CN102480430B - Method and device for realizing message order preservation - Google Patents

Abstract

The invention provides a method and a device for realizing message order preservation. The method includes: A, the queue area control unit on the network communication device stores the received message into the input queue of the shared queue area, and provides the input queue of the shared queue area to the network communication device that first performs forwarding processing The first forwarding core; the first forwarding core polls the input queue of the shared queue area, and sequentially acquires and processes each shared queue area in the input queue of the shared queue area; B, the first forwarding core finishes processing the current shared queue area After receiving all the messages in the corresponding message queue, judge whether it is the last forwarding core that performs forwarding processing, if not, provide the processed shared queue area to the next forwarding core that performs forwarding processing, and perform step C, if Yes, provide the processed shared queue area to the order-preserving processing unit on the network communication device, and execute step D; C, the next forwarding core that performs forwarding processing executes the first forwarding core in step B. D, the order-preserving processing unit obtains the messages in the shared queue area, and serializes them according to the receiving order for sending the messages. By adopting the invention, problems such as caching, timing aging and the like can be avoided while maintaining the sequence of messages.

Description

Method and device for realizing packet order preservation

technical field

The invention relates to data communication technology, in particular to a method and device for realizing message order preservation.

Background technique

As the rate of the physical interface continues to increase, the number of forwarding cores (cores) of the CPU in the network communication device also increases accordingly. At present, the parallel working mode is usually used to make each forwarding core in the CPU process the packet flow in parallel in order to improve the forwarding performance, as shown in FIG. 1 .

Although the parallel working method makes full use of the multi-core parallel forwarding capability and improves the forwarding performance in most cases, for a certain forwarding core, when a message flow (which essentially has the same protocol keyword When the traffic of the packet set) is greater than the current load of the forwarding core, in order to prevent the packets in the packet flow from being discarded, some packets in the packet flow need to be distributed to other forwarding cores for processing. The case where packets of the same packet flow are processed by different forwarding cores. In this way, if the processing time of different forwarding cores does not match, the messages in the same message flow will not be sent in the order of reception, and the data communication system requires that the order of sending messages must be consistent with the order of receiving them. Based on this , the network communication device needs to perform an order-preserving operation on the messages in the same message stream before the messages are sent.

In the prior art, once packet order preservation is involved, issues such as caching and timing aging will be involved accordingly. Since the packet will be distributed to different forwarding cores only when the traffic of a packet flow is greater than the current load of a certain forwarding core, it is impossible to know in advance that the packets in the same packet flow are between different forwarding cores. distribution, and processing delays on different forwarding cores, therefore, the cache capacity required for message order preservation cannot be determined. If the cache capacity is too large, the packets in the packet flow will be discarded prematurely, and if it is too large, it will lead to excessive consumption of cache resources, which will cause system reactions (for example, the receiving end cannot obtain the packet cache). As for the timing aging problem, it is also impossible to determine the required aging time due to the fact that the distribution of messages in the same message flow among different forwarding cores and the processing delay on different forwarding cores cannot be known in advance. If the aging time is set too small, it will cause premature discarding, and if it is too large, it will increase the cache burden.

To sum up, it can be seen that a packet order preservation method that can avoid problems such as caching and timing aging is a technical problem that needs to be solved urgently.

Contents of the invention

The invention provides a method and a device for realizing message order preservation, so as to avoid problems such as buffering and timing aging when the message order is preserved.

A method for realizing message order preservation, comprising:

A. The queue area control unit on the network communication device sequentially stores the received messages into the input queue of the shared queue area, and provides the input queue of the shared queue area to the first forwarding core that performs forwarding processing first, and the first forwarding core The core polls the input queue of the shared queue area, and obtains and processes each shared queue area in the input queue of the shared queue area in turn;

B. After the first forwarding core has processed all the messages corresponding to the message queue in the current shared queue area, it judges whether it is the last forwarding core that performs forwarding processing, and if not, provides the processed shared queue area to the next A forwarding core processing that performs forwarding processing, and executes step C, if so, provides the processed shared queue area to the order-preserving processing unit on the network communication device, and executes step D;

C, the next forwarding core performing forwarding processing performs the operation performed by the first forwarding core in step B;

D. The order-preserving processing unit obtains the messages in the shared queue area, and serializes them according to the received order for sending the messages.

A device for realizing message order preservation, including at least one forwarding core participating in forwarding processing, a message receiving queue and a message sending queue, the key point of which is that the device also includes: a queue area control unit and an order preserving processing unit ;in,

The queue area control unit is used to store the received message into the input queue of the shared queue area, and provide the input queue of the shared queue area to the first forwarding core on the device that first performs forwarding processing;

The first forwarding core polls the input queue of the shared queue area, sequentially obtains and processes the shared queue areas in the input queue of the shared queue area, and after each the first forwarding core processes all the messages of the corresponding message queue in the shared queue area , judging whether it is the last forwarding core that performs forwarding processing, if not, providing the processed shared queue area to the next forwarding core that performs forwarding processing, and the next forwarding core that performs forwarding processing executes the first forwarding core The operation performed, if yes, providing the processed shared queue area to the order-preserving processing unit;

The order-preserving processing unit is used to obtain the shared queue area, and serialize the obtained messages in the shared queue area according to the receiving order, so as to send the processed messages.

As can be seen from the above technical solutions, in the present invention, the first forwarding core that first performs the forwarding process first processes the shared queue area in the input queue of the shared queue area, wherein the shared queue area includes each forwarding core that participates in the forwarding process. Each forwarding core provides the processed shared queue area to the next adjacent forwarding core that performs forwarding processing, and so on, until the last forwarding core that performs forwarding processing provides the processed shared queue area Provided to the order-preserving processing unit, so that the order-preserving processing unit serializes the obtained messages in the shared queue area according to the receiving order, and realizes the purpose of sending messages according to the order in which the messages are received, that is, realizes message preservation sequence. The message sequence protection provided by the present invention, for the shared queue area in the input queue of the shared queue area, is processed by each forwarding core in sequence, for the same shared queue area, serial processing is performed between the forwarding cores, and finally each processed shared The queue area is in accordance with the order of entering the input queue of the shared queue area, and after entering the output queue of the shared queue area, the sequence processing is performed, so problems such as cache and timing aging are not involved.

Description of drawings

Fig. 1 is a schematic diagram of a parallel working mode in the prior art;

Fig. 2 is the basic flowchart provided by the embodiment of the present invention;

FIG. 3 is a detailed flowchart provided by an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of a shared queue area provided by an embodiment of the present invention;

FIG. 5 is an implementation flowchart of step 312 provided by the embodiment of the present invention;

FIG. 6 is a schematic diagram corresponding to the process in FIG. 3 provided by an embodiment of the present invention;

FIG. 7 is a structural diagram of a device provided by an embodiment of the present invention;

FIG. 8 is a schematic diagram of an access channel between two adjacent forwarding cores according to the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

The method provided by the present invention is mainly to avoid problems such as buffering and timing aging involved in packet order preservation in the prior art. FIG. 2 describes the method provided by the embodiment of the present invention.

Referring to FIG. 2, FIG. 2 is a basic flowchart provided by an embodiment of the present invention. This process is mainly applied to network communication devices, where the network communication device may be a router or other devices with routing functions, such as switches, etc., which are not specifically limited in the embodiments of the present invention. Based on this, as shown in Figure 2, the process may include the following steps:

Step 201, the queue area control unit on the network communication device sequentially stores the received messages into the input queue of the shared queue area, and provides the input queue of the shared queue area to the first forwarder on the network communication device that performs forwarding processing first core, the first forwarding core polls the input queue of the shared queue area, sequentially obtains and processes each shared queue area in the input queue of the shared queue area.

In this step 201, the operation of storing the received message in the input queue of the shared queue area by the queue area control unit may refer to the description in steps 302 to 305 in FIG. 3 for details.

Step 202, after the first forwarding core processes all the messages corresponding to the message queue in the current shared queue area, it judges whether it is the last forwarding core to perform forwarding processing, and if not, provides the processed shared queue area to the next The forwarding core that executes forwarding processing, and executes step 203 , if yes, provides the processed shared queue area to the order-preserving processing unit on the network communication device, and executes step 204 .

For details of the operations performed in step 202, refer to the operations in steps 306 to 309 in FIG. 3 .

In addition, in this embodiment, the case where the first forwarding core is the last forwarding core performing forwarding processing is: the network communication device only includes one forwarding core for forwarding processing.

In step 203, the next forwarding core to perform forwarding processing executes the operation performed by the first forwarding core in step 202.

It can be seen from step 202 to step 203 that the first forwarding core processes the shared queue area first, and provides the processed shared queue area to the next forwarding core for processing, and so on until the forwarding core that performs forwarding processing finally provides the processing The shared queue area of is given to the order preservation processing unit, so that the order preservation processing unit executes step 204.

In step 204, the order-preserving processing unit obtains the messages in the shared queue area, and serializes them according to the received order for sending the messages.

The description of this step 204 may refer to the description of steps 311 to 312 in FIG. 3 .

The methods provided by the embodiments of the present invention are briefly described above.

In the above process, all forwarding cores participating in forwarding on the network communication device may be numbered in sequence in advance. Based on this, the first forwarding core in this embodiment may be the forwarding core with the smallest number, and the last forwarding core that performs forwarding processing is correspondingly the forwarding core with the largest number; or, the first forwarding core is the forwarding core with the largest number, and the last A forwarding core that performs forwarding processing corresponds to the forwarding core with the smallest number. The following assumes that the first forwarding core is the forwarding core with the smallest number, and the last forwarding core performing forwarding processing is the forwarding core with the highest number as an example. The principle of other cases is similar. The process shown in FIG. 2 is described in detail through FIG. 3 .

Referring to FIG. 3, FIG. 3 is a detailed flowchart provided by an embodiment of the present invention. In this embodiment, all the forwarding cores participating in the forwarding on the network communication device have completely the same function, and they are all used for message forwarding processing. If all forwarding cores participating in forwarding on the network communication device are numbered in sequence, they are core-1, core-2, ... core-N, wherein the first forwarding core that performs forwarding processing first is the smallest numbered The forwarding core is core-1, and the forwarding core that finally executes forwarding processing is the forwarding core with the largest number, namely core-N.

Based on this, as shown in Figure 3, the process may include the following steps:

In step 301, the message receiver on the network communication device stores the received messages in the message receiving queue according to FIFO (First In First Out) order.

Step 302, when the set time arrives, the queue area control unit on the network communication device polls the message receiving queue, if the message receiving queue is empty, then end the current process; otherwise, execute step 303.

In this step 302, the queue area control unit performs the operation of polling whether the message receiving queue is empty in real time or periodically at fixed time intervals.

Step 303, the queue area control unit obtains all messages from the message receiving queue.

Step 304, the queue area control unit creates a shared queue area, and starts from the message queue corresponding to core-1 in the shared queue area, and puts messages into the message queues corresponding to each forwarding core in sequence.

In this embodiment, the shared queue area includes a message queue corresponding to each forwarding core participating in the forwarding process, wherein the message queue corresponding to each forwarding core is stored by the head of the queue link list created for the forwarding core, and the storage method for FIFO. In this way, when this step 304 is executed, each forwarding core participating in the forwarding process on the network communication device, that is, core-1, core-2, ..., core-N, has a message queue for storing in its corresponding head of the queue list.

The specific implementation of the shared queue area in this embodiment may be: a message chain block composed of message queues stored in parallel in order starting from the message queue corresponding to core-1 is determined as the shared queue area. See Figure 4 for details.

In addition, in this step 304, the queue area control unit can put messages into the message queues corresponding to each forwarding core according to the principle of equal distribution of messages by each forwarding core, specifically: the queue area control unit determines that it needs to put each message The number of messages in the message queue is M; after that, all the obtained messages are divided in order in units of M, and N messages are put into the message queues corresponding to each forwarding core in sequence starting from the message queue corresponding to core-1. arts. For example, there are 3 forwarding cores in the network communication device, namely core-1 to core-3, if the number of all messages obtained by the queue area control unit from the message receiving queue is 12, the order is message 1 to message 12, In this way, it is necessary to store 4 messages in the message queue corresponding to each forwarding core, that is, the message queue corresponding to core-1 stores message 1 to message 4, and the message queue corresponding to core-2 stores message 5 Up to packet 8, the packet queue corresponding to core-3 stores packets 9 to 12.

Certainly, as another implementation of the embodiment of the present invention, in this step 304, the queue area control unit may also transfer the information to the message queue corresponding to each forwarding core according to the current load of each forwarding core and the quantity of all messages obtained. Putting in the message, specifically: step 1, the queue area control unit takes core-1 as the current core. Step 2, the queue area control unit determines the number X of messages to be put into the message queue corresponding to the current core according to the current load of the current core and the number of all received messages. Step 3, the queue area control unit puts X messages into the message queue corresponding to the current core in order; then sets the next forwarding core of core-1 as the current core, and returns to step 2. There are still 3 forwarding cores in the network communication equipment, namely core-1 to core-3, and the number of all messages obtained by the queue area control unit from the message receiving queue is 12, that is, message 1 to message 12 in sequence. For example, if the current load of core-1 is relatively serious and can only carry 2 packets at present, the current load of core-2 can only carry 4 packets, and the current load of core-3 is relatively light and can carry at least 6 packets, then Based on the descriptions of steps 1 to 3 above, packets 1 to 2 can be stored in the packet queue corresponding to core-1, and packets 3 to 6 can be stored in the packet queue corresponding to core-2. Packets 7 to 12 are stored in the packet queue corresponding to core-3. Of course, if the current load of core-2 is relatively light and can carry at least 10 packets, you can directly store packets 3 to 12 in the packet queue corresponding to core-2, and the packet queue corresponding to core-3 The message is not stored in the file, and the specific situation is analyzed in detail, and the embodiment of the present invention does not specifically limit it.

Step 305, the queue area control unit sequentially puts the shared queue area storing messages into the input queue of the shared queue area.

The input queue of the shared queue area in step 305 is essentially a message transmission channel between the queue area control unit and core-1, and is used to store the messages in the shared queue area in the form of FIFO.

It should be noted that, in this embodiment, before step 305 is executed, a corresponding message queue needs to be created in each shared queue area for each forwarding core performing forwarding processing.

Step 306, core-1 polls the input queue of the shared queue area, and obtains each shared queue area in the input queue of the shared queue area in turn.

Step 307, core-1 judges whether the message queue corresponding to itself in the current shared queue area is empty for each shared queue area obtained (recorded as the current shared queue area), if yes, execute step 308; otherwise, process itself For all the messages in the corresponding message queue, step 309 is then executed.

If in step 304, the queue area control unit puts messages into the message queues corresponding to each forwarding core according to the current load of each forwarding core and the quantity of all messages obtained, and, when forming the current shared queue area, core- The load of 1 is relatively serious, and it can no longer carry message processing. Based on this, when this step 307 is executed, the message queue corresponding to core-1 will be empty in the current shared queue area. Considering the embodiment of the present invention Extensiveness, it is necessary to perform the judgment operation in step 307.

In this step 307, processing the message in the message queue can be performed according to the processing operation required by the message, which is not limited here.

Step 308 , core-1 judges whether it is the forwarding core with the largest number, if yes, stores the current shared queue area in the created shared queue area output queue, and executes step 311 , otherwise, executes step 309 .

In step 308, core-1 judges whether it is the forwarding core with the largest number because it determines whether core-1 is the last forwarding core to perform forwarding processing. The case where core-1 is the last forwarding core performing forwarding processing is: core-1 is the only forwarding core of the network communication device.

The output queue in the shared queue area in step 308 is a message transmission channel between the last forwarding core and the order-preserving processing unit, and it is stored in the shared queue area stored therein in a FIFO manner. Since the messages in a shared queue area have been processed after passing through the last forwarding core that performs forwarding processing, after enqueuing the processed shared queue area to the output queue of the shared queue area, the order-preserving processing unit can send The sequence of the processed packets is preserved.

Step 309, core-1 stores the processed shared queue area in the inter-core access channel corresponding to the next forwarding core, and the inter-core access channel can be organized into a shared queue area input queue.

In this embodiment, the inter-core access channel corresponding to the next forwarding core is the message transmission channel between the previous forwarding core adjacent to the next forwarding core and the next forwarding core, which can be stored in the FIFO mode. The shared queue area processed by the preceding forwarding core in an adjacent sequence can therefore be organized into an input queue of the shared queue area, as shown in FIG. 8 for details. Taking the next forwarding core in step 309 as core-2 as an example, then step 309 is specifically: core-1 stores the processed shared queue area in the inter-core access channel corresponding to core-2, wherein core- The inter-core access channel corresponding to 2 is the message transmission channel between core-1 and core-2, which can be stored in the shared queue area processed by core-1 in the form of FIFO, and can be organized into the input queue of the shared queue area.

In step 310, the next forwarding core performs operations similar to those performed by core-1 in steps 306 to 309.

It should be noted that since the number of messages in the message queues corresponding to each forwarding core in each shared queue area will be different, not every core processes the shared queue area in the same order. Therefore, preferably, in the next Before the forwarding core executes the operation performed by core-1 in step 306, the next forwarding core can judge in real time whether the input queue of the shared queue area corresponding to itself is empty, if yes, continue to return to the judgment operation, otherwise, execute the core -1 The operation to perform.

Step 311 , the order preserving processing unit polls the output queue of the shared queue area, and obtains the shared queue area in sequence.

It should be noted that since the operations performed by the order-preserving processing unit and the last forwarding core performing forwarding processing are not in a fixed chronological order, preferably, before step 311 is executed, the order-preserving processing unit can determine the shared queue in real time Whether the area output queue is empty, if yes, continue to return to the judgment operation, otherwise, execute step 311.

Step 312, for each acquired shared queue area, the order-preserving processing unit sequentially accesses the message queue corresponding to each forwarding core from the acquired shared queue area, and sequentially accesses the message queue in the accessed message queue The messages are sequentially stored in the message sending queue of the network communication device.

This step 312 may be implemented in various forms, and FIG. 5 shows one of the forms.

Referring to FIG. 5 , FIG. 5 is a flowchart for implementing step 312 provided by the embodiment of the present invention. As shown in Figure 5, the process may include the following steps:

Step 501, create a message chain L whose initial value is empty.

That is, when step 501 is executed, the message chain L does not store any message.

Step 502, the order preserving processing unit accesses the message queue Q corresponding to the core-S from the acquired shared queue area.

Since the forwarding cores are numbered starting from 1 in this embodiment, S in the initial stage is 1.

In this embodiment, since the message queue corresponding to each forwarding core is stored by the queue link list header of the forwarding core, this step 502 can access the message corresponding to core-S according to the queue link list header of core-s Queue Q.

In step 503, the sequence-preserving processing unit judges whether the accessed message queue Q is empty, and if so, executes step 504; otherwise, puts the messages in the message queue Q into the message chain L in sequence.

Step 504 , the order preserving processing unit judges whether core-S is the forwarding core with the largest number, if not, execute step 505 , and if yes, execute step 506 .

Step 505, set S=S+1, return to step 502.

Step 506, put the message chain L into the message sending queue, and the message sender on the network communication device sends the message chain L in the message sending queue.

So far, the operation of step 312 has been realized through the process shown in FIG. 5 .

Through the description of steps 301 to 312, it can be seen that in this embodiment, the forwarding core on the network communication device does not actually process the messages in parallel, but the core-1 processes the messages in the shared queue area first, And provide the processed shared queue area to core-2 for processing, and so on, until core-N, which is the last forwarding core that performs forwarding processing, provides the processed shared queue area to the order-preserving processing unit, so that the order-preserving processing unit The obtained messages in the shared queue area are serialized according to the receiving order, that is, the purpose of sending the messages according to the receiving order of the messages is realized. At the same time, each forwarding core is processing a different shared queue area, so in a sense, each core is working in parallel.

In order to make the flow shown in FIG. 3 clearer, the embodiment of the present invention provides a specific schematic diagram corresponding to the flow shown in FIG. 3 , as shown in FIG. 6 .

The method provided by the embodiment of the present invention is described above, and the device provided by the embodiment of the present invention is described below.

Referring to FIG. 7, FIG. 7 is a structural diagram of a device provided by an embodiment of the present invention. Wherein, the device may be implemented as a router or other devices with a routing function, such as a switch, which are not specifically limited in this embodiment of the present invention. As shown in Figure 7, the device may include: at least one forwarding core 701 participating in forwarding processing, a message receiving queue 702 and a message sending queue 703, and the key point is that the device also includes: a queue area control unit 704, an order preserving The processing unit 705 and the shared queue area input the queue 706 .

Wherein, the queue area control unit 704 is configured to store the received message into the input queue 706 of the shared queue area, and provide the input queue 706 of the shared queue area to the first forwarding core on the device that first performs forwarding processing;

The first forwarding core polls the input queue 706 of the shared queue area, sequentially acquires the shared queue areas in the input queue of the shared queue area and processes them, and after processing all the messages of the corresponding message queues in a shared queue area, judges whether it is For the last forwarding core that performs forwarding processing, if not, provide the processed shared queue area to the next forwarding core that performs forwarding processing, and the next forwarding core that performs forwarding processing performs the operation performed by the first forwarding core, If yes, provide the processed shared queue area to the order preserving processing unit 705 . In this embodiment, when the first forwarding core processes the shared queue area, it first judges whether the message queue corresponding to itself in the obtained current shared queue area is empty, and if not, executes the process of processing the current shared queue area. If it is empty, directly judge whether it is the last forwarding core that performs forwarding processing, and if so, provide the current shared queue area to the order-keeping processing unit 705, otherwise provide the current shared queue area to the next forwarding core that performs forwarding processing , the operation performed by the first forwarding core is performed by the next forwarding core performing forwarding processing.

The order-preserving processing unit 705 is configured to obtain the shared queue area, and serialize the obtained messages in the shared queue area according to the receiving order, so as to send the processed messages.

In this embodiment, all forwarding cores participating in forwarding processing on the device are numbered in sequence; based on this, the first forwarding core is the forwarding core with the smallest number, and the last forwarding core that performs forwarding processing is the numbered The largest forwarding core; or, the first forwarding core is the forwarding core with the largest number, and the last forwarding core performing forwarding processing is the forwarding core with the smallest number.

In this embodiment, if there is at least one forwarding core participating in forwarding processing on the device, as shown in FIG. 7 , the device further includes: at least one inter-core access channel 707 .

Taking the forwarding cores participating in the forwarding processing on the device as the first forwarding core, the second forwarding core, the third forwarding core...the Nth forwarding core as an example, the shared queue area input queue 706 and the inter-core The relationship between access channels is shown in Figure 8. Wherein, each inter-core access channel is used to store the shared queue area processed by the previous adjacent forwarding core in sequence, therefore, it can also be organized into a shared queue area input queue accordingly.

Based on this, when the first forwarding core determines that it is not the last forwarding core to perform forwarding processing, it provides the processed shared queue area to the next forwarding core that performs forwarding processing, such as the inter-core access channel corresponding to the second forwarding core , the next forwarding core that performs forwarding processing performs operations similar to those performed by the first forwarding core.

Preferably, as shown in FIG. 7, the queue area control unit 704 may include:

The obtaining subunit 7041 is configured to poll the message receiving queue, and once it is found that there are messages in the message receiving queue, obtain all messages from the message receiving queue;

Create a subunit 7042, configured to create a shared queue area, the shared queue area includes a message queue corresponding to each forwarding core participating in the forwarding process;

The processing subunit 7043 is configured to start from the message queue corresponding to the first forwarding core in the shared queue area, and put messages into the message queues corresponding to each forwarding core in sequence;

A subunit 7044 is constructed, configured to put the shared queue area storing messages into the input queue 706 of the shared queue area in sequence.

Wherein, when the processing subunit 7043 puts a message into the message queue corresponding to each forwarding core, it can put a message into the message queue corresponding to each forwarding core according to the principle of equal distribution of messages by each forwarding core; or, according to each The current load of the forwarding core and the quantity of all received messages are put into the message queue corresponding to each forwarding core.

Preferably, as shown in FIG. 7 , the device further includes: an output queue 708 in a shared queue area.

Wherein, the shared queue area output queue 708 is a message transmission channel between the last forwarding core that performs forwarding processing and the order preserving processing unit 705, and is used to store the shared queue area processed by the last forwarding core that performs forwarding processing. The order-preserving processing unit 705 performs order-preserving processing on the processed messages in the shared queue area. Based on this, the last forwarding core that performs forwarding processing provides the processed shared queue area to the order preservation processing unit, including: providing the processed shared queue area to the shared queue area output queue, and the order preservation processing unit 705 polls The output queue in the shared queue area obtains the processed messages in the shared queue area.

In the embodiment of the present invention, as shown in FIG. 7 , the sequence preserving processing unit 705 may include:

The polling subunit 7051 is used to poll the shared queue area output queue 708, and hand over the shared queue area in the shared queue area output queue to the message processing subunit 7052;

The message processing subunit 7052 is used to sequentially access the message queues corresponding to each forwarding core from the obtained shared queue area, and store the messages in the accessed message queues into the message queues in sequence in the sending queue.

So far, the device provided by the embodiment of the present invention has been described.

As can be seen from the above technical solutions, in the present invention, the first forwarding core that first performs the forwarding process first processes the shared queue area in the input queue of the shared queue area, wherein the shared queue area includes each forwarding core that participates in the forwarding process. Each forwarding core provides the processed shared queue area to the next adjacent forwarding core that performs forwarding processing, and so on, until the last forwarding core that performs forwarding processing provides the processed shared queue area Provided to the order-preserving processing unit, so that the order-preserving processing unit serializes the obtained messages in the shared queue area according to the receiving order, and realizes the purpose of sending messages according to the order in which the messages are received, that is, realizes message preservation sequence. The message sequence protection provided by the present invention, for the shared queue area in the input queue of the shared queue area, is processed by each forwarding core in sequence, for the same shared queue area, serial processing is performed between the forwarding cores, and finally each processed shared The queue area is in accordance with the order of entering the input queue of the shared queue area, and after entering the output queue of the shared queue area, the sequence processing is performed, so problems such as cache and timing aging are not involved.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the present invention. within the scope of protection.

Claims (10)

1. A method for realizing message order preservation is characterized in that the method comprises the following steps:

a, a queue area control unit on the network communication equipment sequentially stores received messages into a shared queue area input queue, and provides the shared queue area input queue to a first forwarding core which firstly executes forwarding processing, and the first forwarding core polls the shared queue area input queue, and sequentially acquires and processes each shared queue area in the shared queue area input queue;

b, after the first forwarding core finishes processing all the messages of the corresponding message queue in the current shared queue area, judging whether the first forwarding core is the last forwarding core for executing forwarding processing, if not, providing the processed shared queue area for the next forwarding core for executing forwarding processing, and executing the step C, if so, providing the processed shared queue area for an order-preserving processing unit on the network communication equipment, and executing the step D;

c, the next forwarding core executing the forwarding processing executes the operation executed by the first forwarding core in the step B;

d, the order-preserving processing unit acquires the messages in the shared queue area and carries out serialization processing according to the receiving sequence for sending the messages;

in step a, the step of sequentially storing the received messages into the input queue of the shared queue area by the queue area control unit includes:

a1, when the queue area control unit polls the message receiving queue each time, obtaining all messages from the message receiving queue, where the message receiving queue is used to store the messages received by the network communication device;

a2, the queue area control unit creates a shared queue area, the shared queue area includes a message queue corresponding to each forwarding core participating in forwarding processing, and messages are sequentially put into the message queues corresponding to the forwarding cores in sequence from the message queue corresponding to the first forwarding core in the shared queue area;

a3, the queue area control unit puts the shared queue areas storing messages into the shared queue area input queue in sequence.

2. The method of claim 1, further comprising, prior to execution of the method: numbering all forwarding cores participating in forwarding processing on the network communication equipment according to a sequence;

the first forwarding core is the forwarding core with the smallest number, and the last forwarding core executing forwarding processing is the forwarding core with the largest number; or,

the first forwarding core is the forwarding core with the largest number, and the last forwarding core executing the forwarding processing is the forwarding core with the smallest number.

3. The method according to claim 1, wherein in step a2, the queue area control unit puts packets into the packet queues corresponding to the forwarding cores according to a principle that packets are evenly distributed by the forwarding cores; or,

and the queue area control unit puts the messages into the message queues corresponding to the forwarding cores according to the current loads of the forwarding cores and the obtained number of all the messages.

4. The method of claim 1, further comprising, prior to performing step B: establishing a shared queue area output queue, wherein the shared queue area output queue is used for storing a shared queue area processed by a last forwarding core executing forwarding processing so that an order-preserving processing unit can carry out order-preserving processing on the processed messages in the shared queue area;

in step B, providing the processed shared queue area to the order preserving processing unit includes: and providing the processed shared queue area to the output queue of the shared queue area, and acquiring the processed message in the shared queue area by the order-preserving processing unit through polling the output queue of the shared queue area.

5. The method of claim 4, wherein step D comprises:

d1, polling the output queue of the shared queue area by the order-preserving processing unit to sequentially obtain the shared queue area;

and D2, the order-preserving processing unit sequentially accesses the message queues corresponding to each forwarding core from the obtained shared queue area in sequence, and sequentially stores the messages in the accessed message queues into the message sending queues of the network communication equipment in sequence.

6. An apparatus for implementing packet order preservation, comprising at least one forwarding core participating in forwarding processing, a packet receiving queue and a packet sending queue, the apparatus further comprising: a queue area control unit and an order preserving processing unit; wherein,

the queue area control unit is used for storing the received message into a shared queue area input queue and providing the shared queue area input queue to a first forwarding core which firstly executes forwarding processing on the device;

the first forwarding core is used for polling the input queue of the shared queue area, sequentially acquiring and processing the shared queue area in the input queue of the shared queue area, judging whether the first forwarding core is the last forwarding core for executing forwarding processing after processing all messages of a corresponding message queue in one shared queue area, if not, providing the processed shared queue area to the next forwarding core for executing forwarding processing, executing the operation executed by the first forwarding core by the next forwarding core for executing forwarding processing, and if so, providing the processed shared queue area to the order-preserving processing unit;

the order-preserving processing unit is used for acquiring the shared queue area and serializing the acquired messages in the shared queue area according to the receiving sequence so as to send the processed messages;

wherein the queue area control unit includes:

the acquiring subunit is used for polling a message receiving queue, and acquiring all messages from the message receiving queue once finding that the messages exist in the message receiving queue;

a creating subunit, configured to create a shared queue area, where the shared queue area includes a packet queue corresponding to each forwarding core participating in forwarding processing;

the processing subunit is configured to sequentially place the messages into the message queues corresponding to the forwarding cores in the shared queue area in sequence from the message queue corresponding to the first forwarding core according to a principle that the forwarding cores evenly distribute the messages, or according to the current load of the forwarding cores and the number of all obtained messages;

and the construction subunit is used for sequentially placing the shared queue areas for storing the messages into the shared queue area input queue.

7. The apparatus of claim 6, wherein all forwarding cores on the apparatus that participate in forwarding processing are numbered in order;

the first forwarding core is the forwarding core with the smallest number, and the last forwarding core executing forwarding processing is the forwarding core with the largest number; or,

the first forwarding core is the forwarding core with the largest number, and the last forwarding core executing the forwarding processing is the forwarding core with the smallest number.

8. The apparatus of claim 6 or 7, further comprising:

the shared queue area output queue is used for storing the shared queue area processed by the last forwarding core executing the forwarding processing so that the order-preserving processing unit can carry out order-preserving processing on the processed messages in the shared queue area;

the last forwarding core executing the forwarding processing provides the processed shared queue area to the order-preserving processing unit, and the method comprises the following steps: and providing the processed shared queue area to the output queue of the shared queue area, and acquiring the processed message in the shared queue area by the order-preserving processing unit through polling the output queue of the shared queue area.

9. The apparatus of claim 8, wherein the order preserving processing unit comprises:

the polling subunit is used for polling the output queue of the shared queue area and delivering the shared queue area in the output queue of the shared queue area to the message processing subunit;

and the message processing subunit is used for sequentially accessing the message queues corresponding to the forwarding cores from the acquired shared queue area according to the sequence, and sequentially storing the messages in the accessed message queues into the message sending queue according to the sequence.

10. The apparatus of claim 9, further comprising: at least one inter-core access channel is used for storing the shared queue area processed by the last forwarding core adjacent in sequence and can be organized into an input queue of the shared queue area.