reserved data packet, as shown in Fig. 4 (c), current TCP2 in the sequence of network state quality still
In second position, then current TCP2 sends the not sent data packet of the last TCP subflow in the second position, i.e., currently
TCP2 sends the not sent data packet of the last TCP2 in the second position, and (last time has in the TCP subflow of the second position
Not sent complete data packet, if without not sent complete data packet, current TCP2 sends the bigger data packet of DSN, such as
The data packet that DSN is 121~130), current TCP2 sends the data packet that DSN is 91~100;As shown in Fig. 4 (d), current TCP2
Become also poorer than TCP3 in the sequence of network state quality, current TCP2 is in the position of third, then in current TCP2 transmission
The not sent data packet of the primary TCP subflow in the third place, i.e., current TCP2 send the last time in the third place
TCP3 not sent data packet, current TCP2 send the data packet that DSN is 121~130.
If necessary to reserved data packet, as shown in figure 5, Fig. 5 (a) is that network state quality is ordered as TCP1, TCP2, TCP3
When, TCP1 send DSN be 31~40 data packet, TCP2 send DSN be 81~90 data packet, TCP3 send DSN be 111~
120 data packet;Fig. 5 (b) be when TCP2 does not need the TCP1 reserved data packet for level-one more preferable than TCP2 network state,
TCP2 sends the data packet that DSN is 91~100;Fig. 5 (c) is when TCP2 needs the TCP1 for level-one more preferable than TCP2 network state
When reserved data packet, TCP2 sends the data packet that DSN is 121~130, the calculation formula of the DSN for the data packet that TCP2 is sent are as follows:
DSNpri=DSNHis, pri+Res (2)
Wherein, DSNpriIndicate the DSN, DSN of the data packet that the TCP subflow for being currently at position pri is sentHis, priIt indicates
The DSN for the data packet that TCP subflow of the last time in position pri is sent, Res indicate the quantity or length of current reservations data packet
Degree.
The calculation formula of the amount R es of current reservations data packet are as follows:
Res=(RTTP × BWP+Rud ÷ CWNDpri-1)×(DSNpri-1-DSN0) (3)
Wherein,SRTTpri-1Indicate the smooth past of the TCP subflow for being currently at position pri-1
Return time delay, SRTTpriIndicate the smooth roundtrip time delay for being currently at the TCP subflow of position pri;BWP expression is currently at position
The ratio between the bandwidth of the TCP subflow of pri-1 and the TCP subflow in position pri;Rud indicates the superfluous of the data packet of current receiving end
Remaining size;CWNDpri-1Indicate the congestion window for being currently at the TCP subflow of position pir-1;DSNpri-1Expression is currently at position
Set the DSN of the data packet of the TCP subflow transmission of pir-1;DSN0Indicate the data that the best TCP subflow of current network state is sent
The DSN of packet.
Smooth roundtrip time delay SRTT refers to the time for sending a data packet and receives being averaged for the time of a data packet
It is worth, in the present embodiment, the smooth roundtrip time delay SRTT of the TCP subflow in position pripriCalculation formula are as follows:
SRTTpri=(α × SRTTHis, pri)+((1-α)×RTTpri) (4)
Wherein, α indicates impact factor, for indicating to RTTpriAcceptance level, SRTTHis, priIndicate that the last time is in
The smooth roundtrip time delay of the TCP subflow of position pri, RTTpriIndicate the round-trip delay of the TCP subflow in position pri.This implementation
The SRTT of example considers history RTT while keeping calculating simple, and α is closer to 0, then it represents that SRTT more believes this RTT's
It obtains.
In the present embodiment, the calculation formula of the redundancy size Rud of the data packet of current receiving end are as follows:
Rud=β × Rudhis+(1-β)×ΔRud (5)
Wherein, RudhisIndicate that the backup length in last rx-side buffering area, β indicate impact factor, Δ Rud expression connects
The difference of receiving end buffer area current congestion length and last congestion length, the specific formula for calculation of Δ Rud are as follows:
Δ Rud=DSN0-DATAACK-Rudhis (6)
Wherein, DSN0Indicate the DSN, DSN of the data packet that the best TCP subflow of current network state is sent0It can be from scheduling
It is obtained in device record, DATAACK is the confirmation message of the data packet of TCP subflow rank in MPTCP scheduler;DSN0-DATAACK
As current congestion length.
As shown in Fig. 2, in step 4, if desired reserved data packet, the DSN of data packet transmitted by TCP subflow i
(DSNpri) be the quantity of current reservations data packet and data packet that the last TCP subflow in position pri is sent DSN it
(Res+DSNHis, pri), then also need to judge whether the DSN of data packet transmitted by TCP subflow i is greater than scheduler dispatches
Window sends empty data packet on TCP subflow i, detects the network-like of TCP subflow i if so, the TCP subflow i is waited
State, then be transferred to step 2 and update position of the TCP subflow i in all TCP subflow network state sequencing table sort again.
If it is not, then judgement is currently at DSN (the i.e. DSN of data packet transmitted by the TCP subflow i of position pripri) be
It is no be greater than be currently at the DSN of data packet transmitted by the TCP subflow of position pri+1 (network state i.e. than TCP subflow i be poor
The TCP subflow of level-one);If it is not, then the TCP subflow i send data packet DSN be current reservations data packet quantity with it is upper
The sum of the DSN of data packet that the primary TCP subflow in position pri is sent, if so, the data packet that the TCP subflow i is sent
The DSN of data packet transmitted by TCP subflow of the DSN for the poor level-one of network state than TCP subflow i be (i.e. position pri+1's
The DSN of data packet transmitted by TCP subflow).
Sending window is the window that all TCP subflows are used in conjunction with, the shared transmission buffering of all TCP subflows
Net can be effectively avoided by judging whether the DSN of data packet transmitted by TCP subflow i is greater than scheduler dispatches window in area
Data packet data transmitted by the preferable TCP subflow of receiving end clogging networks state that the poor TCP subflow of network state is sent
Packet.For example, TCP1, TCP2, TCP3 be window size be 10 subflow, and the quality of network state be ordered as TCP1, TCP2,
TCP3, the size for sending window is 40, and when TCP1 sends the data packet that DSN is 1~10, TCP2 sends the number that DSN is 21~30
According to packet, at this point, TCP3 needs to wait, and avoids TCP3 if the DSN for the data packet that TCP3 is sent has been more than transmission window 40
The data packet of transmission data packet transmitted by the obstruction TCP1 and TCP2 of receiving end, if the DSN for the data packet that TCP3 is sent exists
It sends within window, just TCP3 is allowed to send data packet, therefore, efficiently avoid not considering that DSN sends number directly to subflow
The problem of causing receiving end to block according to packet improves the efficiency of transmission of data packet, improves handling capacity.
Above disclosed is only a specific embodiment of the invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, can readily occur in variation or modification,
It is covered by the protection scope of the present invention.
Claims (9)
1. the MPTCP dispatching method in a kind of wireless network based on link-state information, which is characterized in that including following step
It is rapid:
Step 1: all TCP subflows being sorted according to the quality of network state, according to being ordered as, every TCP substream allocation is to be sent
Data packet, every TCP subflow send data packet according to DSN;
Step 2: updating position pri of the TCP subflow i in the sequence of all TCP subflow network state quality;
Step 3: if the network state of TCP subflow i be it is best, send DSN near preceding not sent data packet;If should
The network state of TCP subflow i be not it is best, then be transferred to step 4;
Step 4: judging whether to need to level-one TCP subflow reserved data packet more preferable than the TCP subflow i network state;If being not required to
Reserved data packet is wanted, and the last TCP subflow in position pri has not sent data packet, then sent last in position
The not sent data packet of the TCP subflow of pri is set, the data packet for otherwise selecting DSN bigger is sent;If desired reserved data packet, then
The DSN of the data packet of transmission is the quantity of current reservations data packet and the data that the last TCP subflow in position pri is sent
The sum of DSN of packet;
Step 5: step 2-4 is repeated, until all data packets are sent.
2. the MPTCP dispatching method based on link-state information in a kind of wireless network as described in claim 1, feature exist
In in the step 1, data packet to be sent includes reserved data packet, and the quantity of reserved data packet is network-like with the TCP subflow
State is related, and TCP subflow network state is better, and the quantity of reserved data packet is more.
3. the MPTCP dispatching method based on link-state information in a kind of wireless network as described in claim 1, feature exist
In in the step 2, TCP subflow i updates the TCP subflow i in institute in the ack msg Bao Shijun for receiving the TCP subflow every time
There is the position in the sequence of TCP subflow network state quality.
4. the MPTCP dispatching method based on link-state information in a kind of wireless network as described in claim 1, feature exist
In judging whether the concrete operations for needing reserved data packet in the step 4 are as follows: judge the quantity of last reserved data packet
The sum of whether be less than three times of the sum of last congestion window;
Wherein, the sum of the quantity of last reserved data packet refers to that the last TCP subflow in position pri is than TCP
Reserved the sum of the quantity of data packet of the good all TCP subflows of flow network state;The sum of last congestion window refers to than upper
The sum of the congestion window of the good all TCP subflows of the primary TCP subflow network state in position pri;
If so, needing reserved data packet;If it is not, not needing reserved data packet then.
5. the MPTCP dispatching method in a kind of wireless network as described in claim 1 or 4 based on link-state information, special
Sign is, in the step 4, needs the calculation formula of the quantity of reserved data packet are as follows:
Res=(RTTP × BWP+Rud ÷ CWNDpri-1)×(DSNpri-1-DSN0)
Wherein, Res indicates the quantity of current reservations data packet;SRTTpri-1Expression is currently at position
The smooth roundtrip time delay of the TCP subflow of pri-1, SRTTpriWhen expression is currently at the smooth roundtrip of TCP subflow of position pri
Prolong;BWP expression is currently at the ratio between the bandwidth of the TCP subflow and the TCP subflow for being currently at position pri of position pri-1, i.e.,Rud indicates the redundancy size of the data packet of current receiving end;CWNDpri-1Expression is currently at position
The congestion window of the TCP subflow of pir-1;DSNpri-1Indicate the data packet that the TCP subflow for being currently at position pir-1 is sent
DSN;DSN0Indicate the DSN for the data packet that the best TCP subflow of current network state is sent.
6. the MPTCP dispatching method based on link-state information in a kind of wireless network as claimed in claim 5, feature exist
In the smooth roundtrip time delay SRTT of the TCP subflow in position pripriCalculation formula are as follows:
SRTTpri=(α × SRTTHis, pri)+((1-α)×RTTpri)
Wherein, α indicates impact factor, for indicating to RTTpriAcceptance level, SRTTHis, priIt indicates last and is in position
The smooth roundtrip time delay of the TCP subflow of pri, RTTpriIndicate the round-trip delay of the TCP subflow in position pri.
7. the MPTCP dispatching method based on link-state information in a kind of wireless network as claimed in claim 5, feature exist
In the calculation formula of the redundancy size Rud of the data packet of current receiving end are as follows:
Rud=β × Rudhis+(1-β)×ΔRud
Wherein, RudhisIndicate that the backup length in last rx-side buffering area, β indicate that impact factor, Δ Rud indicate receiving end
The difference of buffer area current congestion length and last congestion length.
8. the MPTCP dispatching method based on link-state information in a kind of wireless network as claimed in claim 7, feature exist
In the specific formula for calculation of the difference Δ Rud of rx-side buffering area current congestion length and last congestion length are as follows:
Δ Rud=DSN0-DATAACK-Rudhis
Wherein, DSN0Indicate that the DSN for the data packet that the best TCP subflow of current network state is sent, DATAACK are MPTCP scheduling
The confirmation message of the data packet of TCP subflow rank in device.
9. the MPTCP dispatching method based on link-state information in a kind of wireless network as described in claim 1, feature exist
In, in the step 4, if desired reserved data packet, then also need to judge data packet transmitted by TCP subflow i DSN whether
Greater than scheduler dispatches window, if so, the TCP subflow i is waited, empty data packet is sent on TCP subflow i, then be transferred to institute
State step 2;
If it is not, then judging whether the DSN of data packet transmitted by TCP subflow i is greater than the TCP for being currently at position pri+1
The DSN of data packet transmitted by subflow;If so, the DSN for the data packet that the TCP subflow i is sent is the net than TCP subflow i
The DSN of data packet transmitted by the TCP subflow of network state difference level-one, if it is not, the DSN for the data packet that then the TCP subflow i is sent
The sum of the DSN of data packet sent for the quantity of current reservations data packet and the last TCP subflow in position pri.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910682837.1A CN110392394B (en) | 2019-07-26 | 2019-07-26 | MPTCP scheduling method based on link state information in wireless network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910682837.1A CN110392394B (en) | 2019-07-26 | 2019-07-26 | MPTCP scheduling method based on link state information in wireless network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110392394A true CN110392394A (en) | 2019-10-29 |
CN110392394B CN110392394B (en) | 2021-04-16 |
Family
ID=68287645
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910682837.1A Active CN110392394B (en) | 2019-07-26 | 2019-07-26 | MPTCP scheduling method based on link state information in wireless network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110392394B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111131017A (en) * | 2019-11-19 | 2020-05-08 | 中国科学院计算技术研究所 | MPTCP (Multi-protocol Transmission control protocol) cross-layer optimization method and system based on multi-cellular wireless access gateway |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103581035A (en) * | 2012-08-01 | 2014-02-12 | 华为技术有限公司 | Method, device and system for multi-path TCP congestion control |
US20150350337A1 (en) * | 2014-05-30 | 2015-12-03 | Apple Inc. | Long-Lived MPTCP Sessions |
EP3162027A1 (en) * | 2014-06-30 | 2017-05-03 | Orange | Multi-path tcp communication method between two terminals |
CN106656856A (en) * | 2016-12-16 | 2017-05-10 | 浙江大学 | Data package scheduling method for resisting blockage of MPTCP receiving buffer zone |
CN107566275A (en) * | 2017-10-18 | 2018-01-09 | 中南大学 | Multi-path transmission method based on the delay inequality opposite sex in data center network |
CN108540380A (en) * | 2017-03-02 | 2018-09-14 | 华为技术有限公司 | More subflow network transfer methods and device |
CN108881008A (en) * | 2017-05-12 | 2018-11-23 | 华为技术有限公司 | A kind of methods, devices and systems of data transmission |
CN109314666A (en) * | 2016-04-12 | 2019-02-05 | Nicira股份有限公司 | Virtual channel endpoint for congestion aware load balancing |
CN109565471A (en) * | 2016-06-18 | 2019-04-02 | 科里维网络有限公司 | High performance intelligent adaptive transport layer is proposed using multichannel |
-
2019
- 2019-07-26 CN CN201910682837.1A patent/CN110392394B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103581035A (en) * | 2012-08-01 | 2014-02-12 | 华为技术有限公司 | Method, device and system for multi-path TCP congestion control |
US20150350337A1 (en) * | 2014-05-30 | 2015-12-03 | Apple Inc. | Long-Lived MPTCP Sessions |
EP3162027A1 (en) * | 2014-06-30 | 2017-05-03 | Orange | Multi-path tcp communication method between two terminals |
CN109314666A (en) * | 2016-04-12 | 2019-02-05 | Nicira股份有限公司 | Virtual channel endpoint for congestion aware load balancing |
CN109565471A (en) * | 2016-06-18 | 2019-04-02 | 科里维网络有限公司 | High performance intelligent adaptive transport layer is proposed using multichannel |
CN106656856A (en) * | 2016-12-16 | 2017-05-10 | 浙江大学 | Data package scheduling method for resisting blockage of MPTCP receiving buffer zone |
CN108540380A (en) * | 2017-03-02 | 2018-09-14 | 华为技术有限公司 | More subflow network transfer methods and device |
CN108881008A (en) * | 2017-05-12 | 2018-11-23 | 华为技术有限公司 | A kind of methods, devices and systems of data transmission |
CN107566275A (en) * | 2017-10-18 | 2018-01-09 | 中南大学 | Multi-path transmission method based on the delay inequality opposite sex in data center network |
Non-Patent Citations (3)
Title |
---|
JIE ZHANG1,DAVID DONOFRIO,JOHN SHALF,MAHMUT KANDEMIR,AND MYOUNGS: ""Nvmmu: A non-volatile memory management unit for heterogeneous gpu-ssd architectures"", 《2015 INTERNATIONAL CONFERENCE ON PARALLEL ARCHITECTURE AND COMPILATION》 * |
刘山: ""基于链路特征的MPTCP子路径数据调度算法研究"", 《湖南大学工程硕士学位论文》 * |
李玮,张大方,黄昆,谢鲲: ""面向大数据处理的高精度多维计数布鲁姆过滤器"", 《电子学报》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111131017A (en) * | 2019-11-19 | 2020-05-08 | 中国科学院计算技术研究所 | MPTCP (Multi-protocol Transmission control protocol) cross-layer optimization method and system based on multi-cellular wireless access gateway |
Also Published As
Publication number | Publication date |
---|---|
CN110392394B (en) | 2021-04-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110139319B (en) | Routing method for minimizing transmission delay of high dynamic delay network | |
CA2882535C (en) | Control device discovery in networks having separate control and forwarding devices | |
US7633880B2 (en) | Access network device for managing queue corresponding to real time multimedia traffic characteristics and method thereof | |
US6480911B1 (en) | Grouping class sensitive queues | |
Radenkovic et al. | Congestion aware forwarding in delay tolerant and social opportunistic networks | |
WO2014077905A1 (en) | Multi-hop error recovery | |
CN101582842A (en) | Congestion control method and congestion control device | |
EP1303944A2 (en) | Multi-path dynamic routing algorithm | |
CN110351200B (en) | Opportunistic network congestion control method based on forwarding task migration | |
Ghaffari et al. | Energy-efficient and QoS-aware geographic routing protocol for wireless sensor networks | |
KR20110016749A (en) | Apparatus for multi-link data distribution/concentration to aggregate bandwidth in multi-connection wireless network | |
Lu | SED: An SDN-based explicit-deadline-aware TCP for cloud Data Center Networks | |
Ni et al. | OCPS: Offset compensation based packet scheduling mechanism for multipath TCP | |
JP4611319B2 (en) | Network architecture | |
Ye et al. | PTP: Path-specified transport protocol for concurrent multipath transmission in named data networks | |
Oida et al. | ARS: an efficient agent-based routing system for QoS guarantees | |
CN110392394A (en) | MPTCP dispatching method in wireless network based on link-state information | |
CN101119307A (en) | Routing method | |
Le et al. | A joint relay selection and buffer management scheme for delivery rate optimization in dtns | |
Liao et al. | A multi-path mechanism for reliable VoIP transmission over wireless networks | |
Nithin et al. | Efficient load balancing for multicast traffic in data center networks using SDN | |
CN115134304A (en) | Self-adaptive load balancing method for avoiding data packet disorder in cloud computing data center | |
Choi et al. | Research challenge on mptcp in 5g/6g networks | |
CN111756557B (en) | Data transmission method and device | |
KR100369562B1 (en) | Emulated weighted fair queueing algorithm for high-speed integrated service networks and the scheduler therefor |
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 |