CN105337708A - DTN network data transmission method using bundle block aggregation on dual-hop asymmetric channel - Google Patents

Abstract

The invention provides a DTN network data transmission method using bundle block aggregation on a dual-hop asymmetric channel. The method is characterized by repackaging, on the premise that the last-hop data is not completely absorbed, a part of data into a bundle according to a link characteristic when a link is communicated in a next hop by using bundle block technology, and accelerating transmission and increasing link utilization efficiency by using a block manner; by using bundle aggregation technology, aggregating multiple data packets to form a large data packet and then transmitting the large data packet. The method decreases the number of the data packets by means of the bundle aggregation so as to reduce the number of feedback packets, shorten data transmission time, and increase throughput. In addition, the method improves the structure of RS so as to enable the RS to change within a small range even in a high bit error rate case. The method effectively solves problems that a DTN is overlong in network transmission and storage waiting time, wastes transmission opportunities, and is low in throughput in the case that an upper link and a lower link are asymmetric in transmission rate and high in bit error rate.

Description

The DTN network data piecemeal aggregation strategy transmission method of double jump nonsymmetric channel

Technical field

The present invention relates to technical field of deep space communication, particularly relate to a kind of DTN network data convergence transmission method.

Background technology

Rong Chi/appearance circuit network (DTN, Delay/Disrupt-TolerantNetworking) equaled to propose for 2003 by Fall, realization for the Internet, space provides a kind of possible solution, main disclosed research institution is DTN research group (DTNRG, DTNResearchGroup) at present.Because interplanetary Internet comprises multiple different communication scenes, its propagation delay, channel conditions there are differences, deep space backbone network and close on the underlying protocol that the localized networks such as network adopt and be also not quite similar.DTN by introducing cover layer between application layer and local bottom-layer network---and bundlelayer provides data transport service for different application, realizes interconnection between localized network, and is supported in localized network and takes respective best performance agreement.Grow due to deep space link propagation delay simultaneously and do not fix, depend on the session interactions such as negotiation, query/response and be difficult to realize, message is packaged into cover layer data cell bundle by DTN application layer, adopts asynchronous transfer mode to carry out empty message forwarding.In order to ensure data transmitting in frequent interrupt network, DTN takes Store and forword pattern, and bundle is stored in permanent storage appliance in node until send when transmission opportunity arrives next time.

DTN protocol architecture mainly comprises BP agreement (bundleprotocol), LTP agreement (LickliderTransmissionProtocol), Saratoga agreement, LTP-T agreement (LTPTransport) and DTTP agreement (Delay-TolerantTransportProtocol) etc.DTN agreement has been used for reference first two protocol architecture and to be correlated with thought, has done corresponding modify, become Future Internet important development direction to its weak point.2008, U.S. NASA implements Deep Impact network test (DINET by name, DeepImpactNetworkExperiment), by Deep Impact spacecraft (DI, DeepImpact) complete the transfer of data of first time based on BP and LTP agreement, demonstrate the application of DTN agreement in spatial network.But it is shorter that DTN protocol architecture proposes the time, and in some Details Of Agreement, such as congestion control, optimized transmission, route, time synchronized etc. are perfect not, have wide Research Prospects.

BP (BundleProtocol) is an important protocol in DTN protocol architecture.BP protocols work, on the agreements such as LTP, TCP, UDP, is the agreement on agreement.The major function of BP agreement has: keeping transmission, end to end reliability, routing policy etc.LTP agreement is the reliable point-to-point host-host protocol of large for propagation delay time, frequent link design of interrupting.

Current DTN Protocol Design follows " STOREandforward " mechanism, and transmitting procedure biases toward storage (STORE), could start down hop transmission after via node only receives entire data.Bundle based on this thought receives and process of transmitting can be used accompanying drawing 1with accompanying drawing 2represent.

From fig. 1with accompanying drawing 2in can find out, " STOREandforward " ensure that the reliability that data are transmitted in the presence of a harsh environment, but owing to could starting the transmission of down hop after needing to wait for data bundle complete reception, even if also need wait when connection appears in down hop link.Transmission opportunity waste precious in limited network may be caused thus, reduce efficiency of transmission.Cause the main cause of this situation to be not mating in two sections link up duration and transmittability, profound reason is that existing DTN agreement is based on " STOREandforward ", lays particular emphasis on storage.

When introducing via node transmission data, owing to transmitting through upper and lower two-hop link in feedback packet (RS) process, transmission speed may be unequal, and the RS number can transmitted in same time is unequal, then may cause congested and postpone.RS may be caused like this can not to pass sending node back in time, and when partial data bag will wait for that its feedback packet arrives sending node, it just can carry out next step transmission, the stand-by period of the transfer of data of increase, reduces throughput.

When transmission range is comparatively far away, and introduce via node again, data transmission error rate may be very high.Under these circumstances, packet (bundle) polymerization must cause the increase of packet.The increase of packet makes packet packet loss and the probability of makeing mistakes increase, and therefore the size of each feedback packet (RS) also can become large, and so feedback packet RS makes mistakes and the probability of packet loss also increases.And because the difference of two sections link rate, RS will run into larger delay in transmitting procedure.

Summary of the invention

In order to solve the problems of the prior art, the present invention proposes a kind of DTN network data piecemeal convergence transmission method of double jump nonsymmetric channel, under the prerequisite that upper hop data do not receive completely, according to its link property, partial data is repackaged into bundle when next hop link is communicated with, utilize partitioned mode to accelerate transmission, improve link utilization efficiency, improve the throughput of transfer of data., improve the structure of RS, the transmission method that bundle is polymerized also improves the throughput of transfer of data when high bit-error meanwhile.

The present invention is achieved through the following technical solutions:

A DTN network data piecemeal convergence transmission method for double jump nonsymmetric channel, is applied to the double jump DTN network comprising source node, via node and destination node; Described method comprises: via node receives the packet that source node sends, wherein, and the not complete reception of described packet; In wait transmission phase, described via node does not need to receive bundle integrality and judges, and starts transmission when down hop transmission opportunity arrives; When link is communicated with, described via node obtains this connecting link information, arranges block data amount size, and is repacked by the packet received from source node and send to destination node; , modify to the structure of RS, even if make RS when high bit-error, its change is also not enough to the throughput affecting transfer of data meanwhile.

As a further improvement on the present invention, described via node obtains this connecting link information, arranges block data amount size, and is repacked by the packet received from source node and send to destination node to be specially:

Step 1: data are (BP) layer from application transfer to bundleprotocol;

Step 2: data are divided into n section at BP layer, then the data every f segment data being loaded BP layer

In bag bundle, data are packaged into after in bundle, are transferred to LTP layer;

Step 3: every m bundle is bundled in a LTP layer data bag block and transmits,

In the process of transmission, block is divided into k segments and transmits;

Step 4: receive a block and must feed back a RS to sending node.

As a further improvement on the present invention, the described structure to RS is modified and is specially: the feedback range binary representation in the reception of RS being stated.

As a further improvement on the present invention, described m is arranged according to different downstream transmission speed ratio and high bit-error situation.

The invention has the beneficial effects as follows: the DTN network data transmission method utilizing bundle piecemeal to be polymerized on double jump nonsymmetric channel that the present invention proposes, first bundle partition is utilized, under the prerequisite that upper hop data do not receive completely, according to its link property, partial data is repackaged into bundle when next hop link is communicated with, utilizes partitioned mode to accelerate transmission, improve link utilization efficiency.Then utilize bundle polymerization technique, multiple packet is aggregated in a large packet and transmits again.Reduced the number of packet by the polymerization of bundle, thus reduce the number of feedback packet, reduce data transmission period, improve throughput.Meanwhile, the structure of the present invention to RS is improved, even if make its size when the error rate is very high, also only changes among a small circle.It is long that the present invention efficiently solves current DTN Internet Transmission memory latency time to a certain extent, waste transmission opportunity, and the problem that two sections link transmission rate is asymmetric and in high bit-error situation, throughput reduces.

Accompanying drawing explanation

Fig. 1 is data transmission and reception schematic diagram under STOREandforward mechanism in prior art;

Fig. 2 is data transmission and reception flow chart under STOREandforward mechanism in prior art;

Fig. 3 is of the present invention based on block data transmission optimization policy data transmission and reception schematic diagram;

Fig. 4 is of the present invention based on block data transmission optimization policy data transmission and reception flow chart;

Fig. 5 is the segmentation of data in DTN protocol stack and polymerization process schematic diagram;

Fig. 6 is RS pack arrangement schematic diagram of the prior art;

Fig. 7 is RS pack arrangement schematic diagram of the present invention.

Embodiment

In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with accompanying drawingand embodiment, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

Junction link performance is optimized existing transmission mechanism, makes full use of transmission opportunity each time, and propose " storeandFORWARD " of the present invention scheme, the process of optimization is by means of piecemeal polymerization technique.Under the prerequisite that upper hop data do not receive completely, according to its link property, partial data is repackaged into packet (bundle) when next hop link is communicated with, utilizes partitioned mode to accelerate transmission, improve link utilization efficiency.Bundle based on piecemeal of the present invention receives and process of transmitting can be used accompanying drawing 3with accompanying drawing 4represent.

By accompanying drawing 3with accompanying drawing 4shown in, be mainly reflected in two places with conventional data transmission process difference: waiting for that transmission phase via node does not need to receive bundle integrality and judges, and start transmission when down hop transmission opportunity arrives; When link is communicated with, via node obtains this connecting link information, arranges message block data amount size, and data packing is sent.

Frequently interrupt for terminal in deep space, the connect hours is not enough, but there is knowability in advance, propose to analyze segments in the transmission of LTP protocol layer, calculate the time required for bundle transmission success, and then block transmission centering can be compared to break the impact of link data transfer.

Distance between spatial network interior joint and node is distant, so the communication link between node and node is unstable, is not suitable for carrying out transfer of data.But node is again in moment motion, and along with the change of topological structure, may originally become closely again apart from far nodal distance, be again attachable between them so this moment.Because space resources is precious especially, so the energy of each node is also limited, likely some node meeting timing is disabled every a period of time, and this also can make Space expanding change.Because node always be motion, so some node may be visible within a period of time.But after a period of time, because relative motion is blocked by some other node, make them invisible, the link interruption between these two nodes will be caused like this, but may they be again attachable as seen afterwards after a while.In spatial network, the connection between node and node may be mechanical periodicity, also may be non-mechanical periodicity.The present invention introduces via node in bundle transmits between the two nodes, analyzes and discusses on the link of space periodicity change, the situation that segments transmits at LTP layer.

If the transmission time is t ' after optimizing, before optimizing, the transmission time is t, and the total size of transfer files is m ₀, jumping transmission rate is up and down v1, v2, and the jumping propagation time is respectively T1, T2 up and down.The Lifetime of jumping in one-period is respectively T up and down _on1, T _on2, opening time is respectively T _off1, T _off2.

1. suppose that a complete cycle link can transfer a bundle, if next hop link UNICOM time is T ' when starting to transmit _on2, as T ' _on2during < 0, represent that link disconnects, if be Δ T opening time.

(1) time disconnected when next hop link is greater than transmission time of Article 1 link, i.e. Δ T > T ₁, the expression formula in transmission time can be obtained:

A. (Δ T-T ₁) V ₁< m ₀time,

$t^{\&prime;}=\mathrm{\&Delta;}T+\frac{m_0}{v_2}+T_2,t=\frac{m_0}{v_1}+T_1+\frac{m_0}{v_2}+T_2---\left(1\right)$

B. (Δ T-T ₁) V ₁>=m ₀time,

$t^{\&prime;}=t=\mathrm{\&Delta;}T+\frac{m_0}{v_2}+T_2---\left(2\right)$

(2) time disconnected when next hop link is greater than transmission time of Article 1 link, i.e. Δ T < T ₁, the expression formula in transmission time can be obtained:

$t^{\&prime;}=max(\frac{m_0}{v_1}+T_1+T_2,\frac{m_0}{v_2}+T_1+T_2),t=\frac{m_0}{v_1}+T_1+\frac{m_0}{v_2}+T_2---\left(3\right)$

(3) time be communicated with when next hop link is less than transmission time of Article 1 link, i.e. T ' _on2< T ₁, the expression formula in transmission time can be obtained:

$t^{\&prime;}=t=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0}{v_2}+T_2---\left(4\right)$

(4) time be communicated with when next hop link is greater than transmission time of Article 1 link, i.e. T ' _on2> T ₁, the expression formula in transmission time can be obtained:

A.v ₁(T ' _on2-T ₁) < m ₀and v ₂> v ₁

$t^{\&prime;}=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0-v_1(T_{on2}^{\&prime;}-T_1)}{V_2}+T_2,t=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0}{V_2}+T_2---\left(5\right)$

B.v ₁(T ' _on2-T ₁) < m ₀and v ₂< v ₁

$t^{\&prime;}=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0-v_2(T_{on2}^{\&prime;}-T_2)}{V_2}+T_2,t=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0}{V_2}+T_2---\left(6\right)$

C.v ₁(T ' _on2-T ₁)>=m ₀and v ₂(T ' _on2-T ₁) < m ₀

$t^{\&prime;}=T_{on2}^{\&prime;}+T_{off2}+\frac{m_0-v_2(T_{on2}^{\&prime;}-T_2)}{V_2}+T_2,t=\frac{m_0}{v_1}+T_1+\frac{m_0}{v_2}+T_2+T_{off2}---\left(7\right)$

D.v ₁(T ' _on2-T ₁)>=m ₀and v ₂(T ' _on2-T ₁)>=m ₀

$t^{\&prime;}=max(\frac{m_0}{v_1}+T_1+T_2,\frac{m_0}{v_2}+T_1+T_2),t=\frac{m_0}{v_1}+T_1+\frac{m_0}{v_2}+T_2---\left(8\right)$

2. can not transfer a bundle when a complete cycle link, then bundle transmission need through repeatedly cycle link.If the ratio of two sections isl cycle is b:a, i.e. a (T _on1+ T _off1)=b (T _on2+ T _off2).Transmission need satisfy condition:

m ₁(T _on1+T _off1)＜T ₁＜m ₁(T _on1+T _off1)+T _on1

m ₂(T _on2+T _off2)＜T ₂＜m ₂(T _on2+T _off2)+T _on2(9)

M ₁, m ₂be the integer being not less than 0, if n ₁, n ₂for the number of times that bundle will transmit at upper-lower section link.

(1) m is worked as ₁=0, m ₂=0, i.e. T ₁< T _on1, T ₂< T _on2time,

$n_1(T_{on1}-T_1)\&CenterDot;v_1=m_0\&DoubleRightArrow;n_1=\frac{m_0}{(T_{on1}-T_1)\&CenterDot;v_1}---\left(10\right)$

a.a(T _on1-T ₁)·v ₁≤b(T _on2-T ₂)·v ₂,

t'＝(n ₁-1)·T _off1+n ₁·T _on1+T _off2+T _on2(11)

b.a(T _on1-T ₁)·v ₁＞b(T _on2-T ₂)·v ₂,

$m_0-\frac{b}{a}(n_1-1)(T_{on2}-T_2)\&CenterDot;v_2=n_3(T_{on2}-T_2)\&CenterDot;v_2$

t'＝(n ₁-1)·T _off1+n ₁·T _on1+(n ₃-1)·T _off2+n ₃·T _on2(12)

And by conventional transmission strategy, the transmission time is:

t＝(n ₁-1)·T _off1+n ₁·T _on1+(n ₂-1)·T _off2+n ₂·T _on2(13)

(2) m is worked as ₁≠ 0 and m ₂when ≠ 0, the expression formula of upper and lower double bounce the number of transmissions can be obtained:

n ₁[(m ₁+1)T _on1-T ₁]·v ₁＝m ₀

n ₂[(m ₂+1)T _on2-T ₂]·v ₂＝m ₀(14)

a. $a(T_{on1}-\frac{1}{m_1+1}{T}_1)\&CenterDot;v_1\&le;b(T_{on2}-\frac{1}{m_2+1}{T}_2)\&CenterDot;v_2,$

t'＝(m ₁+1)[(n ₁-1)T _off1+n ₁T _on1](15)

b. $a(T_{on1}-\frac{1}{m_1+1}{T}_1)\&CenterDot;v_1>b(T_{on2}-\frac{1}{m_2+1}{T}_2)\&CenterDot;v_2,$

$m_0-\frac{b}{a}(n_1-1)(m_1+1)(T_{on2}-\frac{1}{m_2+1}{T}_2)\&CenterDot;v_2=n_3\&lsqb;(m_2+1)T_{on2}-T_2\&rsqb;\&CenterDot;v_2$

t'＝(m ₁+1)[(n ₁-1)·T _off1+n ₁·T _on1]+(m ₂+1)[(n ₃-1)·T _off2+n ₃·T _on2](16)

And by conventional transmission strategy, the transmission time is:

t＝(m ₁+1)[(n ₁-1)·T _off1+n ₁·T _on1]+(m ₂+1)[(n ₂-1)·T _off2+n ₂·T _on2](17)

Do not mate at two sections link rate for dark air transmission feedback packet (RS) and the delay that may cause and congestion problems, present invention employs the transmission optimization strategy of bundle polymerization.Its main thought is, bundle prepares transmission during from BP layer to LTP layer.In LTP layer, the bundle of identical destination can be bundled in a block and transmit again. accompanying drawing 5bundle polymerization and the process of transmission in DTN network.First first source documentpart can be divided into n bundle after arriving BP layer; After n bundle arrives LTP layer, because m bundle is bundled in a block by bundle polymerization, so every m bundle only needs a RS reply, then block is divided into some segments again and transmits.

In DTN network, being bundled to by the application data until of certain length in a protocol Data Unit, being referred to as bundle, is the roundtrip propagation number of times in order to reduce required for a protocol transaction like this.This situation larger to link propagation time delay, has a great impact.Its data bundle is passed to LTP layer by BP, and data cell is packaged into LTPblock.Again by all block segmentations, determine the size of LTP data segment according to the size of the MTU of link sen below, then data segment is packaged as frame in a link layer, and transmit in a data link layer.After data segment arrives receiving terminal, then they are reassembled into original block.At this moment receiving terminal must reply the report section (RS) that receives this block.RS representative be one receive successful ACK (so data all receive) or one receive unsuccessful ACK (needing to retransmit if some data does not receive).

The operation format of LTP acquiescence is that a block comprises a bundle.The asymmetric situation of height in order to the data rate adapting to interplanetary Internet link, the present invention proposes multiple bundle to be aggregated in a block instead of in a block to fill a bundle, thus reduces the number of the RS of feedback.And the reduction of RS number, the congested and delay etc. caused because two sections link rate does not mate can be reduced in theory, and then improve the utilance of link, also greatly can improve the throughput of transfer of data.

Be below the theory analysis of bundle polymerization:

First following symbol is defined:

N: one block is polymerized the number of bundle;

Goodput: effective throughput;

The round trip propagation time of RTT:bundle;

T _b: transmit the time that a bundle needs;

T _r: transmit the time that a RS needs;

T _b: the time of transmission required for a block after polymerization;

T _r': the time that after polymerization, transmission RS needs;

D _b: the size of data of a bundle;

D _head: the header packet information size of a bundle.

(1) first, to general bundle transmission policy, namely block fills the situation of a bundle, and the expression formula obtaining its goodput is:

$Goodput1=\frac{N*(D_b-D_{head})}{N*RTT+N*T_b+N*T_R}=\frac{D_b-D_{head}}{RTT+T_b+T_R}---\left(18\right)$

(2) in the situation of being polymerized based on bundle, install to the situation in a block by multiple bundle, the expression formula obtaining its goodput is:

$Goodput2=\frac{N*(D_b-D_{head})}{RTT+T_B+{T_R}^{\&prime;}}=N*\frac{D_b-D_{head}}{RTT+T_B+{T_R}^{\&prime;}}---\left(19\right)$

Suppose to have N number of bundle (formed objects and size is L _bundle) pass to LTP layer after, be packaged in a block, block is divided into several segments again to be transmitted.The default-length of the segments in LTP layer is L _seg, the error rate is BER, a segment (being also packet) mistakeprobability is P_E, then:

P_E＝1-(1-BER) ^s(S＝8*L _seg)(20)

L _Block＝N*L _bundle+L _head(21)

A block on average can be divided into n _segindividual segments transmission, then

n _seg＝L _Block/L _seg(22)

The part of RS comprises header packet information, also has the scope of the segments lost in transmitting procedure.Header packet information is fixing, and the part of variation is exactly the interval that the information of packet loss sorts in block.The mean size of header packet information is 8bytes, represents that the size required for an interval is probably 16bytes.

That present needs are determined is exactly the number N that in block, segments loses _error, drawn by following equation:

$N_{error}=\underset{i=0}{\overset{n}{\&Sigma;}}i*P\{N=i\}=\underset{i=0}{\overset{n}{\&Sigma;}}i*C_n^i*P\_E^i*{(1-P\_E)}^{n-i}---\left(23\right)$

Obtain the length of RS:

L _RS＝8+(N _error+1)*16(24)

Obtain RS's mistakeprobability is:

$P\_E\_RS=1-{(1-BER)}^{8*L_{RS}}---\left(25\right)$

Two important data can be drawn: a block transmission success needs the average time N of transmission by above formula _bLOCK, and the number of transmissions N that a RS transmission success needs _rS:

$\begin{array}{c}{N}_{BLOCK}=E\left\{N_{BLOCK}\right\}=\underset{J=0}{\overset{+\mathrm{\&infin;}}{\&Sigma;}}J*P\left\{N_{BLOCK}=J\right\}\\ =\underset{J=0}{\overset{+\mathrm{\&infin;}}{\&Sigma;}}J*(\underset{k=0}{\overset{J-1}{\&Sigma;}}\underset{j=0}{\overset{J-k-1}{\&Sigma;}}{C}_{J-k-1}^j*P\_E\_{\mathrm{RS}}^j{(1-P\_E\_RS)}^{J-k-j-1}\\ *C_{J-1}^{k}P\_E^k{\left(1-P\_E\right)}^{J-k}\&lsqb;{\left(1-P\_E^{J-k-j-1}\right)}^{n-1}-{\left(1-P\_E^{J-k-j-1}\right)}^{n-1}\&rsqb;\end{array}---\left(26\right)$

$\begin{array}{c}{N}_{RS}=\underset{k=1}{\overset{M}{\&Sigma;}}k*P\left\{N=k\right\}\\ =\underset{k=1}{\overset{M}{\&Sigma;}}k*\&lsqb;(1-P\_E\_{\mathrm{RS}}^k)-(1-P\_E-{\mathrm{RS}}^{k-1})\&rsqb;\end{array}---\left(27\right)$

Can know from above explanation, the polymerization of bundle can improve the utilance of link really, also greatly can improve the throughput of transfer of data.But in nonideal situation, namely when the error rate is higher, bundle polymerization not only can not bring the lifting of data transmission throughput, can make throughput degradation on the contrary.So the present invention proposes, the structure of RS is modified, make RS can not affect the throughput of the transmission of data along with the increase of the number of bundle polymerization.

RS primarily of report sequence number, checkpoint sequence number, upper confinement boundary, lower limit border, receive statement quantity and receive statement (displacement and length).And these parts except receive statement except, other be all fixing.Receive statement and mainly comprise two parts, displacement and length.Wherein, displacement refers to the start position successfully received from lower line boundary.Length refers to from the original position of displacement, the byte number of success reception continuously.That is, what feed back in RS is the part successfully received, and the part do not fed back is exactly the part needing to retransmit.

By receiving forming and can knowing of statement, each scope not receiving successful segment needs 4 bytes to represent.Large at Documents Comparison, the error rate is higher, when transmission channel speed ratio is smaller.The size of RS may be very large.

One receives statement needs 4 bytes and represents, namely needs 32 to represent.If represent that segments does not have the scope of transmission success with binary system, the figure place that so each scope needs will greatly reduce, and so just greatly can reduce the size receiving statement in RS.

By accompanying drawing 6can find out, except header part in current RS, primary feedback be the scope of the one section one section correct data of transmission.And the packet header of RS after improving is the same with the packet header setting of current RS, but the scope of feedback is then by binary representation, as accompanying drawing 7shown in, 0001 represent be the 1st, 1010 represent be the 10th.Do the size that greatly can reduce RS like this, RS can be avoided to make mistakes due to the increase of size, transmission quantity is increased.

Current RS is made up of RS packet header and scope, and the size in packet header is L _{head_RS}, each scope is by 8 byte representations, and namely 64, the expression formula of RS size is (unit of the size of RS is bytes):

L _RS1＝L _{head_RS}+(N _error+1)*8(28)

If N _bitfor representing figure place needed for this scope, the length that the present invention improves rear RS is:

L _RS2＝L _{head_R}+N _bit*(N _error+1)/8(29)

The improvement to RS adopting the present invention to propose, can ensure that bundle aggregation transfer one improves the throughput of transmission surely in perfect condition, low error rate and high bit-error situation.The deliverytime that bundle can be made to transmit is shorter, and the time delay namely making bundle transmit reduces.

Above content is in conjunction with concrete preferred implementation further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, some simple deduction or replace can also be made, all should be considered as belonging to protection scope of the present invention.

Claims (4)

1. a DTN network data piecemeal convergence transmission method for double jump nonsymmetric channel, is characterized in that: described method is applied to the double jump DTN network comprising source node, via node and destination node; Described method comprises: via node receives the packet that source node sends, wherein, and the not complete reception of described packet; In wait transmission phase, described via node does not need to receive bundle integrality and judges, and starts transmission when down hop transmission opportunity arrives; When link is communicated with, described via node obtains this connecting link information, arranges block data amount size, and is repacked by the packet received from source node and send to destination node; , modify to the structure of RS, even if make RS when high bit-error, its change is also not enough to the throughput affecting transfer of data meanwhile.

2. data aggregate transmission method according to claim 1, is characterized in that: described via node obtains this connecting link information, arranges block data amount size, and is repacked by the packet received from source node and send to destination node to be specially:

Step 1: data are from application transfer to bundleprotocol(BP) layer;

Step 2: data are divided at BP layer nsection, then will be every fsegment data loads in the packet bundle of BP layer, and data are packaged into after in bundle, are transferred to LTP layer;

Step 3: will be every mindividual bundle is bundled in a LTP layer data bag block and transmits, and in the process of transmission, block is divided into kindividual segments transmits;

Step 4: receive a block and must feed back a RS to sending node.

3. data aggregate transmission method according to claim 2, is characterized in that: the described structure to RS is modified and is specially: the feedback range binary representation in the reception of RS being stated.

4. data aggregate transmission method according to claim 2, is characterized in that: described in marrange according to different downstream transmission speed ratio and high bit-error situation.