WO2016000627A1

WO2016000627A1 - Message processing

Info

Publication number: WO2016000627A1
Application number: PCT/CN2015/083142
Authority: WO
Inventors: Mengtao ZHOU; Zhenglin QI; Yihong Xiu
Original assignee: Hangzhou H3C Technologies Co., Ltd.
Priority date: 2014-07-02
Filing date: 2015-07-02
Publication date: 2016-01-07
Also published as: CN105337851A; CN105337851B

Abstract

Upon a forwarding chip of a PE device receiving a load message, the forwarding chip determines whether the load message is a target load message according to a preset message filtering rule, where the target load message is a message whose destination processing device includes the PE device: and the forwarding chip transmits file data included in the load message to a central processing unit of the PE device for loading, upon determining that the load message is the target load message.

Description

MESSAGE PROCESSING

Background

Stacking systems, such as but not limited to Intelligent Resilient Framework 3.0 (IRF3) , may include Control Bridge (CB) devices and Port Extender (PE) device. The PE devices (referred to as “PEs” hereinafter) are connected with the CB devices (referred to as “CBs” hereinafter) and the PEs provide the CB with a port extension function. For example, the CBs may be connected together and virtualized as a single device. The PEs may be virtualized and managed as if they were line cards of the CB devices. The PEs may be initialized by loading version files required for starting the PEs, where the version files are not stored in the PEs and may be transmitted by the CBs to the PEs.

Brief Description of the Drawings

Fig. 1 is a schematic diagram illustrating a system architecture of CBs and PEs in accordance with examples of the present disclosure；

Fig. 2 is a schematic diagram illustrating the structure of PE in accordance with examples of the present disclosure；

Fig. 3 is a schematic flow chart illustrating a message processing method in accordance with examples of the present disclosure；

Fig. 4 is a schematic diagram illustrating an application scenario of a message processing method in accordance with examples of the present disclosure； and

Fig. 5 is a schematic diagram illustrating the message transmission timing of a CB transmitting message in accordance with examples of the present disclosure.

Detailed Description of the Embodiments

Fig. 1 illustrates the architecture of an application system which can include at least one PE and at least one CB, wherein the PE (s) is connected with the CB (s) as an extension port of the CB (s) . The CB can provide the respective PEs connected therewith with files to be loaded by the PEs, including version files required for starting the PEs or patch files or upgrade files, for example. All of these files can be transmitted by the CB to the PEs in “load message” , that is, the CB can transmit to the PEs a load message which carries file data of some file (s) to be loaded by the PEs.

The CB can transmit the load message to the PEs in a Multicast mode. For example, as illustrated in Fig. 1, in a case that a Multicast group includes three PEs PE1, PE2, and PE3, a CB CB1 connected with the three PEs can transmit a load message to the Multicast group, even if the load message is targeted to some PE (s) out of the Multicast group, for example, the PE1 and the PE2. So, the CB1 can transmit the load message to each PE in the Multicast group, and each PE will determine whether the load message transmitted by the CB1 is targeted to itself upon reception of the load message. If the determination result is Yes, the PE will start a loading process, that is, the PE will select the message with a destination processing device including the present PE from the received load messages, where the selection of message by the PE can be referred to “message filtering” . As illustrated in Fig. 1, the PE can perform message filtering via hardware, where the PE1 illustrates an exemplary PE for filtering via hardware, although other PEs can also perform filtering via hardware.

Fig. 2 illustrates a structure of a PE which can include a Central Processing Unit (CPU) 21, a memory 22, a nonvolatile memory 23, and a forwarding chip 24, all of which can communicate with each other through an internal bus. A message filtering rule can be pre-stored in the forwarding chip 24, and the forwarding chip 24 can perform message filtering according to the message filtering rule upon reception of the load message transmitted by the CB.

For example, the message filtering rule preset on the PE can be preset in a User Defined Field (UDF) of the forwarding chip of the PE, and can be configured in the form of an Access Control List (ACL) .

After the PE performs message filtering on the load messages, the PE can transmit the file data in the filtered-out load message to the CPU 21 through the internal bus. The CPU 21 can load the file data into the PE so that the PE can perform a process corresponding to the file data, for example, the PE can be started after the version file data are loaded， or can be upgraded after the upgrade file data are loaded.

The CPU 21 loads the file data into the PE by storing the file data in a special storage space of the memory 22 or the nonvolatile memory 23 (e.g, RAM or ROM) and then executing the stored file data so as to start or upgrade the PE. However, in an example that the CB performs a dividing process on the file data before transmitting the file data to the PE, in other words, the CB divides the file data into a plurality of data packets (e.g., 1000 data packets) and then transmits these data packets via the load messages to the PE, loading the file data may comprises: storing, by the CPU 21, the data packets in the filtered-out load messages into a special storage space of the memory 22 or the nonvolatile memory 23； combining, by the CPU 21, all the data packets so as to form a whole file data； and executing, by the CPU, the combined file data so as to start or upgrade the PE.

Message filtering by the forwarding chip 24 according to the message filtering rule can be referred to filtering via hardware. The forwarding chip can filter the messages rapidly to thereby identify rapidly the message intended for the PE and also alleviate a burden on the CPU because the message will not be further identified by the CPU (to identify whether it is a load message intended for the PE), and the burden on the CPU can be alleviated to thereby enhance the stability of the PE so as to lower a probability of a failure in loading the file data by the CPU.

Fig. 3 illustrates messaging filtering performed by the forwarding chip in the PE upon reception of the load message, where the forwarding chip of the PE receives the load message transmitted by the CB arriving at the PE in 301. Upon the forwarding chip of a PE device receiving a load message, the forwarding chip can determine whether the received load message is a target load message according to the preset message filtering rule, where the target load message is a kind of message which employs the PE device as a destination processing device, and if the PE is the destination processing device of the message, that is, the CB intends to transmit the message to the PE for processing (e.g., loading the file data in the message) ； and if the destination processing devices include the PE, then the load message can be referred to as a target load message, which includes the file data to be loaded by the PE, e.g., version file data, upgrade file data, etc.

For example, if the CB transmits the load message carrying the upgrade file to the PE1 and the PE2 in a Multicast mode, then from the perspective of the CB, the CB intends to have the PE1 but not the PE2 load the file data of the upgrade file, that is, the PE1 is the destination processing device of the load message and can load the file data of the load message, and the PE2 is not the destination processing device of the load message. Moreover the forwarding chip can store the preset message filtering rule in the user defined field in the form of an access control list.

The forwarding chip can transmit the file data included in the load message to the CPU of the PE device in 302 after filtering out the target load message, and the CPU can load the file data into the PE, for example, load the data of the version file to start the PE. Taking loading of the version file as an example, the target load message can be identified rapidly by the forwarding chip and loaded rapidly by the CPU to thereby greatly speed up starting of the PE.

In an example, the PE can determine whether the load message is the target load message upon reception of the load message as follows: the forwarding chip obtains a PE device identifier preset in an identifier field in the load message； and if the PE device identifier is the same as the first PE device information in the preset message filtering rule, then the forwarding chip determines that the load message is the target load message, where the first PE device information indicates that the PE device is the destination processing device of the load message.

In another example, the PE can determine whether the load message is the target load message upon reception of the load message as follows: the forwarding chip obtains a PE device identifier preset in an identifier field in the load message； and if the PE device identifier is different from the second PE device information in the preset message filtering rule, then the forwarding chip determines that the load message is the target load message, where the second PE device information indicates that another PE device other than the PE device is the destination processing device of the load message.

In an example, before the forwarding chip of the PE obtains the PE device identifier preset in the identifier field, the forwarding chip can further obtain message type information preset in a type field in the load message, and if the message type information preset in a type field of the load message is the same as a target type information in the preset message filtering rule , proceeding to obtain the PE device identifier .

How the forwarding chip according to this disclosure filters the load message will be described below in the scenario illustrated in Fig. 4. As illustrated in Fig. 4, there are three PEs connected with a CB, which are a PE1, a PE2, and a PE3； and if the CB intends to transmit a version file to the three PEs, then it can transmit load messages carrying the version file to all of the three PEs in a Multicast mode, that is a PE1, a PE2, and a PE3 belong to a Multicast group , and the respective PEs can identify the messages intended for themselves by performing message filtering above.

The CB can set the message format of the load message before transmitting the load message, where reference can be further made to Fig. 4 for an example of the message format of the load message. The load message can include a message header and a data segment, where the data segment can carry file data, e.g., file data of the version file or file data of an upgrade file； and the message header can include a plurality of fields for defining a destination MAC address, a source MAC address, etc., of the load message. The respective fields in the message header are defined as follows:

DMAC: the field occupying six bytes can carry an MAC address unused in the family of protocols 0180-c2xx-xxxx so as not to collide with the MAC addresses in the existing protocols. For example, the CB can transmit the load message in a Multicast mode, where typically a customized Multicast address can start with 01, whereas a general two-layer Multicast address can start with 0180-c2, and the Multicast address can be customized freely (without colliding with the other protocols) , e.g., 0180-c2ff-fff0.

SMAC: the field occupying six bytes can be an internal MAC address and typically a uni-cast MAC address. For example, the field can be an MAC address of the CB.

802.1Q: the field occupying four bytes can be defaulted as 0x81000001 for which a defaulted VLAN tag and ID are used.

ETYPE: the field occupying two bytes can be of the private type, e.g., 0xDCBA, for example, if the field is 0xDCBA (or can be in another form) , then it indicates that the message is a load message for a version file； and the field ETYPE can be referred to as a type field, and for example, 0xDCBA above can be referred to as message type information.

DEST NODE BITMAP (DNBMP) : each PE device can be provided with a globally unique slot ID in the IRF3 management domain； and if the message is valid for the PE, for example, the destination processing device of the load message is the PE, then the position in the DNBMP field corresponding to the slot ID of the PE can be set to 1.

Fig. 4 illustrates a structure of the DNBMP field above, where the DNBMP field can start with the 19^th byte of the load message because the DMAC, SMAC and other fields preceding the DNBMP field occupy 18 bytes in total. Fig. 4 illustrates three bytes in the DNBMP field, each of which includes eight bits. For example, each of the bits can correspond to one of the PEs, and if one of the PEs is selected as the destination processing device of the load message, then the CB can set the value of the bit corresponding to the PE to 1.

As illustrated in Fig. 4, if slot IDs of the PE1, the PE2, and the PE3 are 1, 2 and 3 respectively, then the slot IDs can correspond to the respective bits in the DNBMP field as follows: the PE1 with the slot ID of 1 corresponds to the 1^st bit of the 19^th byte； the PE2 with the slot ID of 2 corresponds to the 2^nd bit of the 19^th byte； and the PE3 with the slot ID of 3 corresponds to the 3^rd bit of the 19^th byte. If one of the PEs is selected as the processing device of the load message, then the CB can set the value of the bit corresponding to the PE to 1.

For example, the CB intends to transmit the load message to the PE with the slot ID of 2, that is, the PE 2 with the slot ID of 2 is the destination processing device of the load message, then the 2^nd bit of the 19^th byte where the DNBMP field is located can be set 1 while setting the other bits to 0 as illustrated in Fig. 4.

In another example, if the PE with the slot ID of 10 is the destination processing device of the load message, then the bit corresponding to the PE can be calculated by dividing 10 by 8 so the integer quotient is 1, and the remainder is 2, where the integer of 1 indicates pushing backward by one byte from the starting byte of the DNBMP field to the 20^th byte in Fig. 4, and the remainder of 2 indicates the 2^nd bit of the 20^th byte, and this bit corresponds to the PE with the slot ID of 10, so the bit can be set to 1, while the other bits are set to 0.

In an example, the DNBMP field illustrated in Fig. 4 above can be referred to as an identifier field, and the value to which the CB sets the bit corresponding to the PE in the identifier field can be referred to as a PE device identifier. The PE device identifier includes “0” or “1” , for example, where if the PE is the destination processing device of the message, then the CB can set the corresponding PE device identifier to “1” ； and if the PE is not the destination processing device of the message, then the CB can set the corresponding PE device identifier to “0” .

Further to the message format of the load message above, the CB will set the message in the format and transmit the load message to the PE. The example of this disclosure illustrates exemplary transmission of the message by the CB to the PE, but the CB can transmit the message to the PE in another manner. Still referring to the example of Fig. 4, if the number of load messages required for each PE to carry version file data is N, then the CB can transmit the load messages to the PE1, the PE2, and the PE3 illustrated in Fig. 4 in a Multicast mode, but determine different PEs as the destination processing device at different instances of time. For example, the PE1 starts to load the messages at an instance of time t1, the PE2 starts to load the messages at an instance of time t2, and the PE3 starts to load the messages at an instance of time t3, where t1<t2<t3, that is, the CB firstly determines the PE1 as the destination processing device of the load messages, the PE2 as the destination processing device of the load messages after a period of time, and the PE3 as the destination processing device of the load messages after a further period of time. If it takes a total length of time T for each PE to load separately (that is, only this PE loads) the messages, where t2-t1<T, and t3-t1<T, then the forward rate of the messages of the CB can be defined as r＝N/T.

The CB determines the PE1, the PE2, and the PE3 as the destination processing device of the load messages in an order as illustrated in Fig. 5 where the CB transmit the messages in the following several transmission phases:

A first round of transmission:

1. In the period of time from t1 to t2, the CB can set the bit corresponding to the PE1 in the DNBMP fields in the transmitted load messages to 1 and the other bits to 0 to indicate that the PE1 is the destination processing device of the load messages transmitted by the CB in the period of time.

In the period of time from t1 to t2, the number n1 of load messages transmitted by the CB is n1＝ (t2-t1) +r； and

2. In the period of time from t2 to t3, the CB can set the bits corresponding respectively to the PE1 and the PE2 in the DNBMP fields in the transmitted load messages to 1 and the other bits to 0 to indicate that the PE1 and the PE2 are the destination processing devices of the load messages transmitted by the CB in the period of time.

In the period of time from t2 to t3, the number n2 of load messages transmitted by the CB is n2＝ (t3-t2) +r； and

3. In the period of time from t3 to (t1+T) , the CB can set all the bits corresponding respectively to the PE1, the PE2, and the PE3 in the DNBMP fields in the load messages to 1 and the other bits to 0 to indicate that the PE1, the PE2, and the PE3 are the destination processing devices of the load messages transmitted by the CB in the period of time.

In the period of time from t3 to (t1+T) , the number n3 of load messages transmitted by the CB is n3＝ ( (t1+T) -t3) *r；

At this time the CB has transmitted all the N messages, and the states of the three PEs are as follows: the PE1 has loaded all the messages, the PE2 has not loaded the n1 messages in the period of time from t1 to t2, and the PE3 has not loaded the (n1+n2) messages in the period of time from t1 to t3.

Next the CB starts a second round of transmission:

4. In the period of time (t1+T) to (t2+T) , the PE1 has loaded all the messages in this period of time, so the CB will not determine the PE1 as the destination processing device of the messages but set the bits corresponding respectively to the PE2 and the PE3 in the DNBMP fields of the load messages to 1 and the other bits to 0 to indicate that the PE2 and the PE3 are the destination processing devices of the load messages transmitted by the CB in the period of time.

In the period of time from (t1+T) to (t2+T) , the number n4 of load messages transmitted by the CB is n4＝ (t2-t1) *r；

5. In the period of time (t2+T) to (t3+T) , both the PE1 and the PE2 have loaded all the messages in this period of time, so the CB can set the bit corresponding to the PE3 in the DNBMP fields of the load messages to 1 and the other bits to 0 to indicate that the PE3 is the destination processing device of the load messages transmitted by the CB in the period of time.

In the period of time from (t2+T) to (t3+T) , the number n5 of load messages transmitted by the CB is n5＝ (t3-t2) *r.

At this time the three PE devices have loaded all the messages, which actually takes a period of time of:

t＝ (n1+n2+n3+n4+n5) /r＝T+ (t3-t1) .

In the method above of transmitting the messages by the CB, the period of time to load the messages can be shortened, for example, to thereby speed up starting of the PE into which the version file is loaded. For example, if the CB transmits the messages to the PE in a uni-cast mode, then the number of transmitted by the CB is 3N, and it will take a period of time of t＝3N/r＝3T for the three PEs to load all the messages.

It shall be noted that the load messages are transmitted to the different destination processing PEs in the different periods of time in which the messages are transmitted, the CB in the scenario illustrated in Fig. 4 can transmit the load messages concurrently to the PE1, the PE2, and the PE3 in a Multicast mode, and all the three PEs can receive the load messages, so that the respective PEs can perform filtering via hardware as illustrated in Fig. 3 to determine whether the destination processing devices of the load messages transmitted by the CB include the present PE. How the PE determines whether the destination devices of the received load message include the present PE will be described below in two examples in which the message filtering rule on the forwarding chip of the PE is in the form of an ACL.

In an example, the ACL of the message filtering rule can be set in a positive form as depicted in Table 1 below, where the ACL list includes a Rule and a corresponding processing Action, that is, if the characteristic of the load message received by the forwarding chip of the PE satisfies a condition defined in the Rule, then the forwarding chip can perform a process defined in the Action. Since the Action in Table 1 is a positive action, that is, the message is transmitted to the CPU, which means that the message is accepted, the ACL can be referred to as an ACL in the positive form.

Table 1: ACL list in the positive form

The characteristic of the message defined in the Rule in Table 1 above is examined: for example, the Offset in the first row of condition in the Rule represents the number of bytes by which a shift is made rightward from the starting byte of the load message, and referring to the example in Fig. 4, the shift by 16 bytes comes to the field ETYPE, which can be referred to as a type field. The Length represents the length of the field, that is, the length of the field ETYPE is two bytes, the Mask represents a mask, and the Data represents the value of the mask, where the Mask and the Data represent the value of two bytes following the 16 bytes by which the shift is made, i.e., the field of the field ETYPE above. In Table 1, the value of the Data is 0xDCBA, and the first row of condition “16———2——0xffff——0xDCBA” in the Rule in Table 1 represents a condition to be satisfied, i.e., “the value of the field ETYPE in the load message is 0xDCBA” , and the value of the field ETYPE included in the load message can be referred to as message type information to indicate the message type of the load message, whereas the value 0xDCBA of the Data in the ACL in Table 1 can be referred to as target type information to indicate the ACL defines such a condition that the message type information in the load message shall be the target type information, that is, the first row of condition in the Rule defines the message type of the target load message.

In another example, the second row of condition in the Rule is “18+ (Slot ID/8) ———1——1>> (SlotID％8) ——1” , where 18+ (Slot ID/8) represents the following handling will be executed: check “abyte spaced from the starting byte of the field NDBMP by such a number of bytes that is the integer quotient of the slot ID divided by 8” , and 1>> (SlotID％8) represents the following handling will be executed: check a bit represented by the remainder in the byte represented by the integer in that the remainder lies in the byte at the bit. The Data of 1 represents that the bit shall be defined as 1. Under the second row of condition in the Rule, the condition defines a check of the corresponding bit which shall be set to 1, that is, the PE corresponding to the bit is the destination processing device of the load message.

In an example, if the PE with the slot ID of 3 is the destination processing device of the load message, then the bit corresponding to the PE will be checked in that the PE device identifier on the bit shall be 1. The bit can be checked by dividing the slot ID by 8 so that the resulting integer is 0, that is, 18+0＝18, which means a shift rightward by 18 bytes from the starting byte (the starting byte on the left DMAC side) , and then Length＝1 is checked, so that check the contents in one byte following the 18 bytes by which the shift is made rightward. The one byte following the 18 bytes by which the shift is made rightward from the starting byte is the 19^th byte, and the slot ID of 3 divided by the 8 results in the remainder of 3, so the 3^rd bit of the 19^th byte is checked, which shall be 1.

In another example, if the PE with the slot ID of 10 is the destination processing device of the load message, then the bit corresponding to the PE is checked by dividing the slot ID by 8 resulting in the integer of 1, that is, 18+1＝19, that is, one byte following the 19 bytes by which the shift is made rightward from the starting byte is the 20^th byte, and the remainder is 2, so the 2^nd bit of the 20^th byte is checked, and the PE device identifier on the bit shall be 1.

That is, there are two conditions preset in the Rule in Table 1, where one of the conditions is the first row in the Rule to define the condition “16———2——0xffff——0xDCBA” , and the message type information defined by the type field in the load message is 0xDCBA； and the other condition is the second row in the Rule to define the condition “18+ (Slot ID/8) ———1——1>> (SlotID％8) ——1” , and the bit corresponding to the PE in the identifier field of the load message is checked, which shall be 1 under this condition.

In this example, Data-1 in the condition defined in the second row of the Rule in the ACL in Table 1 can be referred to as first PE device information, and the forwarding chip of the PE can obtain the PE device identifier on the bit corresponding to the PE in the DNBMP field in the load message upon reception of the load message, and if the PE device identifier is the same as the first PE device information so that both of them are 1, then the PE can determine that the PE is the destination processing device of the load message, that is, the load message is the target load message； or if the PE device identifier obtained by the PE is 0, and the first PE device information is 1 in Table1, so that they are different, then the PE can determine that the present PE is not the destination processing device.

In another example, the forwarding chip of the PE can firstly obtain the message type information preset in the type field in the load message upon reception of the load message, and if the message type information is the same as the target type information in Table 1, for example, both of them are 0xDCBA, then this may indicate that the message is a load message and it be further determined whether the destination processing devices of the message include the PE, that is, whether the destination processing devices of the message of the type defined by the target type information include the PE itself, according to the first PE device information. For example, after the message type is determined above as 0xDCBA, the PE further obtains the PE device identifier preset in the identifier field in the load message, for example, the PE1 obtains the value of 1 on the bit corresponding to the PE1 in the DNBMP field of the message, and this value of 1 is the same as the first PE device information in the message filtering rule in Table 1, which indicates that the PE is the destination processing device of the message, so the PE determines the message as the target load message. And the PE performs the corresponding Action in Table 1 and transmits the message to the CPU (for example, by copying or forwarding the message) .

In another example, the ACL of the message filtering rule can alternatively be preset in a negative form as depicted in Table 2 below, which differs from Table 1 in that the value of the Data in the condition defined in the second row in the Rule can be referred to as a second PE device information indicating that a PE other than the PE is the destination processing device of the load message, for example, the PE1 checks the PE device identifier on the bit corresponding to the PE1 in the received load message, and if the identifier is the same as the second PE device information which is 0, then it indicates that the PE1 is not the destination processing device of the load message. Correspondingly Table 2 further illustrates that the corresponding Action when the PE device identifier is the second PE device information, that is, the message is rejected, so for example, the PE1 can discard the message without transmitting the file data in the message to the CPU for loading.

Table 2: ACL list in the negative form

On the contrary, if the PE chooses the target load message transmitted to itself, according to the ACL in Table 2, then the PE can operate as follows: for example, the PE1 can obtain the PE device identifier preset on the bit corresponding to the PE1 in the identifier field DNBMP in the load message upon reception of the load message, and if the PE device identifier is 1, that is, the PE device identifier is different from the second PE device information of 0 in the ACL in Table 2, then it indicates that the load message is the target load message.

The forwarding chip of the PE performing message filtering according to the ACL in Table 2 can further obtain the message type information preset in the type field in the load message before the PE device identifier preset in the identifier field in the load message is obtained, and determine that the following handling will be done: determine whether the PE itself is one of the destination processing device (s) of the message according to the message filtering rule, upon determining that the message type information is the same with the target type information in the preset message filtering rule.

Message filtering can be performed by the hardware, i.e., the forwarding chip, in the PE so thereby greatly improve the processing speed in message filtering by the PE and accordingly alleviate the burden on the CPU due to message filtering via hardware so that the PE will operate more stably and a failure in loading will occur less frequently； for example, the PE can identify rapidly that the load message of the CB transmitting the version file is intended for the PE, upon reception of the message, that is, the PE is the destination processing device of the load message, and can load the message rapidly and successfully, to thereby speed up starting of the PE.

If the functions performed by the CPU above are embodied in the form of software functional elements and sold or used as a separate product, then a part of the product can be stored in a computer readable storage medium. Based upon such understanding, the technical teachings of this disclosure can be partly embodied in the form of a software product stored in a storage medium and including several instructions to cause a computer device (e.g., a personal computer, a server, a network device, etc. ) to perform all or a part of the steps in the methods according to the respective examples of the disclosure. The storage medium above can include a U-disk, a mobile hard disk, a Read-Only Memory (ROM) , a Random Access Memory (RAM) , a magnetic disk, an optical disk or various other medium in which program codes can be stored.

The foregoing disclosure is merely illustrative of preferred examples of the disclosure but not o intended to limit the disclosure, and any modifications, equivalent substitutions, adaptations, thereof made without departing from the spirit and scope of the disclosure shall be encompassed in the claimed scope of the appended claims.

Claims

A message processing method performed by a PE device comprising a forwarding chip and a central processing unit, and the method comprising:

determining, by the forwarding chip, upon reception of a load message, whether the load message is a target load message according to a preset message filtering rule, wherein the target load message indicates a kind of message which employs the PE device as a destination processing device； and

transmitting, by the forwarding chip, the file data comprised in the load message to the central processing unit for loading, in a case that the load message is determined as the target load message.
The method according to claim 1, wherein determining whether the load message is the target load message comprises:

obtaining, by the forwarding chip, a PE device identifier preset in an identifier field of the load message； and

determining, by the forwarding chip, that the load message is the target load message, in a case that the PE device identifier is the same as a first PE device information in the preset message filtering rule, wherein the first PE device information indicates that the PE device is one of the destination processing device (s) of the load message.
The method according to claim 1, wherein determining whether the load message is the target load message comprises:

obtaining, by the forwarding chip, a PE device identifier preset in an identifier field of the load message； and

determining, by the forwarding chip, that the load message is the target load message, in a case that the PE device identifier is different from a second PE device information in the preset message filtering rule, wherein the second PE device information indicates that the destination processing device of the load message is a PE device other than the PE device.
The method according to claim 2 or 3, wherein prior to obtaining the PE device identifier, the method further comprises:

obtaining, by the forwarding chip, a message type information preset in a type field of the load message, and

proceeding to obtaining the PE device identifier, in a case that the message type information is the same as a target type information in the preset message filtering rule.
The method according to claim 1, wherein prior to the reception of the load message, the method further comprises:

storing, by the forwarding chip, the preset message filtering rule in a user defined field with the form of an access control list.
A PE device, comprising a forwarding chip in which a preset message filtering rule is stored, and a central processing unit, wherein:

the forwarding chip is to perform the following operations according to the preset message filtering rule:

determine whether a load message is a target load message, upon reception of a load message, wherein the target load message indicates a kind of message which employs the PE device as a destination processing device； and

transmit file data comprised in the load message to the central processing unit, in a case that the load message is determined as the target load message； and

the central processing unit is to load the file data in the load message into the PE device.
The PE device according to claim 6, wherein the forwarding chip determine whether a load message is a target load message according to the preset message filtering rule , comprises:

if a PE device identifier preset in an identifier field of the load message is a first PE device information which indicates that the PE device is one of the destination processing device (s) of the load message, then determining the load message as the target load message and transmitting the load message to the central processing unit.
The PE device according to claim 6, wherein the forwarding chip determine whether a load message is a target load message according to the preset message filtering rule , comprises:

if a PE device identifier preset in an identifier field of the load message is a second PE device information which indicates that the destination processing device of the load message is a PE device other than the PE device, then rejecting the load message.
The PE device according to claim 7 or 8, wherein the forwarding chip further performs the following operation according to the preset message filtering rule :

if a message type information preset in a type field of the load message is the same as a target type information in the preset message filtering rule, proceeding to obtain the PE device identifier.
The PE device according to claim 6, wherein the preset message filtering rule is stored in a user defined field of the forwarding chip with the form of an access control list.