WO2017092609A1 - 具有分等级配置的加速策略的网络传输加速方法和系统 - Google Patents

具有分等级配置的加速策略的网络传输加速方法和系统 Download PDF

Info

Publication number
WO2017092609A1
WO2017092609A1 PCT/CN2016/107196 CN2016107196W WO2017092609A1 WO 2017092609 A1 WO2017092609 A1 WO 2017092609A1 CN 2016107196 W CN2016107196 W CN 2016107196W WO 2017092609 A1 WO2017092609 A1 WO 2017092609A1
Authority
WO
WIPO (PCT)
Prior art keywords
acceleration
kernel
policy
parameters
request
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.)
Ceased
Application number
PCT/CN2016/107196
Other languages
English (en)
French (fr)
Inventor
洪珂
陈雄霖
黄莎琳
陈文生
杨鹏程
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.)
Wangsu Science and Technology Co Ltd
Original Assignee
Wangsu Science and Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wangsu Science and Technology Co Ltd filed Critical Wangsu Science and Technology Co Ltd
Publication of WO2017092609A1 publication Critical patent/WO2017092609A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
    • H04L69/163—In-band adaptation of TCP data exchange; In-band control procedures
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/40—Network security protocols

Definitions

  • the present invention relates to network transmission acceleration technologies, and in particular, to a network transmission acceleration method and system with an accelerated strategy of hierarchical configuration.
  • TCP acceleration refers to a higher performance (compared to standard TCP) for transmitting data over the Internet (Internet) through a set of optimization techniques without modifying specific applications.
  • mainstream TCP acceleration technologies mainly include bilateral TCP optimization and unilateral TCP optimization.
  • TCP unilateral acceleration is widely used because it is easy to deploy, deploys software or devices at one end of a TCP connection, and is fully compatible with standard TCP protocols.
  • the main implementation of the existing TCP unilateral acceleration is to improve the old congestion control algorithm, or design a new congestion control algorithm to perform congestion judgment and processing, thereby effectively utilizing bandwidth and optimizing the transmission rate.
  • TCP Vegas TCP Vegas
  • CUBIC FastTCP
  • Zeta-TCP Zeta-TCP
  • the existing TCP unilateral acceleration uses the same TCP unilateral acceleration scheme for all connections on the server, and cannot be hierarchically configured as needed.
  • CDN Content Delivery Network
  • the server can only adopt the one-way TCP unilateral acceleration scheme, the aggressive TCP unilateral acceleration scheme will bring about an increase in the retransmission ratio, waste of bandwidth, and the TCP unilateral acceleration scheme with lower retransmission ratio cannot be realized.
  • the optimal acceleration effect so some customers do not need special quality acceleration effect considering the bandwidth cost, while some customers are more concerned about the acceleration effect and would sacrifice the bandwidth.
  • the existing acceleration configuration has no The law achieves a balance between the two;
  • the present invention provides a network transmission acceleration method and system with a hierarchically configured acceleration policy, so that the same server can be configured according to different service types and different customer needs. Different acceleration strategies to achieve flexible application of acceleration strategies.
  • the present invention provides a network transmission acceleration method with an accelerated policy of hierarchical configuration, the method comprising: receiving a request from a client; parsing the request to determine a service type of the request; determining that a preset is multiple Whether there is a specific acceleration policy configured for the service type in the kernel of the acceleration policy, wherein the multiple acceleration policies correspond to different kernel acceleration parameters; when it is determined that the core exists for the service type When the specific acceleration policy is configured, the kernel is instructed to use the kernel acceleration parameter corresponding to the specific acceleration policy to issue a packet.
  • each of the plurality of acceleration strategies corresponds to a policy identification.
  • instructing the kernel to apply a kernel acceleration parameter corresponding to the specific acceleration policy to issue a packet includes invoking a policy identifier corresponding to the specific acceleration policy, and responsive to the specific acceleration strategy Corresponding policy identifier is invoked, and an acceleration policy setting interface provided by the kernel transmits a kernel acceleration parameter corresponding to the policy identifier according to the policy identifier, so that the kernel recognizes that the kernel acceleration parameter should be adopted Make a package.
  • the method further includes instructing the kernel to directly issue a kernel acceleration parameter corresponding to a default acceleration policy when there is no specific acceleration policy configured for the service type.
  • the kernel acceleration parameters are not fixed.
  • the type of service is determined based at least on a setting of transmission quality and retransmission ratio.
  • the method further comprises, after the data packet to be transmitted is transmitted, restoring the kernel acceleration parameters employed to perform the packet to a kernel acceleration parameter corresponding to the default acceleration policy.
  • the invention also provides a network transmission acceleration system with a hierarchical configuration acceleration strategy, the system comprising:
  • Means for parsing the request to determine the type of service of the request
  • the present invention has the following advantages: on the same server, different acceleration policies can be configured according to different service types and different customer needs, thereby implementing flexible application of the acceleration policy.
  • FIG. 1 shows a schematic diagram of a core pre-provisioned with a plurality of hierarchically configured acceleration strategies, in accordance with an embodiment of the present invention
  • FIG. 2 shows a flow chart of a network transmission acceleration method with an accelerated policy of hierarchical configuration, in accordance with an embodiment of the present invention
  • FIG. 3 illustrates a specific interaction between a kernel and an application layer of the method shown in FIG. 2 according to an embodiment of the invention. Process
  • FIG. 4 is a detailed process flow diagram of the core of the method shown in FIG. 2, in accordance with an embodiment of the present invention.
  • the kernel is a Linux kernel. But the kernel can also be a variety of Windows kernels and so on.
  • a plurality of different acceleration strategies are preset. This plurality of different acceleration strategies corresponds to different kernel acceleration schemes (eg, different TCP unilateral acceleration schemes) that require different data transmission acceleration effects and different retransmission ratios to meet various business requirements. In other words, this multiple different acceleration strategies correspond to different kernel acceleration parameters.
  • a download type of service needs to complete data transmission quickly, a streaming type business needs to smoothly transmit data, and a small file type business needs to balance speed and retransmission ratio.
  • video playback requires smooth transmission, large file downloads need to be completed faster, and small file downloads focus on the number of data packets transmitted by the first RTT.
  • Small files generally refer to web page class files, and their file sizes are generally below 100 KB.
  • Large files are generally image and resource files, and their single file size is generally larger than 1MB.
  • the plurality of different acceleration policies may be specified to be applicable to different service types, respectively, wherein the service type is determined based at least on the settings of the transmission quality and the retransmission ratio. As shown in FIG.
  • Kernel acceleration parameters corresponding to each acceleration strategy can be configured through an acceleration policy setting interface provided by the kernel, and each acceleration policy corresponds to a policy identifier.
  • the application layer invokes the corresponding policy identifier to invoke the corresponding acceleration policy to send the packet.
  • the policy identifier exists in the kernel.
  • the acceleration policy setting interface passes parameters corresponding to the policy identifier according to the policy identifier, so that the kernel recognizes that the parameters correspond to the policy identifier, and then the kernel These parameters will be used for the next step, such as issuing a package.
  • the kernel acceleration parameters of each acceleration policy may include, for example, a congestion control algorithm that needs to be adopted, and The window value and the change value, as well as the startup parameters of some TCP connection transmissions, etc., may be different for different service types and may vary depending on actual usage.
  • the kernel issues the package with the default policy.
  • the customizable kernel acceleration parameters may include, but are not limited to, parameters defining an initial window of the TCP connection, maximum and minimum window values, parameters that optimize the network packet loss, and may be optimized.
  • the way to set the kernel acceleration parameters is implemented by adding a file assignment interface to the kernel.
  • a directory can be added, and a subdirectory for kernel acceleration parameters corresponding to each acceleration policy (such as a large file acceleration policy, a small file acceleration policy, a video acceleration policy, etc.) is included in the directory.
  • Each acceleration strategy subdirectory may further include a series of kernel parameter setting interface files.
  • the file may be assigned to set the congestion control algorithm in the acceleration policy corresponding to the large file to BIC (binary increase congestion control, Binary Increase). Congestion control algorithm.
  • FIG. 2 shows a flow chart of a network transmission acceleration method with a hierarchical configuration acceleration policy, in accordance with an embodiment of the present invention.
  • the server receives a request initiated by the client.
  • the server can be a CDN server, for example. But the server can also be other servers.
  • Clients can include desktop computers, laptop computers, netbooks, mobile phones, and tablets, and the like.
  • the kernel layer detects that the local device has an application service that opens the response client request port, and establishes a connection with the client after the detection is passed.
  • a TCP connection can be established by a three-way handshake with the client for a TCP connection.
  • the client initiates the request for content, and the kernel layer passes the request to the application layer.
  • the application layer receives the request from the client and parses the request to determine the service type of the request.
  • the service type of the request In the configuration of the application layer, there is a setting of the service type of the request or a domain name transmission quality and retransmission ratio. In the case of the corresponding settings, the corresponding acceleration policy is matched for the request.
  • step 205 after matching the level policy, the application layer will instruct the kernel layer to transmit the instruction of the packet to the kernel layer by using the matched level policy, to instruct the kernel layer to adopt and match the connection.
  • the kernel acceleration parameters corresponding to the level policy are transmitted;
  • the kernel layer receives the instruction and sets the kernel acceleration parameters for the connection.
  • FIG. 3 illustrates a specific interaction flow between a kernel and an application layer of the method illustrated in FIG. 2 according to an embodiment of the invention.
  • step 301 the user requests to reach the kernel, and in step 302 the kernel forwards the request to the application layer for processing.
  • the application layer parses the request to determine the type of service corresponding to the request, wherein the type of service is determined based at least on the settings of the transmission quality and the retransmission ratio.
  • the service type can be small file download, large file download, streaming media play, and the like.
  • the application layer determines whether there is a specific acceleration policy configured for the service type among the plurality of acceleration policies preset in the kernel.
  • the application layer instruction kernel applies a kernel acceleration parameter corresponding to the specific acceleration policy for subsequent data packets. It includes the policy identifier in the application layer calling kernel corresponding to the specific acceleration policy.
  • the acceleration policy setting interface provided by the kernel transmits a parameter corresponding to the policy identifier according to the policy identifier, so that the kernel recognizes that the parameter corresponds to the policy identifier, and then the kernel adopts These parameters are used for the next step, such as sending a package. After the setting is completed, the data packet is sent normally.
  • the data packet is directly normal.
  • the instruction core directly issues the kernel acceleration parameters corresponding to the default acceleration policy for subsequent packets.
  • FIG. 4 illustrates a specific process flow of the core of the method illustrated in FIG. 2, in accordance with an embodiment of the present invention.
  • the kernel layer receives the request submitted by the application layer.
  • the kernel determines whether the connection carrying the request has set an acceleration policy corresponding to the requested service type.
  • step 403 the kernel reads the particular acceleration policy from its preset acceleration policies, and then at step 404 sets the specific acceleration with the level instruction based on the level instructions.
  • Policy-associated kernel acceleration parameters where kernel acceleration parameters The number can include the congestion control algorithm to be used, the congestion window values and change values for each phase, and some startup parameters for connection transmission.
  • the kernel sends the connected packet according to the kernel acceleration parameter until all packets are sent.
  • the kernel restores the kernel acceleration parameters of the connection to the kernel acceleration parameters corresponding to the default acceleration policy.
  • the kernel reads the default acceleration policy.
  • kernel acceleration parameters such as various acceleration parameters, congestion control algorithms, and the like associated with the default acceleration policy are set according to a default acceleration policy.
  • the connected data packet is sent in accordance with the kernel acceleration parameters associated with the default acceleration policy until all data packets have been transmitted.
  • step 410 after all data packets have been transmitted, the transmission ends.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了具有分等级配置的加速策略的网络传输加速方法和系统。该方法包括：接收来自客户端的请求；解析所述请求以确定所述请求的业务类型；判断预置有多个加速策略的内核中是否存在针对所述业务类型来进行配置的特定加速策略，其中所述多个加速策略与不同的内核加速参数相对应；当确定所述内核中存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包。

Description

具有分等级配置的加速策略的网络传输加速方法和系统 技术领域
本发明涉及网络传输加速技术，尤其涉及一种具有分等级配置的加速策略的网络传输加速方法和系统。
背景技术
目前，存在各种网络传输加速技术，例如TCP(传输控制协议，Transmission Control Protocol)加速技术等。其中TCP加速指的是通过一组优化技术，使TCP在通过因特网(Internet)传输数据的性能更高(与标准TCP相比)，而无需修改具体的应用。目前主流的TCP加速技术主要包括双边TCP优化和单边TCP优化两种。尤其地，TCP单边加速由于其易于部署，只需在TCP连接的一端部署软件或设备，且和标准的TCP协议完全兼容，因而受到广泛的应用。现有的TCP单边加速的主要实现方式是改良旧的拥塞控制算法，或者设计新的拥塞控制算法，以进行拥塞的判断与处理，从而有效利用带宽，优化传输速率。拥塞控制算法的示例包括TCP Vegas、CUBIC、FastTCP、Zeta-TCP等。但是由于TCP单边加速技术判断网络拥塞的反馈信息有限，因此为了保证较高的传输速度，通常需要频繁地发送重传包，从而造成了重传比的上升和带宽浪费。而且，现有的TCP单边加速对服务器上的所有连接均采用相同的TCP单边加速方案，而无法根据需要进行分等级配置。但是，对于诸如CDN(内容分发网络，Content Delivery Network)服务器之类的服务器而言，在同一服务器上，可存在多个不同的域名以及多种不同的业务类型。因此，当将现有的TCP单边加速应用于此类服务器时,可存在以下缺点：
首先，由于服务器只能采用一中TCP单边加速方案，而激进的TCP单边加速方案会带来重传比的上升，带宽的浪费，较低重传比的TCP单边加速方案又无法实现最优的加速效果，因此对于有些客户考虑带宽成本不需要特别优质加速效果，而有些客户则更在意加速效果而宁可牺牲带宽，现有的加速配置无 法实现二者的平衡；
其次，服务器上存在具有不同需求的多种不同的业务类型，例如视频播放需要平稳传输、大文件下载需要较快完成传输、小文件下载则重点关注第一个RTT所传数据包个数等，而现有的单边加速技术在同一个服务器上只能设置一种策略，因此无法满足多种业务类型的需求。
虽然以上仅指出了现有的TCP单边加速方案中存在的缺陷，但实际上现有的其他网络传输加速技术也可存在类似的缺陷，因此需要提供一种能够克服以上缺陷的网络传输加速技术。
发明内容
为了克服现有的网络传输加速技术中的缺陷，本发明提供一种具有分等级配置的加速策略的网络传输加速方法和系统，使得在同一服务器上，可根据不同业务类型、不同客户需要来配置不同的加速策略，从而实现加速策略的灵活应用。
具体地，本发明提供一种具有分等级配置的加速策略的网络传输加速方法，该方法包括：接收来自客户端的请求；解析所述请求以确定所述请求的业务类型；判断预置有多个加速策略的内核中是否存在针对所述业务类型来进行配置的特定加速策略，其中所述多个加速策略与不同的内核加速参数相对应；当确定所述内核中存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包。
在一个实施例中，所述多个加速策略中每一个均对应于一策略标识。
在一个实施例中，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包包括调用与所述特定加速策略相对应的策略标识，且响应于与所述特定加速策略相对应的策略标识被调用，所述内核所提供的加速策略设置接口将依据所述策略标识而传递与所述策略标识相对应的内核加速参数，使得所述内核认识到应采用这些内核加速参数来进行发包。
在一个实施例中，所述方法还包括当不存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核直接采用与默认加速策略相对应的内核加速参数来发包。
在一个实施例中，所述内核加速参数并不是固定的。
在一个实施例中，所述业务类型至少基于传输质量和重传比的设置来确定。
在一个实施例中，所述方法还包括当要发送的数据包被发送完毕后，将被采用来进行发包的内核加速参数还原成与默认加速策略相对应的内核加速参数。
本发明还提供了一种具有分等级配置加速策略的网络传输加速系统，所述系统包括：
用于接收来自客户端的请求的装置；
用于解析所述请求以确定所述请求的业务类型的装置；
用于判断预置有多个加速策略的内核中是否存在针对所述业务类型来进行配置的特定加速策略的装置，其中所述多个加速策略与不同的内核加速参数相对应；
用于当确定所述内核中存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包的装置。
与现有技术相比，本发明具有以下优点：在同一服务器上，可根据不同业务类型、不同客户需要来配置不同的加速策略，从而实现加速策略的灵活应用。
附图说明
本发明的以上发明内容以及下面的具体实施方式在结合附图阅读时会得到更好的理解。需要说明的是，附图仅作为所请求保护的发明的示例。在附图中，相同的附图标记代表相同或类似的元素。
图1示出根据本发明的一实施例的预设有多个分等级配置的加速策略的内核的示意图；
图2示出根据本发明的一实施例的一种具有分等级配置的加速策略的网络传输加速方法的流程图；
图3根据本发明的一实施例示出的图2所示的方法的内核与应用层的具体交互 流程；
图4根据本发明的一实施例示出的图2中所示的方法的内核的具体处理流程。
具体实施方式
下面结合具体实施例和附图对本发明作进一步说明，但不应以此限制本发明的保护范围。
图1是根据本发明的一实施例示出预设有多个分等级配置的加速策略的内核的示意图。在一个非限制实施例中，该内核为Linux内核。但该内核也可以是各种Windows内核等等。在根据本发明的内核中，除了设置有常规内核中通常设有的默认加速策略外，还预置有多个不同的加速策略。此多个不同的加速策略与要求不同的数据传输加速效果以及不同的重传比的不同内核加速方案(例如，不同的TCP单边加速方案)相对应，以满足各种不同的业务需求。换言之，此多个不同的加速策略与不同的内核加速参数相对应。例如，下载类型的业务需要快速的完成数据传输，流媒体类型的业务需要平稳地传输数据，小文件类型的业务则需要兼顾速度和重传比等等。具体地，例如视频播放需要平稳传输、大文件下载需要较快完成传输、小文件下载则重点关注第一个RTT所传数据包个数等。其中小文件一般指网页页面类文件，其文件大小一般在100KB以下。大文件一般为图片、资源类文件，其单个文件大小一般都大于1MB。在一个实施例中，可指定该多个不同的加速策略分别适用于不同的业务类型，其中业务类型至少基于传输质量和重传比的设置来确定。如图1所示，作为非限制示例，可指定策略1适用于小文件下载，策略2适用于大文件下载，策略3适用于流媒体播放，策略4适用于移动网络数据传输等等。与各加速策略相对应的内核加速参数可通过内核所提供的加速策略设置接口来进行配置，并且每个加速策略对应于一个策略标识。应用层调用对应的策略标识即可调用对应的加速策略进行发包。策略标识存在于内核中，当应用层调用特定策略标识时，加速策略设置接口会依据该策略标识而传递与该策略标识对应的参数，使得内核认识到这些参数与该策略标识相对应，然后内核会采用这些参数来进行下一步的处理，例如发包等。例如，在针对TCP单边加速的情况下，可以进行配置的每个加速策略的内核加速参数可包括例如需要采用的拥塞控制算法、各个阶段的拥 塞窗口值和变化值以及一些TCP连接传输的启动参数等等，但其对于不同的业务类型可以是不同的，并且可以根据实际使用情况而变化。并且在应用层没有请求特定加速策略的时候，内核以默认策略发包。下面以大文件为例，对可配置的内核加速参数作出进一步的说明。例如，在大文件的情况下，可自定义的内核加速参数可包括但不限于：定义TCP连接的初始窗口、最大、最小窗口值等的参数，对网络丢包有优化作用的参数、可以优化快速恢复的传输的参数、可以控制数据包进行平稳传输的参数、可以利用时间序的判断进行数据包重传的参数等等。
另外，内核加速参数的设置方式是采用在内核中新增文件赋值接口的方式来实现的。比如可新增一目录，并在该目录下包含针对与各加速策略(例如大文件加速策略、小文件加速策略、视频加速策略等)相对应的内核加速参数的子目录。各个加速策略子目录可进一步包含一系列内核参数设置接口文件，在一个示例中对该文件进行赋值则可将对应大文件的加速策略中的拥塞控制算法设置为BIC(二进制增加拥塞控制，Binary Increase Congestion control)算法。
图2示出根据本发明的一实施例的一种具有分等级配置加速策略的网络传输加速方法的流程图。
在步骤201，服务器接收到由客户端发起的请求。作为一个示例，该服务器可以是诸如CDN服务器。但该服务器也可以是其他服务器。客户端可以包括台式计算机、膝上型计算机、上网本、移动电话和平板计算机等等。
在步骤202，内核层检测到本机有开启响应客户端请求端口的应用服务，并在检测通过后与客户端建立连接。例如，如本领域技术人员所公知的，在该请求为TCP请求的情况下，可通过与客户度进行TCP连接的三次握手来建立TCP连接。
在步骤203，客户端发起请求内容，内核层将该请求传递给应用层。
在步骤204，应用层接收到来自客户端的请求并对解析该请求以确定该请求的业务类型，在该应用层的配置中有与该请求的业务类型的设置或域名传输质量与重传比相对应的设置的情况下，为该请求匹配相应的加速策略。
在步骤205，在匹配了等级策略后，应用层将指示内核层采用所匹配的等级策略进行发包的指令传输给内核层，以指令内核层对该连接采用与所匹配的 等级策略相对应的内核加速参数来进行传输；
在步骤206，内核层接收到该指令，设置针对该连接的内核加速参数。
图3根据本发明的一实施例示出图2所示的方法的内核与应用层的具体交互流程。
在步骤301用户请求到达内核，并在步骤302内核将该请求转交给应用层处理。
在步骤303，应用层解析该请求，以确定与该请求相对应的业务类型，其中业务类型至少基于传输质量和重传比的设置来确定。例如，该业务类型可以为小文件下载、大文件下载、流媒体播放等等。
在步骤304，应用层判断在内核中预置的多个加速策略中是否存在针对该业务类型来进行配置的特定加速策略。
在步骤305，在存在针对该业务类型来进行配置的特定加速策略的情况下，该应用层指令内核针对后续的数据包采用与该特定加速策略相对应的内核加速参数来发包。其包括应用层调用内核中与该特定加速策略相对应的策略标识。而响应于应用层调用该策略标识，内核所提供的加速策略设置接口将依据该策略标识而传递与该策略标识对应的参数，使得内核认识到这些参数与该策略标识相对应，然后内核会采用这些参数来进行下一步的处理，例如发包等。设置完毕后，正常发送数据包。
在步骤306，在不存在针对该业务类型来进行配置的加速策略的情况下，直接正常的响应数据包。例如，指令内核针对后续的数据包直接采用与默认加速策略相对应的内核加速参数来发包。
图4根据本发明的一实施例示出图2所示的方法的内核的具体处理流程。
在步骤401，内核层接收到由应用层提交的请求。
在步骤402，内核判断承载该请求的连接是否已设置与该请求的业务类型相对应的加速策略。
如果承载该数据的连接已经经由应用层通知内核采用特定加速策略，则在步骤403内核从其预置的各加速策略中读取该特定加速策略，随后在步骤404根据等级指令设置与该特定加速策略相关联的内核加速参数，其中内核加速参 数可包括要采用的拥塞控制算法、各阶段的拥塞窗口值和变化值以及一些连接传输的启动参数等。
在步骤405，内核按该内核加速参数发送该连接的数据包，直到所有数据包均被发送完毕。
在步骤406，当所有数据包均发送完毕后，内核将该连接的内核加速参数还原成与默认加速策略相对应的内核加速参数。
在步骤407，如果该连接没有设置特定加速策略，则内核读取默认加速策略。
在步骤408，根据默认加速策略来设置与该默认加速策略相关联的各项加速参数、拥塞控制算法等内核加速参数。
在步骤409，按与该默认加速策略相关联的内核加速参数发送连接的数据包，直到所有数据包被发送完毕。
在步骤410，在所有数据包均被发送完毕后，传输结束。
虽然本发明以TCP单边加速的情况为示例作出了进一步的说明，但是本发明也同样适用于其他网络传输加速技术，诸如UDP加速技术等，而不会背离发明所要求保护的精神和范围。

Claims (10)

  1. 一种具有分等级配置的加速策略的网络传输加速方法，所述方法包括：
    接收来自客户端的请求；
    解析所述请求以确定所述请求的业务类型；
    判断预置有多个加速策略的内核中是否存在针对所述业务类型来进行配置的特定加速策略，其中所述多个加速策略与不同的内核加速参数相对应；
    当确定所述内核中存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包。
  2. 如权利要求1所述的方法，其特征在于，所述多个加速策略中每一个均对应于一策略标识。
  3. 如权利要求1所述的方法，其特征在于，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包包括调用与所述特定加速策略相对应的策略标识，且响应于与所述特定加速策略相对应的策略标识被调用，所述内核所提供的加速策略设置接口将依据所述策略标识而传递与所述策略标识相对应的内核加速参数，使得所述内核认识到应采用这些内核加速参数来进行发包。
  4. 如权利要求1所述的方法，其特征在于，所述方法还包括当不存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核直接采用与默认加速策略相对应的内核加速参数来发包。
  5. 如权利要求1所述的方法，其特征在于，所述内核加速参数并不是固定的。
  6. 如权利要求1所述的方法，其特征在于，所述业务类型至少基于传输质量和重传比的设置来确定。
  7. 如权利要求1所述的方法，其特征在于，所述方法还包括当要发送的数据包被发送完毕后，将被采用来进行发包的内核加速参数还原成与默认加速策略相对应的内核加速参数。
  8. 一种具有分等级配置的加速策略的网络传输加速系统，所述系统包括：
    用于接收来自客户端的请求的装置；
    用于解析所述请求以确定所述请求的业务类型的装置；
    用于判断预置有多个加速策略的内核中是否存在针对所述业务类型来进行配置的特定加速策略的装置，其中所述多个加速策略与不同的内核加速参数相对应；
    用于当确定所述内核中存在针对所述业务类型来进行配置的特定加速策略时，指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包的装置。
  9. 如权利要求8所述的系统，其特征在于，所述多个加速策略中每一个均对应于一策略标识。
  10. 如权利要求9所述的系统，其特征在于，用于指令所述内核采用与所述特定加速策略相对应的内核加速参数来进行发包的装置包括用于调用与所述特定加速策略相对应的策略标识的装置，且响应于与所述特定加速策略相对应的策略标识被调用，所述内核所提供的加速策略设置接口将依据所述策略标识而传递与所述策略标识相对应的内核加速参数，使得所述内核认识到应采用这些内核加速参数来进行发包。
PCT/CN2016/107196 2015-12-03 2016-11-25 具有分等级配置的加速策略的网络传输加速方法和系统 Ceased WO2017092609A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510881183.7 2015-12-03
CN201510881183.7A CN105516122B (zh) 2015-12-03 2015-12-03 具有分等级配置的加速策略的网络传输加速方法和系统

Publications (1)

Publication Number Publication Date
WO2017092609A1 true WO2017092609A1 (zh) 2017-06-08

Family

ID=55723759

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/107196 Ceased WO2017092609A1 (zh) 2015-12-03 2016-11-25 具有分等级配置的加速策略的网络传输加速方法和系统

Country Status (2)

Country Link
CN (1) CN105516122B (zh)
WO (1) WO2017092609A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113783740A (zh) * 2021-10-26 2021-12-10 北京字节跳动网络技术有限公司 一种网络优化方法、装置、设备及介质

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105516122B (zh) * 2015-12-03 2019-03-15 网宿科技股份有限公司 具有分等级配置的加速策略的网络传输加速方法和系统
CN107172179B (zh) * 2017-06-05 2020-02-18 网宿科技股份有限公司 一种双边加速传输方法和系统
CN109495388B (zh) * 2018-11-30 2021-04-27 四川安迪科技实业有限公司 基于卫星通信的tcp加速方法
CN111935233B (zh) * 2020-07-13 2022-03-29 杭州鸿雁电器有限公司 路由器的加速方法、装置、存储介质和处理器
CN112822051B (zh) * 2021-01-06 2022-09-16 贵阳迅游网络科技有限公司 基于业务感知的业务加速方法
CN113014509B (zh) * 2021-05-26 2021-09-17 腾讯科技（深圳）有限公司 应用程序加速方法和装置
CN114338819A (zh) * 2021-12-30 2022-04-12 北京天融信网络安全技术有限公司 单边加速方法、装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006062328A1 (en) * 2004-12-08 2006-06-15 Electronics And Telecommunications Research Institute Retransmission and delayed ack timer management logic for tcp protocol
CN101729354A (zh) * 2008-10-18 2010-06-09 飞塔公司 使用隧道加速数据通信
CN102299899A (zh) * 2010-06-24 2011-12-28 清华大学 一种恶劣信道下的tcp加速方法
US20140250237A1 (en) * 2013-03-04 2014-09-04 Gilat Satellite Networks Ltd. Network performance enhancement
CN105516122A (zh) * 2015-12-03 2016-04-20 网宿科技股份有限公司 具有分等级配置的加速策略的网络传输加速方法和系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7783769B2 (en) * 2004-03-31 2010-08-24 Intel Corporation Accelerated TCP (Transport Control Protocol) stack processing
US9648147B2 (en) * 2006-12-29 2017-05-09 Futurewei Technologies, Inc. System and method for TCP high availability

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006062328A1 (en) * 2004-12-08 2006-06-15 Electronics And Telecommunications Research Institute Retransmission and delayed ack timer management logic for tcp protocol
CN101729354A (zh) * 2008-10-18 2010-06-09 飞塔公司 使用隧道加速数据通信
CN102299899A (zh) * 2010-06-24 2011-12-28 清华大学 一种恶劣信道下的tcp加速方法
US20140250237A1 (en) * 2013-03-04 2014-09-04 Gilat Satellite Networks Ltd. Network performance enhancement
CN105516122A (zh) * 2015-12-03 2016-04-20 网宿科技股份有限公司 具有分等级配置的加速策略的网络传输加速方法和系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113783740A (zh) * 2021-10-26 2021-12-10 北京字节跳动网络技术有限公司 一种网络优化方法、装置、设备及介质
CN113783740B (zh) * 2021-10-26 2022-08-16 北京字节跳动网络技术有限公司 一种网络优化方法、装置、设备及介质

Also Published As

Publication number Publication date
CN105516122B (zh) 2019-03-15
CN105516122A (zh) 2016-04-20

Similar Documents

Publication Publication Date Title
WO2017092609A1 (zh) 具有分等级配置的加速策略的网络传输加速方法和系统
CN105763474B (zh) 数据传输方法和装置
CN104067569B (zh) 组合无状态和有状态服务器负载平衡的方法
CN103988483B (zh) 用于网络加速器的动态适应的系统和方法
CN102986189B (zh) 用于为对应于虚通道的网络连接分配服务等级的系统和方法
CN103827828B (zh) 用于在流量指挥站环境中提供负载平衡与数据压缩的灵活性的系统及方法
US10291723B2 (en) Externally initiated application session endpoint migration
EP2543162B1 (en) Selectively disabling reliability mechanisms on a network connection
US12335158B2 (en) Modifying the congestion control algorithm applied to a connection based on request characteristics
US20160119237A1 (en) Wired data-connection aggregation
CN104009938A (zh) 基于路由层面的长连接的方法和系统
EP2208331A1 (en) Parallelizing the tcp behavior of a network connection
US9357014B2 (en) Service-based networking
CN111567011A (zh) 使用跨层信息提高视频服务和WEB服务的QoE的方法
CN104010001A (zh) 移动终端中同类联网请求进行连接通信的方法和系统
CN104717041A (zh) 数据传输方法和装置
US10880413B2 (en) Method and server for establishing a TCP connection
US10291717B2 (en) Prioritizing VDI sessions and redirected devices in software defined networks
CN107666465B (zh) 数据传输方法和装置
Welzl et al. Beneficial transparent deployment of SCTP: the missing pieces
CN104601351A (zh) 网络设备能力配置方法、网络设备及系统
KR101364927B1 (ko) 네트워크의 토렌트 트래픽 선별 차단 방법
CN118714128A (zh) 一种文件快速传输的方法
CN118678400A (zh) 用于网络接入的装置、方法、和非暂态计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16869919

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16869919

Country of ref document: EP

Kind code of ref document: A1