WO2019033891A1

WO2019033891A1 - Rdma-based network traffic determination method and device

Info

Publication number: WO2019033891A1
Application number: PCT/CN2018/096210
Authority: WO
Inventors: 张维
Original assignee: 华为技术有限公司
Priority date: 2017-08-14
Filing date: 2018-07-19
Publication date: 2019-02-21
Also published as: CN109391522B; CN109391522A

Abstract

A remote direct memory access (RDMA)-based network traffic determination method and device. a forwarding device, after receiving one of a plurality of data packets obtained by a network card of a first terminal device by dividing a data block, determines whether the data packet is a target data packet in the plurality of data packets, and if yes, copies the data packet, sends the copied data packet to an analysis server, and forwards the data packet to a second terminal device; otherwise, the forwarding device does not copy the data packet and forwards the data packet to the second terminal device. The described method may reduce the storage capacity occupied by the analysis server and may save resources for transmission between the forwarding device and the analysis server. The embodiments of the present invention are used to solve the problem in existing technology of the storage capacity of an analysis server being largely occupied when the analysis server analyzes RDMA-based network traffic between terminal devices and the problem of resources for transmission between the forwarding device and the analysis server being wasted.

Description

Method and device for determining network traffic based on RDMA

This application claims the priority of the Chinese patent application filed on August 14, 2017, the Chinese Patent Office, Application No. 201710693896.X, entitled "A RDMA-Based Network Traffic Determination Method and Apparatus", the entire contents of which are The citations are incorporated herein by reference.

Technical field

The present application relates to the field of communications technologies, and in particular, to a method and apparatus for determining network traffic based on remote direct memory access (RDMA).

Background technique

RDMA technology is a technique for quickly moving data from a terminal device's memory directly to the memory of another remote terminal through a network to eliminate the impact on the performance of the processor. The main process of the RDMA is: the first terminal device directly sends the data in its own memory to the forwarding device in the network through its own network card, and the forwarding device forwards the data to the second terminal device, the second terminal device The network card directly stores the received data into the memory of the second terminal device, so that the processors of the two terminal devices are not required to participate in the data access process to reduce the impact on the processor performance of the terminal device. In the above process, the forwarding device, when forwarding the received data to the second terminal device, also copies and sends the data to the analysis server, so that the analysis server determines the network according to the data. Traffic, so that the network administrator performs subsequent network management based on the network traffic determined by the analysis server.

Generally, the first terminal device divides the data block to be transmitted into a plurality of data packets for transmission. Therefore, the following method is generally used to determine network traffic: the forwarding device copies and receives the received multiple data packets to the analysis server; the analysis server determines, among the multiple data packets sent by the forwarding device, the data packets belonging to the same data block. All packets, and the sum of the sizes of all the packets determined as the network traffic corresponding to the same data block.

At present, more and more terminal devices transmit data based on RDMA, and the forwarding device copies and sends a large number of data packets to the analysis server, thereby causing a large occupation of the storage capacity of the analysis server, and wasting the forwarding device and the analysis server. Transmission resources between.

Summary of the invention

The present invention provides an RDMA-based network traffic determination method and apparatus, which are used to avoid the storage capacity of the analysis server caused by the analysis of the RDMA-based network traffic between the terminal devices in the prior art. The problem of transmission resources between the device and the analysis server is wasted.

In a first aspect, the present application provides an RDMA-based network traffic determination method.

Receiving, by the network device of the first terminal device, one of a plurality of data packets obtained by dividing a data block by the network card of the first terminal device, and determining whether the data packet is a target data packet in the multiple data packets, if the data packet is The target data packet of the plurality of data packets, the forwarding device copies the data packet, sends the copied data packet to the analysis server, and forwards the data packet to the second terminal device; The data packet is not the target data packet of the plurality of data packets, and the forwarding device does not perform the process of copying the data packet, and forwards the data packet to the second terminal device; wherein the data block is And the first terminal device uses the RDMA mode to transmit the data block to the second terminal device by using the forwarding device, where the data packet includes a length of the data packet and a sequence number of the data packet.

Through the above method, the forwarding device only needs to copy and send the target data packet in the plurality of data packets divided for one data block to the analysis server, and the forwarding device pairs the data packets other than the target data packet among the plurality of data packets. The process of copying and sending to the analysis server is not performed, so that the analysis server determines the network traffic corresponding to the data block according to the target data packet, thereby reducing the occupation of the storage capacity of the analysis server, and saving the forwarding device and the analysis server. Transfer resources between.

In one possible design, the length of the data packet is represented by a payload field in the data packet, and the serial number of the data packet passes through a packet sequence number (PSN) in the data packet. The field is reflected.

In one possible design, the forwarding device may be, but is not limited to, a switch or a router or the like.

In a possible design, the target data packet may be a tail packet of the plurality of data packets. In this scenario, the forwarding device determines whether the data packet is in the multiple data packets. The target data packet may be: the forwarding device identifies a value of an opcode in the data packet, and if the value of the opcode is a value of an opcode corresponding to the tail packet, the forwarding device determines The data packet is a tail packet; if the value of the opcode is not the value of the opcode corresponding to the tail packet, the forwarding device determines that the data packet is not a tail packet.

Through the above method, the forwarding device can accurately identify the tail packets in the plurality of data packets, so that the forwarding device copies the identified tail packets and sends the tail packets to the analysis server, thereby enabling the analysis server. The network traffic corresponding to the data block is determined by the tail packet.

In a possible design, the target data packet may be a first packet or a tail packet of the multiple data packets. In this scenario, the forwarding device determines whether the data packet is the multiple data. The target data packet in the packet may be: the forwarding device identifies the value of the opcode in the data packet, and if the value of the opcode is the value of the opcode corresponding to the first packet, the forwarding device determines the data. The packet is the first packet; if the value of the opcode is the value of the opcode corresponding to the tail packet, the forwarding device determines that the data packet is a tail packet.

The forwarding device can accurately identify the first packet or the tail packet of the plurality of data packets, so that the forwarding device copies the identified first packet or the tail packet and sends the identifier to the analysis server. And causing the analysis server to subsequently determine network traffic corresponding to the data block by using the first packet and the tail packet.

In a possible design, the forwarding device is configured with an access control list (ACL), where the ACL stores the correspondence between the value of the opcode and the first packet, and the correspondence between the value of the opcode and the tail packet.

Through the foregoing method, the forwarding device may determine whether the received data packet is the first packet or the tail packet by querying the correspondence between the value of the opcode in the ACL and the first packet, and the correspondence between the value of the opcode and the tail packet.

In a possible design, the RDMA mode can be divided into three types: send, write, and read, and the forwarding device can determine that the packet is the first packet (send_first) or the tail. a packet (send_last); the forwarding device may also determine that the data packet is a write first packet (write_first) or a write tail packet (write_last); the forwarding device may further determine that the data packet is a read first packet (read_first) or Read the tail package (read_last). Correspondingly, the analysis server can count the transmitted data block, the written data block or the network traffic corresponding to the read data block.

In a second aspect, the present application further provides another method for determining network traffic based on RDMA.

The analyzing server receives a tail packet in a plurality of data packets obtained by dividing the data packet by the forwarding device, where the tail packet includes a length of the tail packet and a sequence number of the tail packet; and then the analyzing Determining, by the server, the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet; wherein the data block is transmitted to the second terminal device by using the RDMA mode by using the forwarding device The data block of the terminal device.

Through the above method, the analysis server only needs to receive the tail packets in the plurality of data packets obtained by dividing each data block sent by the forwarding device, which can reduce the occupation of the storage capacity of the analysis server, and can save the forwarding device and the analysis server. Transmission resources between.

In a possible design, the analysis server may further receive a first packet of the plurality of data packets obtained by the forwarding device for the data block partitioning; wherein the first packet includes the The length of the first packet and the serial number of the first packet. In this scenario, the analyzing server determines the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, and specifically includes the following steps: the analyzing server according to the tail a sequence number of the packet and a sequence number of the first packet, after determining the number of data packets between the first packet and the tail packet, the analysis server is configured according to the length of the first packet, the first packet, and The number of data packets between the tail packets and the length of the tail packets determine the network traffic corresponding to the data blocks.

Through the above method, the analysis server may determine network traffic corresponding to the data block according to the first packet and the tail packet, and in the process, the forwarding device only needs to copy and send multiple data divided for the data block. The first packet and the tail packet in the packet are sent to the analysis server, and the data packets other than the first packet and the tail packet of the plurality of data packets are not copied and sent to the analysis server, thereby reducing the occupation of the storage capacity of the analysis server. And can save transmission resources between the forwarding device and the analysis server.

In a possible design, the analysis server subtracts the serial number of the first packet from the serial number of the tail packet and then subtracts 1 to obtain a determined between the first packet and the tail packet. The number of packets.

In a possible design, the analysis server determines, according to the length of the first packet, the determined number of data packets between the first packet and the tail packet, and the length of the tail packet, The network traffic of the data block can be based on the following formula:

L=Len2+(N+1)*Len1

Where L is the network traffic corresponding to the data block, Len1 is the length of the first packet, Len2 is the length of the tail packet, and N is the number of data packets between the first packet and the tail packet. .

In a possible design, after receiving the first packet of the plurality of data packets, the analysis server identifies and records the quintuple and the QP of the first packet; the analysis server is based on the recorded five-element a group and a QP, determining the received tail packet that is the same as the recorded five-tuple and QP, and using the tail packet as a tail packet in the plurality of data packets of the data block; The quintuple and QP of the first packet are used to indicate the quintuple and QP of the data block to which the plurality of data packets belong.

Through the above method, the analysis server can accurately identify the first packet and the tail packet of the data block, so that the subsequent analysis server determines the network traffic corresponding to the data block according to the first packet and the tail packet. .

In a possible design, the analyzing server determines the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, and specifically includes the following steps:

The analysis server identifies a quintuple and a queue pair QP in the tail package, and determines a quintuple and a QP of the tail package according to a correspondence between the saved quintuple and the value of the QP and the maximum transmission unit MTU. a value of the corresponding MTU; then, the analysis server determines a starting sequence number of the set of the plurality of data packets, and according to the sequence number of the tail package and the set start of the plurality of data packets a sequence number, the number of data packets before the tail packet is determined; finally, the analysis server determines, according to the number of data packets before the tail packet, the quintuple of the tail packet, and the value of the MTU corresponding to the QP And the length of the tail packet determines a network traffic corresponding to the data block.

In a possible design, the analyzing server determines a method for setting a starting sequence number of the plurality of data packets:

When the first terminal device carries the set start sequence number of the plurality of data packets in the sent data packet, the analysis server acquires the start sequence number by using the data packet sent by the forwarding device ;

Alternatively, the analysis server acquires a pre-stored starting sequence number set when the first terminal device divides a plurality of data packets for one data block.

Through the above method, the analysis server can accurately determine the starting sequence number of the set plurality of data packets, so that the analysis server subsequently determines the number of data packets before the tail package.

In a possible design, the analysis server subtracts the set starting sequence number from the serial number of the tail package to obtain the number of data packets before the tail package.

In a possible design, the analysis server determines the corresponding location according to the number of data packets before the tail packet, the quintuple of the tail packet, the value of the MTU corresponding to the QP, and the length of the tail packet. When describing the network traffic of a data block, the following formula can be used:

T=Len+K*MTU

Wherein, T is the network traffic corresponding to the data block, the MTU is the value of the quintuple of the tail packet and the MTU corresponding to the QP, Len is the length of the tail packet, and K is the data packet before the tail packet. The number.

In a third aspect, the embodiment of the present application further provides a forwarding device, where the forwarding device has a function of implementing forwarding device behavior in the foregoing method instance. The functions may be implemented by hardware or by corresponding software implemented by hardware. The hardware or software includes one or more modules corresponding to the functions described above.

In a possible design, the structure of the forwarding device includes a communication interface, a processor, and a memory, the communication interface is configured to receive and send data, and the processor is configured to support the forwarding device to perform the foregoing The corresponding function in the method in one aspect. The memory is coupled to the processor, which stores program instructions and data necessary for the forwarding device.

In a fourth aspect, an embodiment of the present application further provides an analysis server, where the analysis server has a function of analyzing the behavior of the server in the foregoing method instance. The functions may be implemented by hardware or by corresponding software implemented by hardware. The hardware or software includes one or more modules corresponding to the functions described above.

In a possible design, the configuration of the analysis server includes a communication interface, a processor, and a memory, the communication interface is configured to receive and send data, and the processor is configured to support the analysis server to perform the foregoing The corresponding function in the method in the two aspects. The memory is coupled to the processor, which stores program instructions and data necessary for the analysis server.

In a fifth aspect, the embodiment of the present application further provides a network traffic determining system, where the network traffic determining system includes multiple terminal devices, multiple forwarding devices, and an analysis server, and any one of the multiple forwarding devices For the forwarding device of the above third aspect, the analysis server is the analysis server of the fourth aspect described above.

In a sixth aspect, the embodiment of the present application further provides a computer storage medium, where the computer storage medium stores computer executable instructions, where the computer executable instructions are used to cause the computer to execute when called by the computer. Any of the above methods.

In a seventh aspect, the embodiment of the present application further provides a computer program product comprising instructions, when executed on a computer, causing the computer to perform any of the above methods.

In the technical solution of the embodiment of the present application, the forwarding device receives one of a plurality of data packets obtained by dividing the network card of the first terminal device for one data block, and determines whether the data packet is target data in the multiple data packets. a packet, if yes, copying the data packet, transmitting the copied data packet to an analysis server, and forwarding the data packet to the second terminal device; otherwise, the forwarding device does not perform processing for copying the data packet, Forwarding the data packet to the second terminal device only; wherein the data block is a data block that is transmitted by the forwarding device to the second terminal device by using the RDMA mode by the first terminal device. In the method, the forwarding device only needs to copy and send the target data packet in the plurality of data packets divided for the data block to the analysis server, and the data packets other than the target data packet of the plurality of data packets are not copied and sent to the data packet. The processing of the server is analyzed to enable the analysis server to determine network traffic corresponding to the data block according to the target data packet, which can reduce the occupation of the storage capacity of the analysis server, and can save transmission resources between the forwarding device and the analysis server.

DRAWINGS

Figure 1a is a schematic diagram of a data packet provided by the present application;

1 is a schematic structural diagram of a network traffic determining system provided by the present application;

2 is a schematic flowchart of a RDMA-based network traffic determining method provided by the present application;

3 is a schematic structural diagram of a forwarding device provided by the present application;

4 is a schematic structural diagram of an analysis server provided by the present application;

FIG. 5 is a structural diagram of a forwarding device provided by the present application;

FIG. 6 is a structural diagram of an analysis server provided by the present application.

Detailed ways

The present application will be further described in detail below with reference to the accompanying drawings.

The embodiment of the present application provides a method and an apparatus for determining network traffic based on RDMA, so as to avoid the storage capacity of the analysis server caused by the analysis of the RDMA-based network traffic between the terminal devices by the analysis server in the prior art. The problem that the transmission resources between the forwarding device and the analysis server are wasted. The method and the device of the present application are based on the same inventive concept. Since the principles of the method and the device for solving the problem are similar, the implementation of the device and the method can be referred to each other, and the repeated description is not repeated.

Hereinafter, some of the terms in the present application will be explained to be understood by those skilled in the art.

1), a forwarding device, a device that forwards data, such as a switch, a router, and the like.

2) The first terminal device and the second terminal device are devices that use RDMA to transmit data, for example, a handheld device with a wireless connection function, an in-vehicle device, a wearable device, a computing device, and a mobile station (mobile station) , MS) or other processing device or the like connected to the wireless modem, and a mobile terminal that communicates with one or more core networks via the access network.

3) The analysis server may be a big data analysis platform or a host for data analysis, and is configured to determine network traffic corresponding to the data block according to a tail packet of one data block.

4), the network card of the first terminal device and the network card of the second terminal device, when the RDMA communication is used between the first terminal device and the second terminal device, the communication between the data sender and the data receiver is performed. main body. Specifically, the two network cards communicate with each other through a queue pair (QP). The QP of each network card is divided into a receiving QP and a sending QP, which are used for data receiving and sending, respectively.

5) The data packet is obtained by dividing the data block to be transmitted by the terminal device, and can be divided into a first packet, an intermediate data packet, and a tail packet. 1a is a schematic diagram of a data packet, wherein a payload field included in the data packet indicates a length of the data packet; and an opcode in a Base Transport Header (BTH) of the data packet. The field identifies the type of the data packet (ie, the first packet, the tail packet, or the intermediate data packet), and the QP field identifies the QP of the network card of the terminal device that needs to receive the data packet, and the PSN identifies the sequence number of the data packet. The specific meanings of other fields in the data packet are not specifically mentioned in the present application, and therefore will not be described again.

It is known that the RDMA mode is divided into three types: send, write, and read. Therefore, the opcode can specifically identify whether the data packet is the first packet, the intermediate data packet, or the tail packet in the send mode, or the first packet and the middle in the write mode. The data packet or the tail packet is the first packet, the intermediate data packet or the tail packet in the read mode.

6), multiple, means two or more.

7) In the description of the present application, the terms "first", "second" and the like are used to distinguish the purpose of the description, and are not to be construed as indicating or implying relative importance, nor in the order of indication or suggestion.

In order to more clearly describe the technical solutions of the embodiments of the present application, the RDMA-based network traffic determining method and apparatus provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

FIG. 1 shows a possible network traffic determining system applicable to the RDMA-based network traffic determining method provided by the embodiment of the present application, where the network traffic determining system includes: multiple terminal devices, multiple forwarding devices, and an analysis server. The RDMA mode communication is implemented by the multiple forwarding devices between the multiple terminal devices.

The plurality of terminal devices, such as the terminal device A, the terminal device B, the terminal device C, and the terminal device D, as shown in FIG. 1, include a plurality of terminal device pairs that use RDMA communication, that is, a plurality of terminal devices that transmit data and a plurality of terminal devices that receive data, wherein any one of the terminal devices that transmit data (referred to as the first terminal device) is configured to divide the data block to be transmitted into a plurality of data packets, and use the plurality of data packets by the network card Sending a data packet to one of the plurality of forwarding devices (such as forwarding device A, forwarding device B, and forwarding device C shown in FIG. 1), so that the forwarding device will receive the multiple data packets. Forwarding to the terminal device receiving the data (referred to as the second terminal device); after the second terminal device receives the plurality of data packets forwarded by the forwarding device, the network card that is connected to the second terminal device will receive the Multiple packets are stored directly into their own memory. The data packet includes a length and a sequence number in each of the plurality of data packets, where the plurality of data packets includes a first packet, M intermediate data packets, and a tail packet, where M is greater than or equal to 0. An integer; the length of the first packet is the same as the length of the M intermediate data packets, and is a maximum transmission unit (maximum transmission unit, which can be transmitted in the transmission path of the first terminal device and the second terminal device, The value of the MTU), the length of the tail packet is less than or equal to the length of the first packet.

For example, based on the network traffic determining system shown in FIG. 1, it is assumed that the terminal device A and the terminal device B transmit data by using any one of the plurality of forwarding devices by using an RDMA method, and the terminal device Between the C and the terminal device D, the forwarding device uses any one of the plurality of forwarding devices to transmit data by using an RDMA method, that is, the terminal device A and the terminal device B are a terminal device pair that communicates by using RDMA mode. The terminal device C and the terminal device D are a pair of terminal devices that communicate by RDMA. Taking the RDMA method for transmitting data between the terminal device A and the terminal device B through the forwarding device A as an example, the process of using the RDMA mode communication between the terminal devices is described in detail: the terminal device A is to be treated The transmitted data block is divided into a plurality of data packets, and the plurality of data packets are sent to the forwarding device A by the network card, and the forwarding device A forwards the plurality of data packets to the terminal device B. The network card of the terminal device B directly stores the plurality of data packets into its own memory. In this way, data transmission is completed between the terminal device A and the terminal device B.

In the prior art, any one of the plurality of forwarding devices is configured to: after receiving a plurality of data packets sent by a network card of any one of the plurality of terminal devices, the data packet While forwarding to another terminal device, the plurality of data packets are also copied and sent to the analysis server, so that the analysis server determines network traffic corresponding to the data block to which the plurality of data packets belong.

In the present application, in order to reduce the occupation of the storage capacity of the analysis server and save the transmission resources between the forwarding device and the analysis server, the plurality of forwarding devices may not copy all the data packets, but only copy multiple data. The target packet in the package. Specifically, the forwarding device of the plurality of forwarding devices determines, according to each packet received from the first terminal device, whether the data packet is a target data packet in multiple data packets, and if so, Copying the data packet, transmitting the copied data packet to the analysis server, and forwarding the data packet to the second terminal device; otherwise, the process of copying the data packet is not performed, and only the data packet is forwarded to Second terminal device. The target data packet may be a tail packet in the multiple data packets, or a first packet and a tail packet in the multiple data packets.

The analysis server is configured to determine network traffic corresponding to the data block sent by any one of the terminal devices, and display the determined network traffic on a display interface (such as the display interface in FIG. 1), so that the network administrator according to the Analyze the network traffic determined by the server for subsequent network management. In the present application, the determining, by the analysis server, the network traffic specifically includes: receiving, by the forwarding device, a target data packet in a plurality of data packets obtained by dividing a data block, and according to the length and sequence included in the target data packet. Number, determining the network traffic corresponding to the data block.

With the above method provided by the present application, the forwarding device copies and transmits the target data packet of the plurality of data packets into one data block to the analysis server, and does not perform replication on other data packets except the target data packet. And sending the processing to the analysis server, so that the occupation of the storage capacity of the analysis server can be reduced, and at the same time, the transmission resource between the forwarding device and the analysis server can be saved.

An RDMA-based network traffic determining method provided by an embodiment of the present application is applicable to the network traffic determining system shown in FIG. 1. Referring to Figure 2, the specific process of the method includes:

Step 201: The forwarding device receives a data packet sent by the first terminal device, where the data packet is one of a plurality of data packets obtained by dividing a data block by the network card of the first terminal device; The first terminal device transmits the data block to the second terminal device through the forwarding device in an RDMA manner.

Generally, when RDMA is used to transmit data between terminal devices, the data block to be transmitted is divided into multiple data packets for transmission in the network. During the specific transmission, the network card of the first terminal device as the sender sends a plurality of data packets divided for one data block to the forwarding device, so that the forwarding device forwards the plurality of data packets to another terminal device, thereby Complete the data transfer.

Wherein each of the plurality of data packets includes a length of the data packet and a sequence number of the data packet, and specifically, a length of the data packet passes a payload in the data packet ( The payload field (as shown in the payload field in the packet shown in Figure 1a) is embodied, the sequence number of the data packet passing through the PSN field in the data packet (such as the PSN in the BTH in the data packet shown in Figure 1a) Field) embodies.

An optional implementation manner, where the forwarding device can be, but is not limited to, a switch or a router.

Step 202: The forwarding device determines whether the data packet is a target data packet in the multiple data packets, and if the data packet is a target data packet in the multiple data packets, copying the data packet Step 203 is executed. If the data packet is not the target data packet in the plurality of data packets, the process of copying the data packet is not performed, and step 204 is performed.

Since each of the plurality of data packets further includes an opcode field (the opcode field in the BTH in the data packet as shown in FIG. 1a), the data used to represent each data packet The packet type, that is, the value of opcode in each packet indicates that the packet is a first packet, an intermediate packet, or a trailer. Since the target data packet may be a tail packet or a first packet or a tail packet, optionally, when the target data packet is a tail packet in the multiple data packets, the forwarding device determines the data. Whether the packet is the target data packet in the plurality of data packets, the specific method may be: the forwarding device identifies the value of the opcode opcode in the data packet, if the opcode value is an opcode corresponding to the tail packet The value determines that the data packet is a tail packet; if the value of the opcode is not the value of the opcode corresponding to the tail packet, it is determined that the data packet is not a tail packet. When the target data packet is the first packet or the last packet of the multiple data packets, the forwarding device determines whether the data packet is a target data packet in the multiple data packets, and may be: The forwarding device identifies the value of the opcode opcode in the data packet, and if the value of the opcode is the value of the opcode corresponding to the first packet, determining that the data packet is the first packet, if the opcode value is an opcode corresponding to the tail packet The value of the packet is determined to be a tail packet.

An ACL is configured in the forwarding device, where the ACL stores the correspondence between the value of the opcode and the first packet, and the corresponding relationship between the value of the opcode and the tail packet. The forwarding device may determine whether the received data packet is the first packet or the tail packet by querying the correspondence between the value of the opcode in the ACL and the first packet, and the correspondence between the value of the opcode and the tail packet.

Through the foregoing method, the forwarding device can successfully identify a tail packet, or a first packet and a tail packet, of the plurality of data packets divided for one data block, and then copy the tail packet, or the first packet and the After the tail packet is described, the copied data packet is sent to the analysis server, so that the analysis server determines the network traffic corresponding to the data block by using the tail packet, or the first packet and the tail packet.

Since the actual RDMA mode is divided into three types: send, write, and read, the value of the opcode corresponding to the first packet in each mode is different, and the value of the opcode corresponding to the tail packet is also different. Therefore, optionally, the The forwarding device may determine, by using the opcode in the data packet, which type of RDMA mode the first packet or the tail packet is. For example, the forwarding device may determine that the data packet is a send first packet (send_first) or a send tail packet (send_last); the data packet may be determined to be a write first packet (write_first) or a write tail packet (write_last); The data packet is judged to be a read first packet (read_first) or a read end packet (read_last).

Step 203: The forwarding device sends the copied data packet to an analysis server, and forwards the data packet to the second terminal device.

Wherein, in step 203a shown in FIG. 2, the forwarding device forwards the data packet to the second terminal device, so that the network card of the second terminal device stores the data packet to the second The memory of the terminal device, thereby realizing the data transmission of the first terminal device and the second terminal device by using RDMA mode.

At the same time, as shown in step 203b of FIG. 2, the forwarding device sends the copied data packet (target data packet) to the analysis server, so that the analysis server subsequently stores only the data partitioned for one data block. The target data packets in the data packets, so that the storage capacity of the analysis server can be less.

Step 204: The forwarding device forwards the data packet to the second terminal device.

In step 204, the data packet sent by the forwarding device to the second terminal device is other than the target data packet in the plurality of data packets divided by the network card of the first terminal device for one data block. Packet. In this way, since the forwarding device does not perform the copy processing on the data packets other than the target data packet among the plurality of data packets, only the data packets other than the target data packet of the plurality of data packets are sent to the The second terminal device avoids copying and transmitting data packets other than the target data packet in the plurality of data packets to the analysis server, so that the analysis server can be prevented from storing the plurality of data packets. A data packet other than the target data packet, thereby reducing the occupation of the storage capacity of the analysis server, and saving transmission resources between the forwarding device and the analysis server.

Step 205: The analysis server determines network traffic corresponding to the data block according to the length of the target data packet and the sequence number of the target data packet.

Optionally, when the analyzing server performs step 205, the following two methods may be used:

In the first method, the target data packet includes a first packet and a tail packet of the multiple data packets, and the method may be divided into the following three steps:

A1, the analysis server receives a first packet of the plurality of data packets obtained by the forwarding device for the data block partitioning, where the first packet includes a length of the first packet and the first packet Serial number.

A2. The analysis server determines the number of data packets between the first packet and the tail packet according to the sequence number of the tail packet and the sequence number of the first packet.

A3. The analysis server determines network traffic corresponding to the data block according to the length of the first packet, the number of data packets between the first packet and the tail packet, and the length of the tail packet.

In the second method, the target data packet includes a tail packet of the plurality of data packets, and the method may be divided into the following three steps:

B1, the analysis server identifies the quintuple and the QP in the tail package, and determines the quintuple of the tail package and the MTU corresponding to the QP according to the saved quintuple and the correspondence between the QP and the MTU value. Value.

The quintual includes a source Internet Protocol (IP) address, a destination IP address, a source port, a destination port, and a transmission protocol; the QP is a QP field included in each data packet (as shown in FIG. 1a). The illustrated QP field) is used to identify the QP of the network card of the terminal device that needs to receive the data packet.

B2. The analysis server determines a starting sequence number of the set of the plurality of data packets, and determines, according to the sequence number of the tail package and the set start sequence number of the multiple data packets. The number of packets before the tail packet.

B3. The analysis server determines, according to the number of data packets before the tail packet, the quintuple of the tail packet, the value of the MTU corresponding to the QP, and the length of the tail packet, a network corresponding to the data block. flow.

In an optional implementation manner, the forwarding device first identifies the first packet of the multiple data packets because the multiple data packets are sent in the order from the first packet to the tail packet. The first packet is then copied and sent to the analysis server. Therefore, in the first method described above, in order to determine the first packet and the tail packet of the data block, the analysis server further performs the following operations:

After receiving the first packet of the plurality of data packets, the analysis server identifies and records the quintuple and the QP of the first packet; wherein the quintuple and the QP of the first packet are used to indicate the multiple The quintuple and QP of the data block to which the data packet belongs;

The analyzing server determines, according to the recorded quintuple and QP, the received tail packet that is the same as the recorded quintuple and QP, and uses the tail packet as the multiple of the data block The tail package in the packet.

After the first packet of the data block is received by the analysis server, the quintuple and the QP of the data block (that is, the identifier of the data block are recorded) are recorded, and the attribution sent by the subsequent forwarding device is identified. In the tail packet of the data block, the first packet and the tail packet of the data block are accurately determined, and the corresponding data block can be determined according to the first packet and the tail packet in the data block. Network traffic.

Further, the analysis server may determine, by using the foregoing method, a first packet and a tail packet respectively included in each of the plurality of data blocks, thereby determining network traffic corresponding to each of the plurality of data blocks.

Specifically, in the first method, when the analyzing server performs step A2, subtracting 1 from the serial number of the first packet and subtracting 1 from the serial number of the tail packet, the determined first header may be obtained. The number of packets between the packet and the trailer.

As is well known, the length of the first packet is the same as the length of the data packet between the first packet and the trailer. Therefore, specifically, when the analysis server performs step A3, the following formula may be adopted:

L=Len2+(N+1)*Len1

Optionally, when the starting sequence number of the plurality of data packets (that is, the value of the sequence number of the first packet) of the data packet divided by the network card of the first terminal device is a set value, the analysis server may adopt the foregoing The second method performs step 205.

Optionally, in step B1 of the foregoing second method, the analysis server stores a correspondence between a quintuple and a value of a QP and an MTU, and the analysis server determines, according to the correspondence, a five-element of the tail package. The value of the MTU corresponding to the group and the QP, that is, the length of the first packet and the intermediate data packet in the plurality of data packets of the data block are determined. For example, the correspondence between the quintuple and the QP and MTU values saved by the analysis server can be as shown in Table 1 below:

Table 1 Correspondence between quintuple and QP and MTU values

The source IP address, the source port, the destination IP address, the destination port, and the transmission protocol in Table 1 are five-tuples; the source QP and the destination QP are respectively represented in the network card of the terminal device that transmits data and the terminal device that receives the data. QP. Specifically, when the analysis server searches for the value of the quintuple of the tail package and the MTU corresponding to the QP, the analysis server determines that the QP in the tail package meets the source QP and the purpose in the table 1. Any one of QP can be.

In an optional implementation manner, the analysis server needs to update the saved quintuple and the correspondence between the QP and the MTU values in real time. The first terminal device first sends a connection request message to the second terminal device through the forwarding device, and then the second terminal device passes the foregoing, before the first terminal device and the second terminal device perform the data transmission. The forwarding device returns a connection response message, where the connection request message includes a value of an MTU that can be transmitted between the first terminal device and the second terminal device, and a QP of the network card of the first terminal device; The connection response message includes a QP of the network card of the second terminal device. The forwarding device may copy and send the connection request message and the connection response message to the analysis server. Therefore, the analysis server may be configured according to the quintuple of the connection request message, the value of the MTU included in the connection request message, the QP, and the quintuple of the connection response message and the connection response report. The QP included in the text obtains the correspondence between the quintuple and the QP (including the QP of the network card of the first terminal device and the QP of the network card of the second terminal device) and the value of the MTU, and then stores the obtained correspondence relationship into the In the analysis server, the update of the correspondence between the quintuple and the QP and MTU values is completed.

Optionally, in the foregoing step B2, the method for the analysis server to determine the starting sequence number of the multiple data packets that are set may be:

When the first terminal device carries the set start sequence number of the plurality of data packets in the sent data packet, the analysis server acquires the start sequence number by using the data packet sent by the forwarding device ;or

The analysis server acquires a pre-stored starting sequence number set when the first terminal device divides a plurality of data packets for one data block.

Further, in the above step B2, the analysis server determines the number of data packets before the tail package according to the sequence number of the tail package and the set start sequence number of the plurality of data packets. The specific method is: the analysis server subtracts the set start sequence number by using the serial number of the tail package, and obtains the number of data packets before the tail package.

For example, when the start sequence number is 0, the number of data packets before the tail package determined by the analysis server is equal to the sequence number of the tail package; when the start sequence number is 1, the The number of data packets before the tail packet determined by the analysis server is equal to the value obtained by subtracting 1 from the sequence number of the tail package.

Specifically, when the analysis server performs the above step B3, the following formula may be adopted:

T=Len+K*MTU

Wherein, T is the network traffic corresponding to the data block, the MTU is the value of the quintuple of the tail packet and the MTU corresponding to the QP, Len is the length of the tail packet, and K is the data before the tail packet. The number of packages.

According to the RDMA-based network traffic determining method provided by the embodiment of the present application, the forwarding device receives one of the plurality of data packets obtained by dividing the network card of the first terminal device for one data block, and determines whether the data packet is the plurality of data packets. a target data packet in the data packet, if yes, copying the data packet, transmitting the copied data packet to the analysis server, and forwarding the data packet to the second terminal device; otherwise, the forwarding device does not perform the copying device The processing of the data packet, the data packet is forwarded only to the second terminal device; wherein the data block is transmitted to the second terminal device by the forwarding device by using the RDMA mode in the first terminal device. Data block. In the method, the forwarding device only needs to copy and send the target data packet in the plurality of data packets divided for the data block to the analysis server, and the data packets other than the target data packet of the plurality of data packets are not copied and sent to the data packet. The processing of the server is analyzed to enable the analysis server to determine network traffic corresponding to the data block according to the target data packet, which can reduce the occupation of the storage capacity of the analysis server, and can save transmission resources between the forwarding device and the analysis server.

Based on the foregoing embodiment, the embodiment of the present application further provides a forwarding device, where the forwarding device is applied to the network traffic determining system shown in FIG. 1 , and is used to implement the RDMA-based network traffic determining method as shown in FIG. 2 . The function of the forwarding device. Referring to FIG. 3, the forwarding device 300 includes: a receiving unit 301, a processing unit 302, and a sending unit 303, where:

The receiving unit 301 is configured to receive a data packet sent by the first terminal device, where the data packet is one of a plurality of data packets obtained by dividing the network card of the first terminal device into one data block; the data block And the data block transmitted by the forwarding device to the second terminal device by using the RDMA mode in the RDMA mode;

The processing unit 302 is configured to determine whether the data packet is a target data packet in the multiple data packets, and if the data packet is a target data packet in the multiple data packets, copy the data Packet, controlling the sending unit 303 to send the copied data packet to the analysis server, and controlling the sending unit 303 to forward the data packet to the second terminal device; if the data packet is not the plurality of a target data packet in the data packet, the processing of copying the data packet is not performed, and the sending unit 303 is controlled to forward the data packet to the second terminal device;

The data packet includes a length of the data packet and a sequence number of the data packet.

Optionally, the target data packet is a tail packet of the multiple data packets, and the processing unit 302, when determining whether the data packet is a target data packet in the multiple data packets, specifically to:

Identifying the value of the opcode opcode in the data packet, if the value of the opcode is the value of the opcode corresponding to the tail packet, determining that the data packet is a tail packet; if the value of the opcode is not the corresponding one of the tail packet The value of opcode determines that the packet is not a tail packet.

Optionally, the target data packet is a first packet or a tail packet of the multiple data packets, and the processing unit 302 is configured to determine whether the data packet is a target data packet in the multiple data packets. Specifically for:

Identifying the value of the opcode in the data packet, if the value of the opcode is the value of the opcode corresponding to the first packet, determining that the data packet is the first packet; if the value of the opcode is the value of the opcode corresponding to the tail packet, It is determined that the data packet is a tail package.

Optionally, the forwarding device 300 is a switch or a router.

The forwarding device provided by the embodiment of the present application receives one of a plurality of data packets obtained by the network card of the first terminal device for one data block, and determines whether the data packet is a target data packet in the multiple data packets. If yes, copying the data packet, transmitting the copied data packet to the analysis server, and forwarding the data packet to the second terminal device; otherwise, the forwarding device does not perform the process of copying the data packet, only Transmitting the data packet to the second terminal device; wherein the data block is a data block that is transmitted by the forwarding device to the second terminal device by using the RDMA mode by the first terminal device. In this way, the forwarding device only needs to copy and send the target data packet in the plurality of data packets divided for the data block to the analysis server, and the data packets other than the target data packet of the plurality of data packets are not copied and sent to the analysis server. Processing, so that the analysis server determines network traffic corresponding to the data block according to the target data packet, which can reduce the occupation of the storage capacity of the analysis server, and can save transmission resources between the forwarding device and the analysis server.

Based on the above embodiment, the embodiment of the present application further provides an analysis server, where the analysis server is applied to the network traffic determination system shown in FIG. 1 , and is used to implement the RDMA-based network traffic determination method as shown in FIG. 2 . Analyze the capabilities of the server. Referring to FIG. 4, the analysis server 400 includes a receiving unit 401 and a processing unit 402, wherein:

The receiving unit 401 is configured to receive a tail packet of the plurality of data packets obtained by the forwarding device, where the data block is transmitted by the forwarding device by using the RDMA mode by the first terminal device. a data block for the second terminal device; the tail packet includes a length of the tail packet and a sequence number of the tail packet;

The processing unit 402 is configured to determine network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet.

Optionally, the receiving unit 401 is further configured to:

Receiving, by the forwarding device, a first packet of the plurality of data packets obtained by dividing the data block, where the first packet includes a length of the first packet and a sequence number of the first packet;

The processing unit 402 is configured to: when determining the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, specifically:

Determining, according to the sequence number of the tail packet and the sequence number of the first packet, the number of data packets between the first packet and the tail packet;

And determining network traffic corresponding to the data block according to the length of the first packet, the number of data packets between the first packet and the tail packet, and the length of the tail packet.

Optionally, when the processing unit 402 determines network traffic corresponding to the data block, the following formula is adopted:

L=Len2+(N+1)*Len1

Optionally, the processing unit 402 is configured to: when determining network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, specifically:

Identifying the quintuple and the queue pair QP in the tail packet, and determining the quintuple of the tail package and the MTU corresponding to the QP according to the saved quintuple and the correspondence between the QP and the value of the maximum transmission unit MTU value;

Determining a set start sequence number of the plurality of data packets, and determining a data packet before the tail package according to the sequence number of the tail package and the set start sequence number of the multiple data packets Number of

And determining network traffic corresponding to the data block according to the number of data packets before the tail packet, the value of the MTU of the tail packet and the MTU corresponding to the QP, and the length of the tail packet.

T=Len+K*MTU

Optionally, the processing unit 402 is further configured to:

After the receiving unit 401 receives the first packet of the plurality of data packets, identifying a quintuple and a QP of the first packet and recording; wherein the quintuple and QP of the first packet are used to indicate a quintuple and a QP of the data block to which the plurality of data packets belong;

Determining, according to the recorded quintuple and QP, the same tail packet as the recorded quintuple and QP, and using the tail packet as the plurality of data packets in the data block Tail package.

Optionally, the forwarding device is a switch or a router.

The analysis server provided by the embodiment of the present application, after receiving the tail packet in the plurality of data packets obtained by dividing the data packet by the forwarding device, determining, according to the length and the serial number of the tail packet, the corresponding data block Network traffic. In this way, the analysis server only needs to receive the tail packets in the plurality of data packets obtained by dividing each data block sent by the forwarding device, which can reduce the occupation of the storage capacity of the analysis server, and can save between the forwarding device and the analysis server. Transfer resources.

It should be noted that the division of the unit in the embodiment of the present application is schematic, and is only a logical function division. In actual implementation, there may be another division manner. The functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program code. .

Based on the above embodiment, the embodiment of the present application further provides a forwarding device, where the forwarding device is applied to the network traffic determining system shown in FIG. 1 , and is used to implement the RDMA-based network traffic determining method as shown in FIG. 2 . The function of forwarding devices in the middle. Referring to FIG. 5, the forwarding device 500 includes a communication interface 501, a processor 502, and a memory 503, where:

The communication interface 501, the processor 502, and the memory 503 are connected to each other. Optionally, the communication interface 501, the processor 502, and the memory 503 are mutually connected by a bus 504; the bus 504 may be a peripheral component interconnect (PCI) bus or an extended industry standard. Extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 5, but it does not mean that there is only one bus or one type of bus.

The forwarding device 500 implements the RDMA-based network traffic determination method as shown in FIG. 2:

The communication interface 501 is configured to receive and send data;

The processor 502 is configured to control the communication interface 501 to receive a data packet sent by the first terminal device, where the data packet is in a plurality of data packets obtained by dividing the network card of the first terminal device into one data block. a data block is a data block that is transmitted by the forwarding device to the second terminal device by using the RDMA mode by the first terminal device;

Determining whether the data packet is a target data packet in the plurality of data packets, and if the data packet is a target data packet in the plurality of data packets, copying the data packet, and controlling the communication interface 501 Transmitting the copied data packet to the analysis server, and controlling the communication interface 501 to forward the data packet to the second terminal device; if the data packet is not the target data packet in the multiple data packets, Not performing the process of copying the data packet, and controlling the communication interface 501 to forward the data packet to the second terminal device;

Optionally, the target data packet is a tail packet of the multiple data packets, and the processor 502, when determining whether the data packet is a target data packet in the multiple data packets, specifically to:

Optionally, the target data packet is a first packet or a tail packet of the multiple data packets, and the processor 502 is configured to determine whether the data packet is a target data packet in the multiple data packets. Specifically for:

Optionally, the forwarding device 500 is a switch or a router.

The memory 503 is configured to store a program or the like. In particular, the program can include program code, the program code including computer operating instructions. The memory 503 may include a RAM, and may also include a non-volatile memory, such as at least one disk storage. The processor 502 executes an application stored in the memory 503 to implement the foregoing functions, thereby implementing an RDMA-based network traffic determining method as shown in FIG. 2.

Based on the above embodiment, the embodiment of the present application further provides an analysis server, where the analysis server is applied to the network traffic determination system shown in FIG. 1 , and is used to implement the RDMA-based network traffic determination method as shown in FIG. 2 . The function of the analysis server. Referring to FIG. 6, the analysis server 600 includes a communication interface 601, a processor 602, and a memory 603, where:

The communication interface 601, the processor 602, and the memory 603 are connected to each other. Optionally, the communication interface 601, the processor 602, and the memory 603 are connected to each other by a bus 604; the bus 604 may be a peripheral component interconnect (PCI) bus or an extended industry standard. Extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 6, but it does not mean that there is only one bus or one type of bus.

The analysis server 600 implements the RDMA-based network traffic determination method as shown in FIG. 2:

The communication interface 601 is configured to receive and send data;

The processor 602 is configured to control the communication interface 601 to receive a tail packet in a plurality of data packets obtained by dividing the data block by the forwarding device, where the data block is an RDMA mode used by the first terminal device. a data block transmitted by the forwarding device to the second terminal device; the tail packet includes a length of the tail packet and a sequence number of the tail packet;

And determining network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet.

Optionally, the processor 602 is further configured to:

Controlling the communication interface 601 to receive a first packet of the plurality of data packets obtained by the forwarding device for the data block partitioning; wherein the first packet includes a length of the first packet and the first packet Serial number;

The processor 602 is specifically configured to: when determining the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet,

Optionally, when the processor 602 determines network traffic corresponding to the data block, the following formula is adopted:

L=Len2+(N+1)*Len1

Optionally, the processor 602 is specifically configured to: when determining network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet,

T=Len+K*MTU

Optionally, the processor 602 is further configured to:

After receiving the first packet of the plurality of data packets, the communication interface 601 identifies and records the quintuple and the QP of the first packet; wherein the quintuple and QP of the first packet are used to indicate a quintuple and a QP of the data block to which the plurality of data packets belong;

Optionally, the forwarding device is a switch or a router.

The memory 603 is configured to store a program or the like. In particular, the program can include program code, the program code including computer operating instructions. The memory 603 may include a RAM, and may also include a non-volatile memory such as at least one disk storage. The processor 602 executes an application stored in the memory 603 to implement the foregoing functions, thereby implementing an RDMA-based network traffic determining method as shown in FIG. 2.

In summary, the RDMA-based network traffic determining method and apparatus provided by the embodiment of the present application, the forwarding device receives one of a plurality of data packets obtained by dividing the network card of the first terminal device for one data block, and determining the data. Whether the packet is a target data packet in the plurality of data packets, and if so, copying the data packet, transmitting the copied data packet to an analysis server, and forwarding the data packet to the second terminal device; otherwise The forwarding device does not perform the process of copying the data packet, and only forwards the data packet to the second terminal device; wherein the data block is transmitted by the forwarding device by using the RDMA mode in the first terminal device. A data block for the second terminal device. In the method, the forwarding device only needs to copy and send the target data packet in the plurality of data packets divided for the data block to the analysis server, and the data packets other than the target data packet of the plurality of data packets are not copied and sent to the data packet. The processing of the server is analyzed to enable the analysis server to determine network traffic corresponding to the data block according to the target data packet, which can reduce the occupation of the storage capacity of the analysis server, and can save transmission resources between the forwarding device and the analysis server.

Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system, or computer program product. Thus, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware. Moreover, the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

It is apparent that those skilled in the art can make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the embodiments of the present application. Thus, it is intended that the present invention cover the modifications and variations of the embodiments of the present invention.

Claims

A method for determining network traffic based on remote direct data access RDMA, comprising:

The forwarding device receives the data packet sent by the first terminal device, where the data packet is one of a plurality of data packets obtained by dividing the network card of the first terminal device into one data block; the data block is the first terminal The data block transmitted by the device to the second terminal device by using the forwarding device by using the RDMA method;

Determining, by the forwarding device, whether the data packet is a target data packet in the plurality of data packets, and if the data packet is a target data packet in the plurality of data packets, copying the data packet, and copying the data packet The obtained data packet is sent to the analysis server, and the data packet is forwarded to the second terminal device; if the data packet is not the target data packet in the plurality of data packets, the forwarding device does not perform the copying Processing the data packet, forwarding the data packet to the second terminal device;

The data packet includes a length of the data packet and a sequence number of the data packet.
The method according to claim 1, wherein said target data packet is a tail packet of said plurality of data packets, and said forwarding device determines whether said data packet is a target of said plurality of data packets Packets, including:

The forwarding device identifies the value of the opcode opcode in the data packet. If the value of the opcode is the value of the opcode corresponding to the tail packet, determining that the data packet is a tail packet; if the value of the opcode is not The value of the opcode corresponding to the tail packet determines that the data packet is not a tail packet.
The method according to claim 1, wherein the target data packet is a first packet or a tail packet of the plurality of data packets, and the forwarding device determines whether the data packet is the plurality of data packets. Target packets in, including:

The forwarding device identifies the value of the opcode opcode in the data packet. If the value of the opcode is the value of the opcode corresponding to the first packet, the data packet is determined to be the first packet, and if the value of the opcode is the tail packet, The value of the opcode determines that the packet is a trailer.
A method for determining network traffic based on remote direct data access RDMA, comprising:

The analysis server receives the tail packet of the plurality of data packets obtained by the forwarding device for the data block, wherein the data block is transmitted by the forwarding device to the second terminal device by using the RDMA mode by the first terminal device. a data block; the tail packet includes a length of the tail packet and a sequence number of the tail packet;

The analysis server determines network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet.
The method of claim 4, wherein the method further comprises:

The analysis server receives a first packet of the plurality of data packets obtained by the forwarding device for the data block division, where the first packet includes a length of the first packet and a sequence of the first packet number;

The analyzing server determines the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, including:

The analyzing server determines the number of data packets between the first packet and the tail packet according to the sequence number of the tail packet and the sequence number of the first packet;

The analysis server determines network traffic corresponding to the data block according to the length of the first packet, the number of data packets between the first packet and the tail packet, and the length of the tail packet.
The method according to claim 5, wherein when the analysis server determines the network traffic corresponding to the data block, the following formula is adopted:

L=Len2+(N+1)*Len1

Where L is the network traffic corresponding to the data block, Len1 is the length of the first packet, Len2 is the length of the tail packet, and N is the number of data packets between the first packet and the tail packet. .
The method of claim 4, wherein the analyzing server determines network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, including:

The analysis server identifies a quintuple and a queue pair QP in the tail package, and determines a quintuple and a QP of the tail package according to a correspondence between the saved quintuple and the value of the QP and the maximum transmission unit MTU. The value of the corresponding MTU;

Determining, by the analysis server, a starting sequence number of the plurality of data packets that are set, and determining the tail package according to a sequence number of the tail package and a starting sequence number of the plurality of data packets that are set The number of previous packets;

The analysis server determines the network traffic corresponding to the data block according to the number of data packets before the tail packet, the quintuple of the tail packet, the value of the MTU corresponding to the QP, and the length of the tail packet.
The method according to claim 7, wherein when the analysis server determines the network traffic corresponding to the data block, the following formula is adopted:

T=Len+K*MTU

Wherein, T is the network traffic corresponding to the data block, the MTU is the value of the quintuple of the tail packet and the MTU corresponding to the QP, Len is the length of the tail packet, and K is the data packet before the tail packet. The number.
The method of claim 5 or claim 6, wherein the method further comprises:

After receiving the first packet of the plurality of data packets, the analysis server identifies and records the quintuple and the QP of the first packet; wherein the quintuple and the QP of the first packet are used to indicate the multiple The quintuple and QP of the data block to which the data packet belongs;

The analyzing server determines, according to the recorded quintuple and QP, the received tail packet that is the same as the recorded quintuple and QP, and uses the tail packet as the multiple of the data block The tail package in the packet.
A forwarding device, comprising:

a memory for storing program instructions;

a communication interface for receiving and transmitting data;

a processor for invoking program instructions stored in the memory to perform the following method:

Controlling, by the communication interface, a data packet sent by the first terminal device, where the data packet is one of a plurality of data packets obtained by dividing a data block by the network card of the first terminal device; The first terminal device uses the RDMA mode to transmit the data block to the second terminal device by using the forwarding device;

Determining whether the data packet is a target data packet in the plurality of data packets, if the data packet is a target data packet in the plurality of data packets, copying the data packet, and controlling the communication interface Copying the obtained data packet to the analysis server, and controlling the communication interface to forward the data packet to the second terminal device; if the data packet is not the target data packet in the plurality of data packets, not executing Copying the data packet, and controlling the communication interface to forward the data packet to the second terminal device;

The data packet includes a length of the data packet and a sequence number of the data packet.
The forwarding device according to claim 10, wherein said target data packet is a tail packet of said plurality of data packets, and said processor determines whether said data packet is said plurality of data packets When the target packet is used, it is specifically used to:

Identifying the value of the opcode opcode in the data packet, if the value of the opcode is the value of the opcode corresponding to the tail packet, determining that the data packet is a tail packet; if the value of the opcode is not the corresponding one of the tail packet The value of opcode determines that the packet is not a tail packet.
The forwarding device according to claim 10, wherein the target data packet is a first packet or a tail packet of the plurality of data packets, and the processor determines whether the data packet is the plurality of packets When the target packet in the packet is used, it is specifically used to:

Identifying the value of the opcode opcode in the data packet, if the value of the opcode is the value of the opcode corresponding to the first packet, determining that the data packet is the first packet; if the value of the opcode is the value of the opcode corresponding to the tail packet And determining that the data packet is a tail package.
An analysis server, comprising:

a memory for storing program instructions;

a communication interface for receiving and transmitting data;

a processor for invoking program instructions stored in the memory to perform the following method:

Controlling, by the communication interface, a tail packet of a plurality of data packets obtained by dividing the data packet by the forwarding device, where the data block is transmitted to the second terminal device by using the RDMA mode by using the forwarding device a data block of the terminal device; the tail packet includes a length of the tail packet and a sequence number of the tail packet;

And determining network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet.
The analysis server according to claim 13, wherein the processor is further configured to:

Controlling, by the communication interface, a first packet of the plurality of data packets obtained by the forwarding device for the data block division; wherein the first packet includes a length of the first packet and the first packet serial number;

The processor, when determining the network traffic corresponding to the data block according to the length of the tail packet and the sequence number of the tail packet, specifically used to:

Determining, according to the sequence number of the tail packet and the sequence number of the first packet, the number of data packets between the first packet and the tail packet;

And determining network traffic corresponding to the data block according to the length of the first packet, the number of data packets between the first packet and the tail packet, and the length of the tail packet.
The analysis server according to claim 14, wherein when the processor determines network traffic corresponding to the data block, the following formula is adopted:

L=Len2+(N+1)*Len1

Where L is the network traffic corresponding to the data block, Len1 is the length of the first packet, Len2 is the length of the tail packet, and N is the number of data packets between the first packet and the tail packet. .
The analysis server according to claim 13, wherein the processor is specifically configured to determine network traffic corresponding to the data block according to a length of the tail packet and a sequence number of the tail packet. :

Identifying the quintuple and the queue pair QP in the tail packet, and determining the quintuple of the tail package and the MTU corresponding to the QP according to the saved quintuple and the correspondence between the QP and the value of the maximum transmission unit MTU value;

Determining a set start sequence number of the plurality of data packets, and determining a data packet before the tail package according to the sequence number of the tail package and the set start sequence number of the multiple data packets Number of

And determining network traffic corresponding to the data block according to the number of data packets before the tail packet, the value of the MTU of the tail packet and the MTU corresponding to the QP, and the length of the tail packet.
The analysis server according to claim 16, wherein when the processor determines network traffic corresponding to the data block, the following formula is adopted:

T=Len+K*MTU

Wherein, T is the network traffic corresponding to the data block, the MTU is the value of the quintuple of the tail packet and the MTU corresponding to the QP, Len is the length of the tail packet, and K is the data packet before the tail packet. The number.
The analysis server according to claim 14 or 15, wherein the processor is further configured to:

After the first interface of the plurality of data packets is received by the communication interface, the quintuple and the QP of the first package are identified and recorded; wherein the quintuple and the QP of the first package are used to indicate the a quintuple and a QP of the data block to which the plurality of data packets belong;

Determining, according to the recorded quintuple and QP, the same tail packet as the recorded quintuple and QP, and using the tail packet as the plurality of data packets in the data block Tail package.
A computer storage medium, wherein the computer storage medium stores computer executable instructions, the computer executable instructions, when called by the computer, for causing the computer to perform any of claims 1-9 The method described in the item.