[go: up one dir, main page]

CN102739508B - A kind of method and system of express network data transmission - Google Patents

A kind of method and system of express network data transmission Download PDF

Info

Publication number
CN102739508B
CN102739508B CN201110093822.5A CN201110093822A CN102739508B CN 102739508 B CN102739508 B CN 102739508B CN 201110093822 A CN201110093822 A CN 201110093822A CN 102739508 B CN102739508 B CN 102739508B
Authority
CN
China
Prior art keywords
rtt
stream
router
server
formula
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.)
Active
Application number
CN201110093822.5A
Other languages
Chinese (zh)
Other versions
CN102739508A (en
Inventor
王洪波
李阳阳
罗圣美
程时端
金怡爱
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.)
ZTE Corp
Beijing University of Posts and Telecommunications
Original Assignee
ZTE Corp
Beijing University of Posts and Telecommunications
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 ZTE Corp, Beijing University of Posts and Telecommunications filed Critical ZTE Corp
Priority to CN201110093822.5A priority Critical patent/CN102739508B/en
Publication of CN102739508A publication Critical patent/CN102739508A/en
Application granted granted Critical
Publication of CN102739508B publication Critical patent/CN102739508B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了一种高速网络数据传输的方法,该方法包括:服务器确定当前流的拥塞窗口超过门限值时,根据路由器获取的RTT平均值获取加式增加因子,并根据获取的加式增加因子调整拥塞窗口。本发明还公开了一种高速网络数据传输的系统,采用本发明所述的方法及系统,减少时延相对较小的流的带宽资源,提高时延较大的流的带宽资源,从而增加不同时延的流之间带宽占用率的公平性,提高高速带宽网络的服务质量。

The invention discloses a method for high-speed network data transmission. The method includes: when the server determines that the congestion window of the current flow exceeds a threshold value, obtains an additive increase factor according to the average value of RTT obtained by a router, and increases the factor according to the obtained additive value. Factor to adjust the congestion window. The present invention also discloses a system for high-speed network data transmission. By adopting the method and system described in the present invention, the bandwidth resources of streams with relatively small delays are reduced, and the bandwidth resources of streams with relatively large delays are increased, thereby increasing the The fairness of bandwidth occupancy between concurrently delayed flows improves the service quality of high-speed bandwidth networks.

Description

一种高速网络数据传输的方法及系统Method and system for high-speed network data transmission

技术领域 technical field

本发明涉及网络数据传输领域,特别是指一种高速网络数据传输的方法及系统。The invention relates to the field of network data transmission, in particular to a method and system for high-speed network data transmission.

背景技术 Background technique

随着网络技术的不断发展,出现了越来越多的带宽大于1Gbps的高速网络,这样的网络通常具有高带宽时延积(BDP,Bandwidth Delay Product)的特点,会带来非常大的拥塞窗口,传统的传输控制协议(TCP)已经不再适应网络发展的要求。其中,所述拥塞窗口是指一个往返时延(RTT)内最多发送的数据包数。With the continuous development of network technology, more and more high-speed networks with bandwidth greater than 1Gbps appear. Such networks usually have the characteristics of high bandwidth delay product (BDP, Bandwidth Delay Product), which will bring a very large congestion window. , the traditional transmission control protocol (TCP) is no longer suitable for the requirements of network development. Wherein, the congestion window refers to the maximum number of data packets sent within one round-trip time delay (RTT).

目前,标准的TCP协议使用加式增加乘式下降(AIMD)算法来调整拥塞窗口w的大小:Currently, the standard TCP protocol uses the Additive Increase Multiplicative Descent (AIMD) algorithm to adjust the size of the congestion window w:

在拥塞避免阶段,还没有出现拥塞,可以通过公式(1)来调整拥塞窗口大小, In the congestion avoidance phase, there is no congestion yet, and the congestion window size can be adjusted by formula (1),

在拥塞阶段,因为出现拥塞,需要减少数据包发送的数量,通过公式(3)调整拥塞窗口大小,w←w-bw(2)。In the congestion phase, due to congestion, the number of data packets sent needs to be reduced, and the congestion window size is adjusted by formula (3), w←w-bw(2).

在慢启动阶段,需要快速提高带宽速率,可以通过慢启动参数以及公式(2)来调整拥塞窗口大小,w←w+c(3);In the slow start phase, it is necessary to quickly increase the bandwidth rate, and the congestion window size can be adjusted through the slow start parameter and formula (2), w←w+c(3);

其中,根据TCP协议中规定,加式增加(AI)因子a=1,乘式下降(MD)因子b=0.5,慢启动参数c=1,a、b、c的单位为最大报文段长度(MSS)。Wherein, according to the provisions in the TCP protocol, the additive increase (AI) factor a=1, the multiplicative decrease (MD) factor b=0.5, the slow start parameter c=1, and the unit of a, b, and c is the maximum segment length (MSS).

但是,根据上述方法调整拥塞窗口过程中会存在一个问题,假设由于丢包拥塞窗口减少到了w=1,发送端进入了拥塞避免阶段,网络中的服务器通过公式(1)线性的增加拥塞窗口;如果拥塞窗口要达到X,每次拥塞窗口增加1,则需要X-1个往返时间。这样的话,如果想要达到10Gbps的速度,假设往返时间是100ms,10Gbps对应的窗口值为83333,则需要83332次往返时间,那么达到上述速度需要8333秒,这在实际应用中是不能接受的。However, there will be a problem in adjusting the congestion window process according to the above method, assuming that the congestion window is reduced to w=1 due to packet loss, the sending end enters the congestion avoidance stage, and the server in the network increases the congestion window linearly by formula (1); If the congestion window is to reach X, each time the congestion window increases by 1, it takes X-1 round trip times. In this case, if you want to achieve a speed of 10Gbps, assuming that the round-trip time is 100ms, and the window value corresponding to 10Gbps is 83333, then 83332 round-trip times are required, and it takes 8333 seconds to reach the above-mentioned speed, which is unacceptable in practical applications.

针对TCP协议中AIMD算法在高带宽时延积网络环境下,存在的响应缓慢以及窗口扩展性缓慢的问题,Floyd提出了一种基于包丢失的AIMD窗口调整算法,即高速TCP(HSTCP)算法。HSTCP算法的主要思想是:设定一个拥塞窗口门限值wlow,当拥塞窗口值w超过wlow时,根据公式(1)及公式(3),将加式增加因子和乘式下降因子设置为w的函数,即a(w)和b(w)。这样,加式增加因子a(w)随着w的增大而增大,乘式下降因子b(w)随着w的增大而减小,a(w)与b(w)呈线性变化。如此,就避免了在高速网络环境下,拥塞窗口缓慢的线性增长;并且,在出现拥塞时,避免大幅度的降低慢启动的阈值。Aiming at the problems of slow response and slow window expansion of the AIMD algorithm in the TCP protocol in a high-bandwidth-delay product network environment, Floyd proposed an AIMD window adjustment algorithm based on packet loss, namely the High Speed TCP (HSTCP) algorithm. The main idea of the HSTCP algorithm is: set a congestion window threshold value w low , when the congestion window value w exceeds w low , according to formula (1) and formula (3), set the additive increase factor and the multiplicative decrease factor is a function of w, namely a(w) and b(w). In this way, the additive increase factor a(w) increases with the increase of w, the multiplicative decrease factor b(w) decreases with the increase of w, and a(w) and b(w) change linearly . In this way, the slow linear growth of the congestion window in a high-speed network environment is avoided; and, when congestion occurs, a large reduction in the slow start threshold is avoided.

但是,HSTCP算法存在一个问题:不同RTT的流在传输过程中存在带宽利用率不公平的问题,时延较小的流拥塞窗口增加很快,会抢占带宽资源。However, there is a problem with the HSTCP algorithm: the flow of different RTTs has the problem of unfair bandwidth utilization during the transmission process, and the congestion window of the flow with small delay increases rapidly, which will seize bandwidth resources.

根据分析,不同RTT的HSTCP流相互竞争时的吞吐量关系为:According to the analysis, the throughput relationship when HSTCP flows with different RTTs compete with each other is:

TT 11 TT 22 == WW 11 // RTTRTT 11 WW 22 // RTTRTT 22 == [[ RTTRTT 22 RTTRTT 11 ]] 11 11 -- dd

其中,根据HSTCP算法规定,d为0.82,假设存在2条流,并且RTT1∶RTT2=1∶2,那么,两条流的吞吐量之比T1∶T2=1∶47。Wherein, according to the HSTCP algorithm, d is 0.82, assuming that there are 2 streams, and RTT1:RTT2=1:2, then the throughput ratio of the two streams is T1:T2=1:47.

可以看出,HSTCP算法中,不同时延的数据流在竞争时存在明显的不公平性,这对时延比较大的用户服务质量造成较大的影响,不能享有高速带宽的优势。It can be seen that in the HSTCP algorithm, there is obvious unfairness in the competition between data streams with different delays, which has a great impact on the service quality of users with relatively large delays, and cannot enjoy the advantages of high-speed bandwidth.

发明内容 Contents of the invention

有鉴于此,本发明的主要目的在于提供一种高速网络数据传输的方法及系统,能减少时延相对较小的流的带宽资源,提高时延较大的流的带宽资源,从而增加不同时延的流之间带宽占用率的公平性,提高高速带宽网络的服务质量。In view of this, the main purpose of the present invention is to provide a method and system for high-speed network data transmission, which can reduce the bandwidth resources of streams with relatively small delays and increase the bandwidth resources of streams with large delays, thereby increasing the The fairness of bandwidth occupancy between delayed flows improves the service quality of high-speed bandwidth networks.

为达到上述目的,本发明的技术方案是这样实现的:In order to achieve the above object, technical solution of the present invention is achieved in that way:

本发明提供了一种高速网络数据传输的方法,该方法包括:The invention provides a method for high-speed network data transmission, the method comprising:

服务器确定当前流的拥塞窗口超过门限值时,根据路由器获取的往返时延RTT平均值获取加式增加因子,并根据获取的加式增加因子调整拥塞窗口。When the server determines that the congestion window of the current flow exceeds the threshold value, it obtains an additive increase factor according to the average round-trip delay RTT obtained by the router, and adjusts the congestion window according to the obtained additive increase factor.

上述方案中,所述根据路由器获取的RTT平均值获取加式增加因子之前,该方法进一步包括:路由器记录接收流的RTT,确定有流的RTT更新,获取RTT平均值。In the above solution, before obtaining the additive increase factor according to the average RTT obtained by the router, the method further includes: the router records the RTT of the received flow, determines that there is an RTT update of the flow, and obtains the average RTT.

上述方案中,所述确定有流的RTT更新,获取RTT平均值包括:路由器实时获取自身接收流的RTT,有流的RTT变化时,则确定有流的RTT更新,更新路由器时延统计表中所述流的RTT,并根据RTT平均值计算公式及路由器时延统计表中的RTT,获取RTT平均值。In the above scheme, the determination of the RTT update of the flow and obtaining the RTT average value include: the router obtains the RTT of the flow received by itself in real time, and when the RTT of the flow changes, it is determined that the RTT update of the flow is updated, and the router delay statistics table is updated. The RTT of the flow, and according to the RTT average calculation formula and the RTT in the router delay statistics table, obtain the RTT average value.

上述方案中,所述服务器确定当前流的拥塞窗口超过门限值之前,该方法进一步包括:服务器向路由器发送流,记录并更新服务器时延统计单元中自身发送流的RTT。In the above solution, before the server determines that the congestion window of the current flow exceeds the threshold value, the method further includes: the server sends the flow to the router, records and updates the RTT of the flow sent by itself in the server delay statistics unit.

上述方案中,所述根据RTT平均值获取加式增加因子包括:根据服务器时延统计单元中流的RTT、RTT平均值及公平因子计算公式,获取所述流对应的公平因子,将公平因子与高速传输控制协议HSTCP协议中的加式增加因子相乘,获取新的加式增加因子。In the above scheme, the acquisition of the additive increase factor based on the RTT average value includes: obtaining the fairness factor corresponding to the flow according to the RTT, the RTT average value, and the fairness factor calculation formula of the flow in the server delay statistics unit, and combining the fairness factor with the high-speed Multiply the additive increase factors in the transmission control protocol HSTCP to obtain a new additive increase factor.

本发明还提供了一种高速网络数据传输的系统,该系统包括:窗口调整模块、参数获取模块;The present invention also provides a high-speed network data transmission system, which includes: a window adjustment module and a parameter acquisition module;

参数获取模块,用于获取RTT平均值,发送给窗口调整模块;The parameter acquisition module is used to obtain the RTT average value and send it to the window adjustment module;

窗口调整模块,用于在确定当前流的拥塞窗口超过门限值时,根据RTT平均值获取加式增加因子,并根据获取的加式增加因子调整拥塞窗口。The window adjustment module is configured to obtain an additive increase factor according to the RTT average value and adjust the congestion window according to the acquired additive increase factor when it is determined that the congestion window of the current flow exceeds a threshold value.

上述方案中,所述窗口调整模块包括:服务器时延更新单元、加式增加因子获取单元;服务器时延更新单元,记录自身发送流的RTT,保存在服务器时延统计单元中,更新服务器时延统计单元中的RTT;加式增加因子获取单元,用于在确定当前流的拥塞窗口超过门限值时,根据服务器时延统计单元中流的RTT、RTT平均值以及公平因子计算公式,获取所述流对应的公平因子,将公平因子与HSTCP协议中的加式增加因子相乘,获取新的加式增加因子。In the above scheme, the window adjustment module includes: a server delay update unit, an additive increase factor acquisition unit; a server delay update unit, which records the RTT of the flow sent by itself, stores it in the server delay statistics unit, and updates the server delay The RTT in the statistical unit; the additive increase factor acquisition unit is used to obtain the RTT, RTT average value and fair factor calculation formula of the flow in the server delay statistical unit when determining that the congestion window of the current flow exceeds the threshold value. The fairness factor corresponding to the flow. Multiply the fairness factor with the additive increase factor in the HSTCP protocol to obtain a new additive increase factor.

上述方案中,所述参数获取模块包括:路由器时延更新单元、平均值获取单元;路由器时延更新单元,用于记录接收流的RTT,保存在路由器时延统计表中,确定有流的RTT更新,通知平均值获取单元。In the above scheme, the parameter acquisition module includes: router delay update unit, average value acquisition unit; router delay update unit, used to record the RTT of the received flow, save it in the router delay statistics table, and determine the RTT of the flow Update, notify the average value acquisition unit.

上述方案中,所述路由器时延更新单元,具体用于实时获取自身接收流的RTT,有流的RTT变化时,则确定有流的RTT更新,更新路由器时延统计表中所述流的RTT;平均值获取单元,用于根据路由器时延统计表中的RTT及RTT平均值计算公式,获取RTT平均值。In the above scheme, the router delay update unit is specifically used to obtain the RTT of the flow received by itself in real time, and when the RTT of the flow changes, it is determined that the RTT of the flow is updated, and the RTT of the flow described in the router delay statistics table is updated. ; The average value obtaining unit is used to obtain the RTT average value according to the RTT and the RTT average value calculation formula in the router delay statistics table.

由此可见,采用本发明所述的方法及系统,通过时延的平均值来调整加式增加因子,从而使时延相对较小的流的窗口增长较小,使时延相对较大的流的窗口增长较大,减少时延相对较小的流的带宽资源,提高时延较大的流的带宽资源,从而增加不同时延的流之间带宽占用率的公平性,提高高速带宽网络的服务质量。It can be seen that, using the method and system of the present invention, the additive increase factor is adjusted through the average value of time delay, so that the window growth of the flow with relatively small time delay is small, and the window of the flow with relatively large time delay The window increases greatly, reducing the bandwidth resources of flows with relatively small delays and increasing the bandwidth resources of flows with large delays, thereby increasing the fairness of bandwidth occupancy among flows with different delays, and improving the efficiency of high-speed bandwidth networks. service quality.

附图说明 Description of drawings

图1为本发明实现高速网络数据传输的方法实施例示意图;Fig. 1 is a schematic diagram of an embodiment of a method for realizing high-speed network data transmission in the present invention;

图2为数据传输应用场景示意图;Figure 2 is a schematic diagram of a data transmission application scenario;

图3为本发明实现路由器获取RTT平均值的方法实施例示意图;Fig. 3 is the schematic diagram of the embodiment of the method that realizes that router obtains RTT average value in the present invention;

图4为本发明实现高速网络数据传输的系统实施例示意图。FIG. 4 is a schematic diagram of an embodiment of a system for realizing high-speed network data transmission according to the present invention.

具体实施方式 detailed description

本发明的基本思想是:服务器确定当前流的拥塞窗口超过门限值时,根据路由器获取的RTT平均值,获取加式增加因子,并根据获取的加式增加因子调整拥塞窗口。The basic idea of the present invention is: when the server determines that the congestion window of the current flow exceeds the threshold value, it obtains the additive increase factor according to the RTT average value obtained by the router, and adjusts the congestion window according to the acquired additive increase factor.

下面通过具体实施例与附图来对本发明进行详细说明,本发明提供了一种高速网络数据传输的方法,如图1所示,具体步骤如下:The present invention will be described in detail below through specific embodiments and accompanying drawings. The present invention provides a method for high-speed network data transmission, as shown in Figure 1, and the specific steps are as follows:

步骤101、服务器判断当前流的拥塞窗口是否超过门限值wlow,若超过,则执行步骤102;否则,执行步骤104;Step 101, the server judges whether the congestion window of the current flow exceeds the threshold w low , if so, executes step 102; otherwise, executes step 104;

这里,所述门限值wlow为HSTCP协议规定的值;Here, the threshold w low is a value specified by the HSTCP protocol;

本发明实施例中,网络数据传输应用场景示意图如图2所示,服务器将一条或多条流发送给路由器,由路由器发送给用户,多个服务器组成服务器集群,路由器同时接收服务器集群发送的多条流。In the embodiment of the present invention, a schematic diagram of network data transmission application scenarios is shown in Figure 2. The server sends one or more streams to the router, and the router sends them to the user. Multiple servers form a server cluster, and the router receives multiple streams sent by the server cluster at the same time. flow.

通常,服务器向路由器发送流,记录并更新服务器时延统计单元中自身发送流的RTT,判断当前流的拥塞窗口是否超过门限值,所述服务器时延统计单元为保存服务器发送的流及对应的RTT的对应关系的一种软件存储形式;所述记录服务器时延统计单元中自身发送流的RTT包括:服务器根据TCP协议实时读取自身发送的流的RTT,并保存在服务器时延统计单元中,若有新的流发出,则用新的服务器时延统计单元保存新的流及对应的RTT;所述更新服务器时延统计单元中自身发送流的RTT包括:服务器读取自身发送的流的RTT,若有流的RTT变化,则将与所述流对应的服务器时延统计单元中所述流对应的RTT的值更新为变化后的值。Usually, the server sends streams to the router, records and updates the RTT of the streams sent by itself in the server delay statistics unit, and judges whether the congestion window of the current stream exceeds the threshold value. The server delay statistics unit saves the streams sent by the server and the corresponding A kind of software storage form of the corresponding relationship of the RTT; The RTT of the self-sending flow in the recording server delay statistics unit includes: the server reads the RTT of the flow sent by itself in real time according to the TCP protocol, and saves it in the server delay statistics unit , if a new stream is sent, then use the new server delay statistics unit to save the new stream and the corresponding RTT; the RTT of the self-sent stream in the update server delay statistics unit includes: the server reads the stream sent by itself If the RTT of the flow changes, the RTT value corresponding to the flow in the server delay statistics unit corresponding to the flow is updated to the changed value.

具体的,所述判断当前流的拥塞窗口是否超过门限值wlow包括:服务器读取自身发送的流的窗口的数量,即:拥塞窗口的大小,与HSTCP协议规定的门限值wlow比较,若自身发送的流的拥塞窗口大小超过门限值,则说明当前链路带宽较高,需要调整拥塞窗口,执行步骤102;否则,仍按照标准TCP协议传输。Specifically, the judging whether the congestion window of the current flow exceeds the threshold value w low includes: the server reads the number of windows of the flow sent by itself, that is: the size of the congestion window, and compares it with the threshold value w low specified by the HSTCP protocol , if the congestion window size of the flow sent by itself exceeds the threshold value, it means that the current link bandwidth is high, and the congestion window needs to be adjusted, and step 102 is performed; otherwise, the transmission is still carried out according to the standard TCP protocol.

步骤102、根据路由器获取的RTT平均值,获取加式增加因子;Step 102, according to the RTT average value obtained by the router, obtain the additive increase factor;

服务器接收路由器发送的所有流的RTT平均值β,根据β以及公平因子计算公式,即公式(4),获取公平因子η,所述公平因子计算公式为:The server receives the RTT average value β of all flows sent by the router, and obtains the fair factor η according to β and the fair factor calculation formula, i.e. formula (4), and the fair factor calculation formula is:

ηη == [[ RTTRTT ii ββ ]] 11 dd -- -- -- (( 44 ))

其中,RTTi为服务器时延统计单元中流对应的RTT;Among them, RTT i is the RTT corresponding to the flow in the server delay statistics unit;

然后,将所述公平因子与HSTCP协议中的加式增加因子相乘,获取新的加式增加因子,即:a(w)=a(w)*ηThen, the fairness factor is multiplied with the additive increase factor in the HSTCP protocol to obtain a new additive increase factor, namely: a(w)=a(w)*η

其中,路由器获取RTT平均值的过程如图3所示,包括以下步骤:Wherein, the process of the router obtaining the RTT average value is shown in Figure 3, including the following steps:

步骤102a、路由器记录接收流的RTT;Step 102a, the router records the RTT of the received flow;

具体的,所述记录包括:路由器接收由服务器集群发送的流,获取流的RTT并保存至本地的路由器时延统计表中,若接收的流有新增,则获取新增流的RTT,并将所述流的RTT添加到路由器时延统计表中;所述获取流的RTT可以采用多种方法:可以由服务器在发送的流中携带所述流的RTT,路由器直接读取所述流中的RTT;也可以是路由器根据TCP协议的规定读取所述流的RTT,在此不一一列举。其中,所述路由器时延统计表为路由器接收的流及相应RTT的对应关系表。Specifically, the record includes: the router receives the flow sent by the server cluster, obtains the RTT of the flow and saves it in the local router delay statistics table, if the received flow is newly added, obtains the RTT of the newly added flow, and Add the RTT of the flow to the router delay statistics table; the RTT of the acquisition flow can adopt multiple methods: the server can carry the RTT of the flow in the flow sent, and the router directly reads the RTT of the flow in the flow The RTT of the flow; it may also be the RTT of the router reading the flow according to the provisions of the TCP protocol, which will not be listed here. Wherein, the router delay statistics table is a correspondence table of flows received by the router and corresponding RTTs.

步骤102b、路由器判断是否有流的RTT更新;Step 102b, the router judges whether there is an RTT update of the flow;

具体的,所述判断包括:路由器实时获取自身接收的流的RTT,若有流的RTT变化,则说明RTT有更新,更新路由器时延统计表中所述流的RTT;所述更新包括:将路由器时延统计表中,有RTT变化的流对应的RTT变更为更新后的RTT,然后执行步骤102c;否则,执行步骤102a;Specifically, the judgment includes: the router obtains the RTT of the flow received by itself in real time, if there is a change in the RTT of the flow, it indicates that the RTT has been updated, and the RTT of the flow in the router delay statistics table is updated; the update includes: In the router delay statistics table, the RTT corresponding to the flow with RTT change is changed to the updated RTT, and then step 102c is executed; otherwise, step 102a is executed;

步骤102c、计算RTT平均值。Step 102c, calculating the average value of RTT.

具体的,读取路由器时延统计表中的所有流的RTT,根据RTT平均值计算公式,即公式(5),获取RTT平均值,并将获取的RTT平均值发送给服务器。Specifically, read the RTTs of all flows in the router delay statistics table, obtain the average RTT value according to the formula (5) for calculating the average RTT value, and send the obtained average RTT value to the server.

ββ == ΣΣ ii == 11 nno RTTRTT ii nno -- -- -- (( 55 ))

步骤103、根据获取的加式增加因子调整拥塞窗口,结束当前处理流程;Step 103, adjust the congestion window according to the obtained additive increase factor, and end the current processing flow;

具体的,服务器将加式增加因子设置为所述根据RTT平均值获取的加式增加因子,也就是新的加式增加因子,就可自行调整拥塞窗口。这样,对于时延较小的流,其加式增加因子也较小,相对于时延较大的流,其拥塞窗口的增长较小,从而提高使时延较长的流的带宽利用率。Specifically, the server sets the additive increase factor as the additive increase factor obtained according to the average value of the RTT, that is, the new additive increase factor, and can adjust the congestion window by itself. In this way, for a flow with a small delay, its additive increase factor is also small, and compared with a flow with a large delay, the increase of its congestion window is small, thereby improving the bandwidth utilization rate of a flow with a long delay.

步骤104:仍根据标准TCP协议向路由器传输流。Step 104: still transmit the flow to the router according to the standard TCP protocol.

基于以上方法,本发明还提供了一种高速网络数据传输的系统,该系统包括:窗口调整模块401、参数获取模块402;其中,窗口调整模块401位于服务器中,参数获取模块402位于路由器中或路由器出口处;Based on the above method, the present invention also provides a system for high-speed network data transmission, the system includes: a window adjustment module 401, a parameter acquisition module 402; wherein, the window adjustment module 401 is located in the server, and the parameter acquisition module 402 is located in the router or Router exit;

窗口调整模块401,用于确定当前流的拥塞窗口超过门限值,根据参数获取模块获取的RTT平均值,获取加式增加因子,并根据获取的加式增加因子调整拥塞窗口;The window adjustment module 401 is used to determine that the congestion window of the current flow exceeds the threshold value, obtains the additive increase factor according to the RTT average value obtained by the parameter acquisition module, and adjusts the congestion window according to the acquired additive increase factor;

参数获取模块402,用于获取RTT平均值,并发送给窗口调整模块401。The parameter acquisition module 402 is configured to acquire the average value of the RTT and send it to the window adjustment module 401 .

所述窗口调整模块401包括:服务器时延更新单元4011、加式增加因子获取单元4012;The window adjustment module 401 includes: a server delay update unit 4011, an additive increase factor acquisition unit 4012;

服务器时延更新单元4011,用于记录并更新服务器时延统计单元中自身发送流的RTT;所述记录包括:服务器根据TCP协议实时读取自身发送的流的RTT,并保存在服务器时延统计单元中,若有新的流发出,则用新的服务器时延统计单元保存新的流及对应的RTT;所述更新包括:服务器读取自身发送的流的RTT,若有流的RTT变化,则将与所述流对应的服务器时延统计单元中所述流对应的RTT的值更新为变化后的值;The server delay update unit 4011 is used to record and update the RTT of the stream sent by itself in the server delay statistics unit; the record includes: the server reads the RTT of the stream sent by itself in real time according to the TCP protocol, and saves it in the server delay statistics In the unit, if there is a new flow to send out, then use the new server delay statistics unit to save the new flow and the corresponding RTT; the update includes: the server reads the RTT of the flow sent by itself, if there is a change in the RTT of the flow, Then update the RTT value corresponding to the flow in the server delay statistics unit corresponding to the flow to the changed value;

加式增加因子获取单元4012,用于在确定当前流的拥塞窗口超过门限值时,根据RTT平均值、服务器时延统计单元中的流及对应的RTT,以及公平因子计算公式,获取流的公平因子,将公平因子与HSTCP协议中的加式增加因子相乘,获取新的加式增加因子。The additive increase factor obtaining unit 4012 is used to obtain the flow rate according to the RTT average value, the flow in the server delay statistics unit and the corresponding RTT, and the fairness factor calculation formula when it is determined that the congestion window of the current flow exceeds the threshold value. Fairness factor. Multiply the fairness factor with the additive increase factor in the HSTCP protocol to obtain a new additive increase factor.

所述参数获取模块402包括:路由器时延更新单元4021、平均值获取单元4022;The parameter acquisition module 402 includes: a router delay update unit 4021, an average value acquisition unit 4022;

路由器时延更新单元4021,用于记录接收流的往返时延RTT,确定有流的RTT更新,通知平均值获取单元4022。The router delay update unit 4021 is configured to record the round-trip delay RTT of the received flow, determine that there is an RTT update of the flow, and notify the average value acquisition unit 4022 .

平均值获取单元4022,用于根据路由器时延统计表中记录的RTT及平均值计算公式,获取RTT平均值。The average value acquisition unit 4022 is configured to obtain the average value of RTT according to the RTT recorded in the router delay statistics table and the average calculation formula.

路由器时延更新单元4021具体用于,路由器实时获取自身接收流的RTT,有流的RTT变化时,则确定有流的RTT更新,更新路由器时延统计表中所述流的RTT;所述更新包括:将路由器时延统计表中,该流对应的RTT变更为更新后的RTT。The router time delay updating unit 4021 is specifically used to obtain the RTT of the flow received by the router in real time. When the RTT of the flow changes, it is determined that the RTT of the flow is updated, and updates the RTT of the flow in the router time delay statistics table; Including: changing the RTT corresponding to the flow in the router delay statistics table to the updated RTT.

以上所述,仅为本发明的较佳实施例而已,并非用于限定本发明的保护范围,凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等,均应包含在本发明的保护范围之内。The above description is only a preferred embodiment of the present invention, and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention shall be included in the within the protection scope of the present invention.

Claims (6)

1. the method for an express network data transmission, it is characterised in that the method includes:
Server sends stream to router, records and updates self transmission stream in server delay statistics unit Round-trip delay RTT;
Server determines when the congestion window of current stream exceedes threshold value, according in server delay statistics unit The RTT meansigma methods of the RTT of stream, router acquisition and fairness index computing formulaObtain institute State the fairness index η that stream is corresponding, by fairness index with high speed processing system HSTCP agreement adds formula Increasing fac-tor, obtaining the new formula that adds increases the factor, and adjusts congested according to the formula increase factor that adds obtained Window, wherein, described d=0.82.
Method the most according to claim 1, it is characterised in that the described RTT obtained according to router Meansigma methods obtains before adding the formula increase factor, and the method farther includes:
Router records receives the RTT of stream, determines that the RTT having stream updates, and obtains RTT meansigma methods.
Method the most according to claim 2, it is characterised in that described determine that the RTT having stream updates, Obtain RTT meansigma methods to include:
Router obtains self in real time and receives the RTT of stream, when having the RTT change of stream, it is determined that have stream RTT updates, more the RTT of stream described in new router delay statistics table, and public according to RTT mean value calculation RTT in formula and router delay statistics table, obtains RTT meansigma methods.
4. the system of an express network data transmission, it is characterised in that this system includes: window adjusts mould Block, parameter acquisition module;
Parameter acquisition module, is used for obtaining RTT meansigma methods, is sent to window adjustment module;
Server time delay updating block, sends the RTT of stream for recording self, is saved in server time delay system In meter unit, update the RTT in server delay statistics unit;
Add formula and increase factor acquirement unit, during for exceeding threshold value at the congestion window determining current stream, root According to RTT, RTT meansigma methods flowed in server delay statistics unit and fairness index computing formulaObtain the fairness index η that described stream is corresponding, by fairness index and adding in HSTCP agreement Formula increases fac-tor, and obtaining the new formula that adds increases the factor, and gathers around according to the formula increase factor adjustment that adds obtained Plug window, wherein, described d=0.82.
System the most according to claim 4, it is characterised in that described parameter acquisition module includes: road By device time delay updating block, meansigma methods acquiring unit;
Router time delay updating block, for recording the RTT receiving stream, is saved in router delay statistics table In, determine that the RTT having stream updates, notify meansigma methods acquiring unit.
System the most according to claim 5, it is characterised in that
Described router time delay updating block, receives the RTT of stream specifically for obtaining self in real time, has stream During RTT change, it is determined that have the RTT of stream to update, the more RTT of stream described in new router delay statistics table;
Meansigma methods acquiring unit, based on according to RTT and the RTT meansigma methods in router delay statistics table Calculate formula, obtain RTT meansigma methods.
CN201110093822.5A 2011-04-14 2011-04-14 A kind of method and system of express network data transmission Active CN102739508B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201110093822.5A CN102739508B (en) 2011-04-14 2011-04-14 A kind of method and system of express network data transmission

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201110093822.5A CN102739508B (en) 2011-04-14 2011-04-14 A kind of method and system of express network data transmission

Publications (2)

Publication Number Publication Date
CN102739508A CN102739508A (en) 2012-10-17
CN102739508B true CN102739508B (en) 2016-09-28

Family

ID=46994326

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201110093822.5A Active CN102739508B (en) 2011-04-14 2011-04-14 A kind of method and system of express network data transmission

Country Status (1)

Country Link
CN (1) CN102739508B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103475600B (en) * 2013-09-12 2016-05-11 北京邮电大学 A kind of multipath fairness transfer control method
CN104135401B (en) * 2014-07-14 2018-05-01 京信通信系统(中国)有限公司 The time delay detecting method and device of a kind of end to end network
CN105024946A (en) * 2015-06-16 2015-11-04 清华大学 A method, device and system for bandwidth allocation based on sliding window
CN111601177B (en) * 2020-04-13 2021-03-23 北京创享苑科技文化有限公司 Online continuous inspection method based on data packet round-trip delay distribution

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101056259A (en) * 2007-05-21 2007-10-17 中南大学 A collaborative congestion control method for high-speed network
CN101075960A (en) * 2007-06-22 2007-11-21 清华大学 Method for controlling fluid based on router virtual display load feedback in high-speed network
CN101969371A (en) * 2010-10-19 2011-02-09 高杰 User data packet protocol-based method for realizing high-speed reliable transmission of network data
CN101977151A (en) * 2010-11-09 2011-02-16 中南大学 Congestion control method used for wide area network environment with high packet loss rate
CN102006230A (en) * 2010-11-26 2011-04-06 中南大学 Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7974195B2 (en) * 2003-06-12 2011-07-05 California Institute Of Technology Method and apparatus for network congestion control
JP4793652B2 (en) * 2004-10-21 2011-10-12 日本電気株式会社 Communication quality measuring apparatus and measuring method thereof

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101056259A (en) * 2007-05-21 2007-10-17 中南大学 A collaborative congestion control method for high-speed network
CN101075960A (en) * 2007-06-22 2007-11-21 清华大学 Method for controlling fluid based on router virtual display load feedback in high-speed network
CN101969371A (en) * 2010-10-19 2011-02-09 高杰 User data packet protocol-based method for realizing high-speed reliable transmission of network data
CN101977151A (en) * 2010-11-09 2011-02-16 中南大学 Congestion control method used for wide area network environment with high packet loss rate
CN102006230A (en) * 2010-11-26 2011-04-06 中南大学 Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
《一种基于HSTCP改进的公平性算法》;刘文远等;《小型微型计算机系统》;20080131;第29卷(第1期);第31页到第34页 *

Also Published As

Publication number Publication date
CN102739508A (en) 2012-10-17

Similar Documents

Publication Publication Date Title
CN103999409B (en) Method for Link Buffer Size and Queue Length Estimation for Bandwidth Varying Mobile Data Networks
CN104272680B (en) signaling congestion
CN105024940B (en) Heterogeneous network TCP jamming control methods based on link circuit self-adapting
CN110730469B (en) Method for predicting bandwidth based on extended Kalman wireless network and congestion control thereof
CN112104562A (en) Congestion control method and device, communication network and computer storage medium
Poojary et al. Analytical model for congestion control and throughput with TCP CUBIC connections
CN102739508B (en) A kind of method and system of express network data transmission
CN103428104B (en) A kind of jamming control method based on content center network
CN101969432B (en) Random backoff based control method of TCP (Transmission Control Protocol) congestion window
Barrera et al. Statistical approach for congestion control in gateway routers
Jiang et al. CLTCP: an adaptive TCP congestion control algorithm based on congestion level
Weigle et al. Performance of competing high-speed TCP flows
Chen et al. Fluid-flow Analysis of TCP Westwood with RED
Xu et al. Adaptive congestion control in infrastructure wireless LANs with bounded medium access delay
Ameur et al. Evaluation of gateway-based shaping methods for HTTP adaptive streaming
Ogura et al. A new tcp congestion control supporting rtt-fairness
Su et al. DVPTCP: A delay-driven virtual parallel TCP for high-speed and lossy networks
Estevez et al. A Carrier-Ethernet oriented transport protocol with a novel congestion control and QoS integration: Analytical, simulated and experimental validation
Xu et al. Stability and fairness of rate estimation-based AIAD congestion control in TCP
Geng et al. Delay-constrained rate control for real-time video streaming over wireless networks
CN102624614A (en) Stable congestion control method of high-speed network
CN104079498B (en) Satellite network active queue management method based on packet size
Manikandan et al. Active queue management based congestion control protocol for wireless networks
Keceli et al. Fair and efficient transmission control protocol access in the IEEE 802.11 infrastructure basic service set
Zhang et al. Adaptive fast TCP

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant