CN102104548B - Method and device for receiving and processing data packets - Google Patents

Abstract

The present invention provides a data packet receiving and processing method, comprising: receiving a data packet, extracting stream classification operation information during the process of receiving the data packet, and slicing the data packet according to a preset size to obtain a data slice, and each time a data slice is received Then store it in the storage space corresponding to the data packet; when the data packet is received, perform a flow classification operation on the data packet according to the flow classification operation information, and after obtaining the flow classification operation result, extract relevant information to initiate a follow-up operation . The invention also provides a device for receiving and processing data packets. By adopting the method of the invention, the processing time of long packets can be significantly reduced, and the delay of the system can be reduced.

Description

Method and device for receiving and processing data packets

technical field

The present invention relates to the field of data communication. Specifically, it is a method and device for receiving and processing data packets.

Background technique

High-speed data services require that the processing delay of the system be as small as possible. When a data packet enters the data service system, it generally needs to go through three operations: reception, flow classification and storage. During this process, the time spent on receiving and storing the data packet and the storage space resources are proportional to the length of the data packet. According to the traditional method, the start of each step of operation needs to wait for the completion of the previous step. For a data packet with a long length, such as 9600byte, it is very expensive in terms of time and on-chip RAM resources of the device, and even becomes a problem for the system. bottleneck.

When the system adopts the traditional processing method, the time spent on receiving a data packet is the sum of the time t1 for receiving the data packet, the time t2 for the flow classification operation of the data packet, and the time t3 for storing the data packet in the cache. in:

The time t1 of data packet reception: t1=len/v1, wherein, len is the packet length of data packet, and v1 is the input port rate of this data packet;

The time t2 of the data packet flow classification operation can be regarded as a fixed value regardless of the length of the data packet;

The time t3 at which the data packet is stored in the cache: t3=len/v2, where len is the packet length of the data packet, and v2 is the storage rate of the system cache.

Therefore, the time consumed by the whole process is t1+t2+t3=len/v1+t2+len/v2. For a fixed system, v1 and v2 are fixed values, so the time taken to determine the entire process is the length len of the data packet.

When the maximum packet length supported by the system is 9600byte, the data packet input port is 1Gbps, and the storage rate of the system cache is 20Gbps, the maximum delay from the beginning to the end of the cache is: 9600×8/(10 ⁹ )+9600×8 /(2×10 ¹⁰ )+t2, that is, 8.064×10 ^-5 +t2(s).

In addition, for the reception of data packets, the on-chip RAM resources of the device required by the data port should at least support the space of two full packets of data with the largest packet length, that is, 9600×8×2=153.6kb. If the number of ports input by the system is more than In the case of one, as the number of ports increases, the overhead of this resource becomes progressively unacceptable.

Contents of the invention

The technical problem to be solved by the present invention is to provide a data packet receiving and processing method and device to reduce data processing delay.

In order to solve the above problems, the present invention provides a device for receiving and processing data packets, including an input control module, a message processing module and a packet management module, wherein:

The input control module is used to: receive data packets, extract stream classification operation information during the process of receiving data packets, slice data packets according to preset sizes, and send each data slice to the packet management module, after receiving the data packet, sending the flow classification operation information to the packet processing module; and receiving the flow classification operation result of the data packet, and sending the flow classification operation result to the packet management module ;

The packet processing module is configured to: perform a flow classification operation on the data packet according to the flow classification operation information, and send the obtained flow classification operation result to the input control module;

The packet management module is configured to: when receiving the data slice of the data packet sent by the input control module, store it in the storage space corresponding to the data packet; and, after receiving the flow classification operation result of the data packet , to extract relevant information and initiate follow-up operations.

Further, the above device may also have the following feature, the preset size is the smallest unit or a multiple of the smallest unit for system cache operation.

Further, the above-mentioned device may also have the following features, the input control module is further configured to: send the data slice to the packet management module, and send the flow classification operation information to the packet processing module , and, when sending the flow classification operation result to the packet management module, carry the port identification information of the input port of the data packet;

The message processing module is further configured to save the port identification information when receiving the flow classification operation information and port identification information, and carry the port identification information when sending the flow classification operation result to the input control module information;

The packet management module is further configured to: when receiving the data slice and port identification information, or receive the flow classification operation result and port identification information, perform the data slice or flow classification operation according to the port identification information The result is sent to the corresponding data port for processing.

Further, the above device may also have the following features, the flow classification operation result includes packet modification information and packet descriptor information;

The package management module is further configured to: store the message modification information of the data package in the storage space corresponding to the data package, and the extracting related information includes extracting the message descriptor information of the data package.

Further, the above-mentioned device can also have the following features, the input control module is also used for: in the process of receiving the data packet, save the data of the first N bytes of the data packet, from the N bytes after the data packet Start slicing, and the N is the difference between the preset size and the message modification information; after receiving the flow classification operation result of the data packet, combine the first N bytes of data with the data packet The message modification information of the packet is composed of a packet header slice and sent to the packet management module;

The packet management module is further configured to: when storing the data slice of the data packet, reserve the first storage block of the storage space corresponding to the data packet; and, when receiving the packet header slice of the data packet, Store it in the first storage block of the storage space corresponding to the data packet.

Further, the above-mentioned device can also have the following features, the input control module is also used for: after receiving a data packet, receive the next data packet during the process of waiting for and receiving the flow classification operation result of the data packet operation, and send the data slice of the next data packet to the packet management module;

The packet management module is further configured to: after receiving the data slice of the data packet and waiting for the flow classification operation result of the data packet, when receiving the data slice of the next data packet of the data packet, Store it in the storage space corresponding to the next data packet.

The present invention also provides a data packet receiving and processing method, including:

Receive data packets, in the process of receiving data packets, extract the flow classification operation information, slice the data packets according to the preset size to obtain data slices, and store each received data slice into the storage space corresponding to the data packet;

After the data packet is received, perform a flow classification operation on the data packet according to the flow classification operation information, and after obtaining a flow classification operation result, extract relevant information to initiate a follow-up operation.

Further, the above method may also have the following feature, the preset size is the smallest unit or a multiple of the smallest unit for system cache operation.

Further, the above-mentioned method can also have the following characteristics, and the method also includes:

When processing data packets of multiple input ports, the port identification information is used to distinguish data slices, flow classification operation information and flow classification operation results of data packets of different input ports.

Further, the above method may also have the following features, the flow classification operation result includes message modification information and message descriptor information; the extracting related information includes extracting the message descriptor information of the data packet;

The method further includes storing the message modification information of the data packet in a storage space corresponding to the data packet.

Further, the above method may also have the following characteristics, the method also includes: during the process of receiving the data packet, saving the data of the first N bytes of the data packet, and starting to slice after the N bytes of the data packet , the N is the difference between the preset size and the message modification information; when storing the data slice of the data packet, retain the first storage block of the storage space corresponding to the data packet;

After obtaining the flow classification operation result of the data packet, the first N bytes of data and the message modification information of the data packet are formed into a packet header slice and stored in the first storage block of the storage space corresponding to the data packet .

Further, the above method may also have the following characteristics, the method also includes, after receiving a data packet, during the process of waiting for and receiving the flow classification operation result of the data packet, receiving, slicing and Data slice storage operations.

By adopting the method of the present invention, compared with the existing scheme, since the data packet does not need to wait for the receiving to be completed, the storage operation can be performed, and the storage time and the receiving time are multiplexed to a certain extent. Therefore, the time spent in the whole process is roughly the time of receiving a data packet and the storage time of a data slice. For the long packet, the storage time of the data packet is basically saved, therefore, the processing time of the long packet can be significantly reduced, and the delay of the system can be reduced.

At the same time, since it is not necessary to completely receive a data packet before initiating a storage operation, but only needs to be able to support the data size of the data cache operation, the operation from receiving to buffering can be completed. Therefore, the on-chip RAM resource overhead of the device is determined by The space for the two longest packets is significantly reduced to two data slices.

In addition, the present invention stores part of the message modification information of the flow classification operation into the external storage space together with the message data entity without exceeding the processing capacity of the external storage space, and then directly transmits it to the output module without going through the queue. The management part reduces the burden of queue management storage.

Description of drawings

Fig. 1 is a system block diagram of the present invention;

Fig. 2 is the interface schematic diagram of IPC and PP of the present invention;

Fig. 3 is the interface schematic diagram of IPC and PMG of the present invention;

Fig. 4 is the processing schematic diagram of IPC of the present invention;

FIG. 5 is a flow chart of the PMG process of the present invention.

Detailed ways

Since the traffic classification operation needs the verification information of the entire data packet, it must wait until the end of the data packet enters the system to start, so the time of receiving the data packet and the flow classification operation cannot be saved. What can be changed is that the data packet is stored in the cache. operate.

The core idea of the present invention is: to provide a data packet receiving and processing method, so that the data packet storage operation and the data packet receiving operation start at the same time, and the operation of storing the data packet in the cache does not need to wait until the receiving operation and the flow classification operation are completed. Therefore, the time for storing the data packet is saved, and the processing delay and resource consumption of the system are reduced.

The present invention provides a data packet receiving and processing device, as shown in Figure 1, comprising three parts: an input control module (IPC), a message processing module (PP) and a packet management module (PMG), wherein:

The IPC is used to: receive data packets, extract flow classification operation information during the process of receiving data packets, slice the data packets according to the preset size, and send each data slice to the PMG when the data is received After the packet is received, send the flow classification operation information to the PP; and receive the flow classification operation result of the data packet, and send the flow classification operation result to the PMG;

The PP is configured to: perform a flow classification operation on the data packet according to the flow classification operation information, and send the obtained flow classification operation result to the IPC;

The PMG is used to: when receiving the data slice of the data packet sent by the input control module, store it in the storage space corresponding to the data packet; and, after receiving the flow classification operation result of the data packet, extract Relevant information initiates follow-up operations. Data packets can be stored in DDR (Double DataRate, double-rate synchronous dynamic random access memory); and complete the reassembly of data packets in DDR.

The IPC is further configured to: send the data slice to the PMG, and, send the traffic classification operation information to the PP, and, when sending the traffic classification operation result to the PMG, carry the Port identification information of the input port of the data packet;

The PP is further configured to store the port identification information when receiving the flow classification operation information and port identification information, and carry the port identification information when sending the flow classification operation result to the IPC;

The PMG is further configured to: when receiving the data slice and port identification information, or receiving the flow classification operation result and port identification information, send the data slice or flow classification operation result according to the port identification information to the corresponding data port for processing.

Wherein, the PMG is further configured to: store the message modification information of the data packet in the storage space corresponding to the data packet, and the extracting related information includes extracting the message descriptor information of the data packet.

Wherein, the input control module is also used for: in the process of receiving the data packet, saving the first N bytes of data of the data packet, and starting to slice after the N bytes of the data packet, and the N is the The difference between the preset size and the message modification information; after receiving the flow classification operation result of the data packet, the first N bytes of data and the message modification information of the data packet are used to form a packet header slice Send to the package management module;

The packet management module is further configured to: when storing the data slice of the data packet, reserve the first storage block of the storage space corresponding to the data packet; and, when receiving the packet header slice of the data packet, Store it in the first storage block of the storage space corresponding to the data packet.

The port identification information is a port number.

The interface between IPC and PP is shown in Figure 2. The interface from IPC to PP includes effective signals, traffic classification operation information, and port numbers; the interface from PP to IPC includes traffic classification operation results, port numbers, and effective signal interfaces. The valid signal interface is used to transmit enabling control signals, the port number interface is used to transmit port identification information, the traffic classification operation information interface is used to transmit traffic classification operation information, and the traffic classification operation result interface is used to transmit traffic classification operation information.

The interface between IPC and PMG is shown in Figure 3, including effective signal, data channel, start signal, end signal, flow classification identifier and packet descriptor information interface, wherein the effective signal interface is used to transmit enable control signal, data The channel is used to transmit data slices and packet header slices (see the embodiment for specific meanings), the start signal interface is used to transmit the start signal of the data packet, the end signal interface is used to transmit the end signal of the data packet, and the flow classification identification interface is used to Indicates whether there is a traffic classification operation result, and the packet descriptor information interface is used to transmit packet descriptor information.

The operation process at the IPC is as follows:

Step 101, the IPC receives the data from the input port, slices the data packet according to the preset size, and extracts the information required for the traffic classification operation according to the requirements of the PP, referred to as the traffic classification operation information;

Wherein, the preset size is the smallest unit of system cache operation or an integer multiple thereof.

Step 102, every time the IPC receives a data slice, it sends the data slice and port number to the PMG for caching;

Step 103, when the data packet is received, the flow classification operation information and port number of the data packet are sent to the PP;

Step 104, the IPC waits for and receives the packet flow classification operation result returned by the PP, and sends it to the PMG to complete the transmission of the packet; while the IPC waits for the packet flow classification operation result returned by the PP, it can Receive and buffer subsequent data packets.

The operation process at PP is as follows:

Step 201, the PP receives the flow classification operation information and the port number of the data packet sent by the IPC;

Step 202, the PP performs the flow classification operation of the data packet according to the flow classification operation information;

Wherein, the PP needs to guarantee the sequence (first in, first out) of the flow classification operation of the data packets.

Step 203, PP sends the result of the flow classification operation and the port information corresponding to the data packet back to the IPC;

The operations at the PMG are as follows:

Step 301, PMG receives the start signal of the data packet sent by IPC, and applies for storage space;

Step 302, sequentially storing the data slices sent by the IPC until the end signal of the data packet is received;

Step 303, PMG carries out the storage operation (repeating step 301 and step 302) of follow-up data packet, and monitors the flow classification operation result of data packet simultaneously;

Step 304, if the flow classification operation result of the data packet is received, store the message modification information of the data packet in the storage space corresponding to the data packet, and extract the corresponding information to initiate the next operation.

The corresponding information includes packet descriptor information of the data packet.

During the whole process, the data stored in the cache by the data packet is actually divided into two parts: the entity of the data packet and the result of the flow classification operation of the data packet. The storage of the entity part of the data packet and the reception of the data packet are carried out simultaneously, while The storage of the result of the class classification operation is to wait for the result of the flow classification operation to be sent back to the IPC after receiving the data packet.

The present invention is not only applicable to single-port input systems, but also applicable to multi-port input systems. It only needs to add port identification information to separate the data between each port number, and the processing method of each port is the same as that of a single port. are handled in the same way. In the case of a single port, when transmitting information (data slices, flow classification operation information, flow classification operation information) between modules, it is not necessary to carry port identification information.

In order to illustrate the technical solution of the present invention more clearly, a specific process for realizing the solution will be introduced below, but not as a limitation of the present invention.

Assume that there are 20 1Gbps data ports on the IPC side; the data required for PP processing is the first 128 bytes of the data packet and the CRC check information of the data packet, and the PP returns a 48-byte message modification information to be added in front of the data message, supporting For subsequent message modification, 66bit message descriptor information is transmitted for subsequent message scheduling operations; the data size of each storage operation of the DDR controller is 64byte (512bit), so the data bit width from IPC to PMG can be selected 64byte (512bit), the size of the data slice is 64byte (512bit); 2048byte is the smallest space unit in DDR; the data packet length supported by the system is 64byte-9600byte.

The processing flow of IPC is shown in Figure 4, and the specific steps for a single port to process data packets are as follows:

After the IPC receives the initial data of the data packet, it stores the first 16 bytes of the data packet into the preset FIFO (First Input First Output, first in, first out), which is called the header slice FIFO;

Since the packet modification information is 48 bytes, which is less than the size of one cache operation, 16 bytes of data are required to form data slices together for cache operations. Since the message modification information should be stored in front of the message data (to facilitate the message modification operation), the first 16 bytes of the message are pre-intercepted and cached, and when the message modification information is sent back to the IPC, it is assembled into a packet header slice , and send it to the PMG for caching, and transfer the first slice of the packet and at the same time transfer the message description information corresponding to the data packet. Of course, the 16-byte data may not be pre-stored, and the packet modification information is directly formed into a packet header slice and sent to the PMG for caching.

The IPC extracts the first 128 bytes of the data packet and stores them in a fixed storage space (differentiated by the port number). After the data packet is received, it transmits the CRC check result to the PP and initiates a flow classification operation;

IPC transmits data slices in units of 64 bytes from the 17byte of the data packet, and sends them to PMG in sequence, adds the SOP signal to the first data slice as the start of a new packet, and adds the SOP signal to the last slice of the data packet EOP signal, indicating that the entity information of the data packet has been received;

Monitor the flow classification operation result of the data packet, assemble the received flow classification operation result and the first data in the header slice FIFO into the first slice of the data packet, and pass the corresponding 66-bit packet descriptor information to PMG together Store and initiate the next step.

The result of the flow classification operation is divided into two parts: 1. Message modification information, this part does not need to be passed to the queue management part of the system, and is directly stored in the external cache (in DDR) together with the message to reduce the storage of queue management Burden, when the data packet is dequeued, it is directly passed to the output module of the system for message modification; 2. Message descriptor information (66bit), information required for queue management and processing.

The IPC receives the result of the traffic classification operation, and mainly performs the following operations: There are two data channels from the IPC to the PMG, one is a data channel, and the data and modification information of the message are transmitted, and then cached by the PMG to the DDR, and the other is the data channel. After the PMG, the text descriptor information is added with the PMAU identifier corresponding to the message (the DDR storage space information of the message), and passed to the subsequent queue management together. Among them, the message modification information (48byte) and the reserved first 16byte data of the message form a data slice (packet header slice of the message) for transmission. While transmitting this slice, the message descriptor information is synchronously Channel, which transmits the message descriptor information of the data packet to the PMG.

In the process of PP, the specific steps of the processing flow are as follows:

Receive flow classification operation information and port number of packets sent by IPC;

Carry out flow classification operations on data packets in sequence;

Synchronously match the flow classification operation result of the data packet with the port of the data packet, and then transmit the flow classification operation result and the port number back to the IPC.

The processing flow of PMG is shown in Figure 5, and the specific steps are as follows:

After PMG receives a certain data slice, it sends the data slice to the corresponding data port through the port number of the data slice;

PMG identifies the type of data slice: 1. Start slice with SOP; 2. End slice with EOP; 3. Packet header slice with message modification information and descriptor information; 4. Ordinary data without the above information slice;

PMG receives the start slice with SOP, applies for a data storage space unit to cache the data packet, and reserves the first address of the data storage space unit for the first slice of the packet, and stores it from the second address The data of the initial slice; after the storage of the data slice is completed, record the address of the data stored in the port next time (the current address plus 1); put the first address of the data packet storage space in a FIFO (packet first slice address FIFO) In , it is provided for the first slice of the storage package to be used later;

PMG receives a normal data slice without any control information, and stores it in the DDR according to the address prompted by the port information of the data slice; after the storage is completed, record the address of the next data stored in the port (add 1 to the current address) ; If the current address space is full and the data packet has not been stored yet, apply for a new address space to continue storing; the applied address space is managed in the form of a linked list, and the address corresponding to the first node is used as the identifier of the entire packet to provide later use;

PMG receives the start slice with EOP, and records the data packet flag to indicate that the packet entity of the data packet has been stored and can receive the information of the next data packet; specifically, when PMG stores EOP, it will The first address of the storage space of the data packet is stored in a specific FIFO, and each port has a FIFO, which is used to record the first address of the corresponding storage space of all messages. When the PMG receives the packet header slice of a message of the port, it Take out an address from the corresponding FIFO, store the packet header slice into this address, this is the sign that the message cache is completed, and then complete the address information (as the message identifier) and the corresponding message descriptor information together After that, it is passed to the subsequent queue management module for the next operation.

PMG receives the packet header slice with message modification information and message descriptor information, and at the same time extracts the first address of the packet header slice address FIFO, and stores the packet header slice in this address, thus completing the data reorganization ; Simultaneously extract the information of the corresponding data packet (the first address stored in the data packet and the information of the data packet descriptor, etc.), and initiate the next operation.

In this embodiment, since the size of the data slice is the minimum unit of system operation, each data slice occupies one data storage space unit. When the size of the data slice is an integer multiple of the minimum unit of system operation, how much does each data slice occupy A data storage space unit is called a storage block. At this time, the first storage block is reserved for storing the first slice of the packet, and the data slice of the data packet entity is stored from the second storage block. If the size of the message modification information exceeds the storage space of one storage block, the first multiple storage blocks of the storage space corresponding to the data packet are reserved for storing the message modification information.

Since the minimum unit of operation is a data slice in units of 64bytes throughout the process, each port can easily distinguish the port information attached to each data slice as long as each port polls and transmits in units of data slices; the data in the port The division of packets is mainly based on the sequence of data packet control information and data packets themselves; in PMG, the write address of each port needs to be managed separately. According to the control information, operate according to the storage operation process mentioned above. The reassembly of the data packet can be completed in the DDR space.

The present invention also provides a data packet receiving and processing method, including:

Receive data packets, in the process of receiving data packets, extract the flow classification operation information, slice the data packets according to the preset size to obtain data slices, and store each received data slice into the storage space corresponding to the data packet;

After the data packet is received, perform a flow classification operation on the data packet according to the flow classification operation information, and after obtaining a flow classification operation result, extract relevant information to initiate a follow-up operation.

Wherein, the preset size is the smallest unit or a multiple of the smallest unit for system cache operation.

Wherein, when processing data packets of multiple input ports, the port identification information is used to distinguish data slices of data packets of different input ports, flow classification operation information and flow classification operation results.

Wherein, the flow classification operation result includes message modification information and message descriptor information; the extracting related information includes extracting the message descriptor information of the data packet; the method further includes, The message modification information is stored in the storage space corresponding to the data packet.

The method further includes: during the process of receiving the data packet, saving the first N bytes of data of the data packet, and starting to slice after N bytes of the data packet, where N is the preset size and The message modifies the information difference; when storing the data slice of the data packet, retaining the first storage block of the storage space corresponding to the data packet; after obtaining the flow classification operation result of the data packet, storing The first N bytes of data and the message modification information of the data packet form a packet header slice stored in the first storage block of the storage space corresponding to the data packet.

The method further includes, after receiving a data packet, performing receiving, slicing and data slice storage operations of the next data packet while waiting for and receiving the flow classification operation result of the data packet.

The data packet receiving and processing method provided by the present invention allows the data packet storage operation and the data packet receiving operation to start at the same time, and the operation of storing the data packet in the cache does not need to wait until the receiving operation and the flow classification operation are completed, thereby saving data packet storage The operation time reduces the processing delay and resource consumption of the system.

In addition, the present invention can also reduce the RAM resource overhead of the on-chip RAM of the device in the process of receiving and storing data packets. Since the flow classification operation only needs the header of the data packet (that is, the fixed-length data in the data packet) and the result of the CRC check when the data packet is received, the required cache is not very large. The biggest consumption of the on-chip RAM resource of the device is in the flow of data packet reception-buffering. If the existing solution is used to initiate the cache operation after the data packet is received, the system needs at least two long-packet on-chip RAM resources to complete this process. When adopting the scheme of the present invention, receiving-caching is carried out at the same time, that is, as long as the data packet receives enough data for one cache operation, the cache processing of the data packet is carried out, and the device on-chip RAM resource of this process is reduced to two cache operations. The size of the data.

In addition, the present invention stores the part of the flow classification result that does not require queue management—the message modification information—in the external space together with the message entity. Although the burden on the external space is increased, the storage burden on the queue management is reduced. Relatively speaking, the capability of the message management module (PMG) is stronger than that of the queue management module, and the cost is relatively low.

Claims (10)

1. a receives data packets processing unit, is characterized in that, comprises input control module, message processing module (MPM) and bag administration module, wherein:

Described input control module is used for: receive packet, in the process receiving packet, extract traffic classification operation information, by default size, packet is cut into slices, often receive a data slicer and send it to described bag administration module, after receives data packets, described traffic classification operation information is sent to described message processing module (MPM); And, receive the traffic classification operating result of described packet, described traffic classification operating result is sent to bag administration module;

Described message processing module (MPM) is used for: carry out traffic classification operation according to described traffic classification operation information to described packet, and the traffic classification operating result obtained is sent to described input control module;

Described bag administration module is used for: when receiving the data slicer of the described packet that described input control module sends, be stored into the memory space that this packet is corresponding; And, after receiving the traffic classification operating result of described packet, extract relevant information and initiate subsequent operation;

Described traffic classification operating result comprises message amendment information and message descriptor information;

Described bag administration module also for: by the message of described packet amendment information be stored in memory space corresponding to described packet, described extraction relevant information comprises the message descriptor information extracting described packet.

2. device as claimed in claim 1, is characterized in that, described default size is the minimum unit of system cache operation or the multiple of minimum unit.

3. device as claimed in claim 1, is characterized in that,

Described input control module also for: described data slicer is sent to described bag administration module, with, described traffic classification operation information is sent to described message processing module (MPM), with, when described traffic classification operating result being sent to bag administration module, carry the port identification information of the input port of described packet;

Described message processing module (MPM) also for, when receiving described traffic classification operation information and port identification information, preserve described port identification information, when sending described traffic classification operating result to described input control module, carry described port identification information;

Described bag administration module also for: receive described data slicer and port identification information, or when receiving described traffic classification operating result and port identification information, according to described port identification information, described data slicer or traffic classification operating result are sent to corresponding FPDP process.

4. device as claimed in claim 1, is characterized in that,

Described input control module also for: in the process receiving packet, preserve N byte data before described packet, cut into slices from after N number of byte of described packet, described N is that described default size and described message revise information difference; After receiving the traffic classification operating result of described packet, information of being revised by the message of described front N byte data and described packet forms wraps first section and sends to described bag administration module;

Described bag administration module also for: when being stored by the data slicer of described packet, retain the first memory block of memory space corresponding to described packet; And, when receiving the bag head section of described packet, be stored in the first memory block of memory space corresponding to described packet.

5. the device as described in as arbitrary in Claims 1-4, is characterized in that,

Described input control module also for: after receiving a packet, waits for and receive in the traffic classification operating result process of described packet, carrying out the reception operation of next packet, and the data slicer sending next packet described is to bag administration module;

Described bag administration module also for: receiving the data slicer of described packet, wait in the process of traffic classification operating result of described packet, when receiving the data slicer of next packet of described packet, be stored to the memory space that next packet described is corresponding.

6. a receives data packets processing method, is characterized in that, comprising:

Receive packet, in the process receiving packet, extract traffic classification operation information, by default size, section is carried out to packet and obtain data slicer, often receive a data slicer and be then stored into memory space corresponding to this packet;

After described receives data packets, according to described traffic classification operation information, traffic classification operation is carried out to described packet, after obtaining traffic classification operating result, extract relevant information and initiate subsequent operation;

Described traffic classification operating result comprises message amendment information and message descriptor information; Described extraction relevant information comprises the message descriptor information extracting described packet;

Described method also comprises, and the message of described packet amendment information is stored in memory space corresponding to described packet.

7. method as claimed in claim 6, is characterized in that, described default size is the minimum unit of system cache operation or the multiple of minimum unit.

8. method as claimed in claim 6, it is characterized in that, described method also comprises:

When processing the packet of multiple input port, port identification information is used to distinguish the data slicer of the packet of different input port, traffic classification operation information and traffic classification operating result.

9. method as claimed in claim 6, is characterized in that,

Described method also comprises: in the process receiving packet, preserves N byte data before described packet, and cut into slices from after N number of byte of described packet, described N is that described default size and described message revise information difference; When being stored by the data slicer of described packet, retain the first memory block of memory space corresponding to described packet;

After obtaining the traffic classification operating result of described packet, information of described front N byte data and the message of described packet being revised forms wraps first section and is stored in the first memory block of memory space corresponding to described packet.

10. the method as described in as arbitrary in claim 6 to 9, is characterized in that,

Described method also comprises, and after receiving a packet, waits for and receives in the traffic classification operating result process of described packet, carrying out the reception of next packet, section and data slicer storage operation.