WO2014026554A1 - 报文拥塞处理方法及装置 - Google Patents

报文拥塞处理方法及装置 Download PDF

Info

Publication number
WO2014026554A1
WO2014026554A1 PCT/CN2013/080824 CN2013080824W WO2014026554A1 WO 2014026554 A1 WO2014026554 A1 WO 2014026554A1 CN 2013080824 W CN2013080824 W CN 2013080824W WO 2014026554 A1 WO2014026554 A1 WO 2014026554A1
Authority
WO
WIPO (PCT)
Prior art keywords
packet
length
parameter
congestion
fragmentation
Prior art date
Application number
PCT/CN2013/080824
Other languages
English (en)
French (fr)
Inventor
吴立群
王志忠
刘衡祁
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Priority to US14/422,176 priority Critical patent/US9992116B2/en
Priority to EP13829109.1A priority patent/EP2887591A4/en
Publication of WO2014026554A1 publication Critical patent/WO2014026554A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/11—Identifying congestion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/30—Flow control; Congestion control in combination with information about buffer occupancy at either end or at transit nodes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/32—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/622—Queue service order
    • H04L47/623—Weighted service order
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/32—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames
    • H04L47/326—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames with random discard, e.g. random early discard [RED]

Definitions

  • the present invention relates to the field of network communication technologies, and in particular, to a packet congestion processing method and apparatus.
  • BACKGROUND In the process of network communication, if the data traffic sent by the user is not restricted, the data bursting by a large number of users may cause network congestion; excessive congestion may cause great harm to network resources, and measures must be taken to remove it.
  • Congestion avoidance is a flow control mechanism that monitors the usage of network resources and removes network overload by adjusting the traffic of the network. For example, the network actively discards packets when the network is congested. When discarding packets, the device minimizes packet discarding and delay while maximizing the throughput and utilization efficiency of the network through the effective combination of a specific packet loss policy and a remote flow control mechanism.
  • the Weighted Random Early Detection (WRED) algorithm is mainly used in the network processor to implement packet congestion processing.
  • the method mainly uses the average queue length and the set average queue upper limit and lower limit comparison to determine the probability of discarding, and the average queue length is a function of the queue length, reflecting the trend of the queue.
  • the probability of discarding is determined by comparing the average queue length with the set average queue upper limit and lower limit. Using this method to discard packets requires more resources and longer waiting time.
  • this method calculates the packet length corresponding to the packet, the packet length of the current packet needs to be obtained in advance, and the packet enters the network processor according to the packet fragmentation form.
  • a random access memory (RAM) resource is used to store the packet fragmentation, and the entire packet is received to obtain the packet length before the congestion processing can be performed.
  • the number of interfaces provided by the current network processor is increasing. The packet fragmentation of each interface is interleaved in most cases. This method consumes RAM resources proportional to the number of interfaces. When the number of interfaces exceeds a certain range, the consumption of such RAM resources is unacceptable.
  • the main object of the present invention is to provide a packet congestion processing method and apparatus, which are intended to efficiently solve the congestion problem of communication network packets.
  • the invention discloses a packet congestion processing method, which comprises the following steps: Receiving packet fragmentation and fragmentation parameters of the packet, and obtaining a congestion parameter required for the packet congestion processing according to the fragmentation parameter; and the packet fragmentation is a packet start (SOP) packet classification And determining, according to the fragmentation parameter and the congestion parameter, an average queue length and a length threshold of the packet, and processing the packet.
  • SOP packet start
  • the step of obtaining an average queue length and a length threshold of the packet according to the fragmentation parameter and the congestion parameter, and the step of processing the packet further includes the following steps: when the message ends (EOP) When the fragment arrives, the average queue length of the packet is corrected and updated according to the difference between the actual packet length and the configured packet length in the fragment parameter.
  • EOP message ends
  • the step of obtaining the average queue length and the length threshold of the packet according to the fragmentation parameter and the congestion parameter, and the step of processing the packet specifically includes: according to the fragmentation parameter and the congestion parameter, Obtaining an average queue length of the packet; obtaining a weighted random early detection (WRED) policy group according to the packet queue number in the fragmentation parameter, according to the queue priority and the policy group parameter in the WRED policy group, Obtaining an upper limit length threshold and a lower limit length threshold of the average queue length; determining a relationship between the average queue length and the upper limit length threshold and a lower limit length threshold; if the average queue length is greater than the upper limit length threshold, allowing Disabling the SOP packet fragmentation; if the average queue length is less than the lower threshold length threshold, allowing the SOP packet to be fragmented; if the average queue length is not less than the lower threshold length threshold and not greater than
  • the upper limit length threshold is used to randomly discard the SOP packet fragment according to the WRED policy group; according to the fragmentation of
  • the step of randomly discarding the SOP packet fragment according to the WRED policy group specifically includes: generating Calculating and acquiring the discarding probability of the SOP packet fragment according to the WRED policy group; determining whether the random number is smaller than the discarding probability; If yes, the SOP message fragmentation is allowed to be discarded; if not, the SOP message fragmentation is allowed to pass.
  • the step of correcting and updating the queue length of the packet according to the difference between the actual packet length and the configured packet length in the fragment parameter further includes a step And discarding, according to the processing result of the packet, the packet that is allowed to be discarded, and transmitting the allowed packet and the average queue length of the updated packet.
  • the present invention also discloses a packet congestion processing apparatus, including: a congestion input module, configured to receive a packet fragmentation and a fragmentation parameter of a packet, and obtain a congestion parameter required for the packet congestion processing according to the fragmentation parameter.
  • the congestion processing module is configured to: when the packet fragmentation is SOP packet fragmentation, obtain an average queue length and a length threshold of the packet according to the fragmentation parameter and the congestion parameter, where the packet is received. Process it.
  • the congestion processing module is further configured to: when the EOP packet fragment arrives, correct and update the average of the packet according to the difference between the actual packet length and the configured packet length in the fragmentation parameter. Queue length.
  • the congestion processing module specifically includes: a parameter acquisition sub-module, configured to: obtain an average queue length of the packet according to the fragmentation parameter and the congestion parameter; and according to the packet queue in the fragmentation parameter
  • the WRED policy group obtains the upper limit length threshold and the lower limit length threshold of the average queue length according to the queue priority and the policy group parameter in the WRED policy group.
  • the packet decision submodule is set to - The relationship between the average queue length and the upper limit length threshold and the lower limit length threshold; if the average queue length is greater than the upper limit length threshold, allowing the SOP packet to be fragmented; if the average queue length is smaller than the The lower limit length threshold allows the SOP message to be fragmented; And if the average queue length is not less than the lower limit length threshold and is not greater than the upper limit length threshold, the SOP packet fragment is randomly discarded according to the WRED policy group; and all the packet fragments of the packet are fragmented. The processing result of the fragmentation of the SOP message will be passed or discarded.
  • the packet decision sub-module specifically includes: a length threshold determining unit, configured to determine a relationship between the average queue length and the upper limit length threshold and a lower limit length threshold; a first result unit, configured to be at the average When the queue length is greater than the upper limit length threshold, the SOP packet fragmentation is allowed to be discarded; the second result unit is configured to allow the SOP packet fragmentation when the average queue length is less than the lower limit length threshold.
  • a length threshold determining unit configured to determine a relationship between the average queue length and the upper limit length threshold and a lower limit length threshold
  • a first result unit configured to be at the average When the queue length is greater than the upper limit length threshold, the SOP packet fragmentation is allowed to be discarded
  • the second result unit is configured to allow the SOP packet fragmentation when the average queue length is less than the lower limit length threshold.
  • the third result unit is configured to: when the average queue length is not less than the lower limit length threshold and not greater than the upper limit length threshold, randomly discard the SOP packet fragment according to the WRED policy group; And the result unit is configured to allow all the packets of the packet to be allowed to pass or discard according to the processing result of the fragmentation of the SOP packet; the third result unit specifically includes: a random parameter obtaining subunit, setting Calculating and obtaining the discarding probability of the SOP packet fragment according to the WRED policy group to generate a random number; a unit, configured to determine whether the random number is smaller than the discarding probability; and a result obtaining subunit, configured to allow the SOP packet fragmentation to be discarded when the random number is less than the discard probability; When the number is not less than the discard probability, the SOP message fragmentation is allowed to pass.
  • the packet congestion processing apparatus further includes: a congestion output module, configured to discard the message allowed to be discarded according to the processing result of the packet, and transmit the allowed message and the updated station The average queue length of the message.
  • the invention obtains the congestion parameter required for the packet congestion processing according to the received packet fragmentation and the fragmentation parameter, and according to the fragmentation parameter and the congestion parameter when the packet fragmentation is SOP packet fragmentation. Obtaining the average queue length and the length threshold of the packet, and processing the packet has the beneficial effect of efficiently solving the communication network packet, improves the operation rate of processing the congestion packet, and greatly saves the internal buffer space.
  • FIG. 1 is a schematic flow chart of a first embodiment of a packet congestion processing method according to the present invention
  • FIG. 2 is a schematic flowchart of a second embodiment of a packet congestion processing method according to the present invention
  • FIG. 4 is a schematic diagram showing the working principle of WRED in the packet congestion processing method of the present invention
  • FIG. 5 is a packet congestion processing method according to the present invention. In the method, if the average queue length of the packet is not less than the lower limit length threshold and is not greater than the upper limit length threshold, the flow chart of the SOP packet fragmentation is randomly discarded according to the WRED policy group.
  • FIG. 6 is a packet congestion processing method according to the present invention.
  • FIG. 3 is a schematic structural diagram of a first embodiment of a packet congestion processing apparatus according to the present invention
  • FIG. 8 is a schematic structural diagram of an embodiment of a congestion processing module in a packet congestion processing apparatus according to the present invention
  • FIG. 10 is a schematic structural diagram of an embodiment of a packet rejection sub-module in a congestion processing apparatus
  • FIG. 10 is a packet congestion processing apparatus according to the present invention
  • FIG. 11 is a schematic view of a second embodiment of a structure of congestion packet processing apparatus of the present invention.
  • FIG. 1 is a schematic flowchart of a packet congestion processing method according to a first embodiment of the present invention. As shown in FIG.
  • the packet congestion processing method of the present invention includes the following steps: Step S01, packet fragmentation of received packets, and a fragmentation parameter, the congestion parameter required for the packet congestion processing is obtained according to the fragmentation parameter; in a preferred embodiment, the packet sent by the upper layer related module and the fragment of the packet are received, and the packet is divided into The fragment includes a packet start (Start of Packet, SOP for short) packet fragmentation, an End of Packet (E0P) packet fragmentation, and other valid packet fragments, and the related fragmentation parameters. Including the message valid signal, the message configuration length, etc., the real length of the message is obtained by the slot interval between the SOP message fragment and the EOP message fragment; and according to the message valid signal to the corresponding internal RAM.
  • Step S01 packet fragmentation of received packets, and a fragmentation parameter, the congestion parameter required for the packet congestion processing is obtained according to the fragmentation parameter; in a preferred embodiment, the packet sent by the upper layer related module and the fragment of the packet are received, and the packet is
  • Step S02 Determine whether the packet fragmentation is an SOP packet fragmentation; if yes, execute step S03; Step S03, obtain an average queue length and a length threshold of the packet according to the fragmentation parameter and the congestion parameter. , processing the message.
  • the packet after receiving the information such as the real length of the packet, the fragment of the packet, the valid signal of the packet, the length of the packet configuration, and the related fragmentation information after the processing, it is determined whether the packet is in the packet. Contains SOP packet fragmentation; if yes, calculates the queue length by configuring the packet length. After the queue length is determined, the average queue length corresponding to the current packet is calculated based on the queue length, the weight of the queue, and the average queue length. After the average queue length corresponding to the current packet is obtained, the current packet upper limit length threshold and the average queue lower length threshold are combined to discard the current packet. Specifically, if the average queue length is greater than the average queue upper limit, the packet will be discarded.
  • the packet will not be discarded.
  • the length threshold which is greater than the average queue lower length threshold, begins to randomly discard incoming packets.
  • the WRED algorithm is used for discarding decisions. That is, the discard probability is calculated according to the WRED algorithm, and the obtained discard probability is compared with the generated random number. If the generated random number is smaller than the discard probability, the packet will be discarded. If the generated random number is greater than the discard probability, Then this message will not be discarded. All message fragments of the entire message are passed or discarded according to the calculation result of the SOP fragment.
  • the length of the packet is the length of the configuration, not the actual length of the packet. If the SOP packet is discriminated and discarded, the subsequent fragments of the packet refer to the SOP fragment. The judgment result is all discarded; if it is passed, the subsequent fragments of the message are all passed by referring to the judgment result of the SOP slice.
  • the invention obtains the congestion parameter required for the packet congestion processing according to the received packet fragmentation and fragmentation parameters, and when the packet fragmentation is only the SOP packet fragmentation, according to the fragmentation parameter and congestion
  • the parameter obtains the average queue length and the length threshold of the packet, and the method for processing the packet has the beneficial effect of effectively solving the congestion of the communication network packet, improves the operation rate of processing the congestion packet, and greatly saves the internal cache. space.
  • FIG. 2 is a schematic flowchart of a second embodiment of a packet congestion processing method according to the present invention.
  • the difference between this embodiment and the first embodiment of the packet congestion processing method of the present invention is that only step S04 is added;
  • the packet congestion processing method of the present invention obtains an average queue length and a length threshold of the packet according to the fragmentation parameter and the congestion parameter in step S03, and processes the packet after the step of processing the packet.
  • Step S04 When the EOP packet fragment arrives, the queue length of the packet is corrected and updated according to the difference between the actual packet length and the configured packet length in the fragment parameter.
  • the packet congestion processing method of this embodiment if the packet fragmentation is both the SOP fragment and the EOP fragment, the difference between the actual packet length and the configured packet length is used to correct and update the queue length of the current packet.
  • the actual packet length is obtained by using a slot interval between the EOP packet fragment and the SOP packet fragment.
  • the average queue length corresponding to the current packet is calculated according to the actual packet length, the weight of the queue calculation, and the average queue length of the packet fragment.
  • the slice is configured to perform subsequent calculations.
  • the SOP packet When the SOP packet is fragmented, it is determined whether the packet is discarded. When the EOP fragment is received, the queue length is corrected. In this embodiment, when the EOP packet fragment arrives, the method for correcting and updating the queue length of the packet according to the difference between the actual packet length and the configured packet length in the fragment parameter has improved processing SOP congestion. The beneficial effect of the message's operating rate.
  • FIG. 3 is a schematic flowchart of an embodiment of processing a packet according to a fragmentation parameter and a congestion parameter according to the fragmentation parameter and the congestion parameter in the packet congestion processing method of the present invention.
  • This embodiment only performs step S03.
  • step S03 the average queue length and the length threshold of the packet are obtained according to the fragmentation parameter and the congestion parameter, and the step of processing the packet is performed.
  • the method includes the following steps: Step S11: Acquire an average queue length of the packet according to the fragmentation parameter and the congestion parameter.
  • Step S12 Obtain a weighted random early detection WRED policy group according to the packet queue number in the fragmentation parameter, and obtain an upper limit length of the average queue length according to the queue priority and the policy group parameter in the WRED policy group. Threshold and lower limit length thresholds;
  • the WRED algorithm is used to process message congestion. Referring to FIG. 4, FIG. 4 is a schematic diagram of the WRED working principle in the packet congestion processing method of the present invention. As shown in FIG.
  • the WRED policy group is obtained according to the received packet queue number, and the policy group parameter is obtained according to the policy group.
  • the parameters include the queue calculation weight and the queue length threshold. According to the WRED policy group and the packet priority, the parameters required for the current packet congestion processing are obtained in the corresponding internal RAM, such as the average queue upper limit length threshold and the average queue lower limit length threshold. , discard curve slope parameters, etc.
  • Step S13 determining a relationship between the average queue length and the upper limit length threshold and the lower limit length threshold; if the average queue length is greater than the upper limit length threshold, performing step S14; if the average queue length is less than the lower limit Step S15 is performed, if the average queue length is not less than the lower limit length threshold and is not greater than the upper limit length threshold, step S16 is performed; Step S14, the SOP packet fragmentation is allowed to be discarded; S15, the SOP packet fragmentation is allowed to pass; Step S16, the SOP packet fragment is randomly discarded according to the WRED policy group; and the relationship between the average queue length and the upper limit length threshold and the lower limit length threshold is determined.
  • the SOP packet fragmentation is allowed to be discarded; if the average queue length is less than the lower limit length threshold, the SOP packet fragmentation is allowed to pass; If the average queue length is not If the threshold is less than the lower threshold and is not greater than the upper threshold, the SOP packet fragment is randomly discarded according to the WRED policy group.
  • the discarding the SOP packet fragment according to the WRED policy group specifically includes: generating a random number, and calculating and acquiring the discarding probability of the SOP packet fragment according to the WRED policy group If the random number is smaller than the discard probability, the SOP packet fragmentation is allowed to be discarded; if the random number is not less than the discard probability, the SOP packet fragmentation is allowed to pass.
  • Step S17 According to the processing result of the SOP packet fragmentation, all the packets of the packet are allowed to be fragmented or discarded. All the message fragments in the message are processed according to the processing result of the SOP message fragmentation; if the SOP message fragmentation judgment is discarded, the subsequent fragments of the message are referred to the judgment of the SOP fragment.
  • the result is all discarded; if the SOP message fragmentation is judged, the subsequent fragments of the message are all passed by referring to the judgment result of the SOP fragment.
  • the method for obtaining the average queue length of the SOP packet fragments according to the fragmentation parameter and the congestion parameter, and processing the packet according to the average queue length has the beneficial effect of efficiently solving the communication network packet, thereby saving the effect. Internal cache space.
  • FIG. 5 is a flowchart of an embodiment of randomly dropping SOP packets according to a WRED policy group, if the average queue length of the packet is not less than the lower threshold and not greater than the upper threshold, in the packet congestion processing method of the present invention.
  • the step of randomly discarding the SOP packet according to the WRED policy group includes: Step S21: Generate a random number, calculate and obtain a discarding probability of the SOP packet fragment according to the WRED policy group; generate a random number R before randomly discarding the SOP packet fragment, and according to WRED
  • the algorithm calculates the discarding probability Pb .
  • Step S22 determining whether the random number is smaller than the discarding probability; if yes, executing step S23; if not, executing step S24; step S23, allowing the SOP message to be fragmented; step S24, allowing the The SOP message was sharded.
  • the SOP packet fragment is randomly discarded by arbitrating the random number R and the discard probability Pb. Specifically, if the random number R is smaller than the discard probability Pb, the SOP packet fragmentation is discarded; if the random number R is smaller than the discard probability If Pb does not hold, the SOP packet fragment is not discarded. If the SOP packet is discarded, the packet is discarded. If the SOP packet is not discarded, the packet is not discarded. In this embodiment, by generating a random number and calculating the discarding probability according to the WRED policy group, the method of randomly discarding the packet according to the relationship between the random number and the discarding probability has the beneficial effect of further improving the congestion processing efficiency.
  • FIG. 6 is a schematic flowchart of a third embodiment of a packet congestion processing method according to the present invention.
  • the difference between this embodiment and the second embodiment of the packet congestion processing method of the present invention is that only step S05 is added;
  • the packet congestion processing method of the present invention corrects and updates the difference between the actual packet length and the configured packet length in the fragmentation parameter when the EOP packet fragment arrives in step S04.
  • the step of the average queue length of the packet further includes the following steps: Step S05: Discarding the packet that is allowed to be discarded according to the processing result of the packet, and transmitting the allowed packet and the updated packet The length of the queue.
  • the average queue length of the updated message is transmitted to the relevant module according to the processing result of the packet discarding, and is set to update the average queue length stored in the RAM, and directly discard the packet to be discarded according to the discarding judgment result.
  • Discard send normal packets to the lower-level module, and wait for the processing of the lower-level module.
  • the method for processing the packet according to the packet processing result and transmitting the average queue length of the updated packet according to the packet processing result further improves the operation rate of processing the congestion packet and saves the internal cache space. Beneficial effect.
  • FIG. 7 is a schematic structural diagram of a packet congestion processing apparatus according to a first embodiment of the present invention.
  • the packet congestion processing apparatus of the present invention includes: a congestion input module 01, configured to receive a packet of a packet. And a fragmentation parameter, the congestion parameter required for the packet congestion processing is obtained according to the fragmentation parameter; in a preferred embodiment, the congestion input module 01 receives the packet and the packet fragment sent by the upper layer related module.
  • the packet fragmentation includes a packet start (Start of Packet, SOP for short) packet fragmentation, an End of Packet (EOP) packet fragmentation, and other valid packet fragments.
  • the related fragmentation parameter includes a packet valid signal, a packet configuration length, and the like, and obtains a real length of the packet by using a slot interval between the SOP packet fragment and the EOP packet fragment; and the congestion input module 01 is configured according to The message valid signal is sent to the corresponding internal RAM to obtain the parameters required for the current packet congestion processing, such as the average queue upper limit length threshold, the average queue lower limit length threshold, the queue calculation weight, and the average queue length. And discarding the curve slope parameter, etc.; after the congestion input module 01 reads the parameters required for the current packet congestion processing from the internal RAM, the series parameters, the packet configuration length, and the counting of all the fragments of the packet are obtained.
  • the real length of the packet is transmitted to the congestion processing module 02, such as the true length of the packet, the fragmentation of the packet, the valid signal of the packet, the length of the packet configuration, and the related fragmentation information after processing.
  • the congestion processing module 02 is configured to: when the packet fragment is SOP packet fragmentation, obtain an average queue length and a length threshold of the packet according to the fragmentation parameter and the congestion parameter, where the packet is received. Process it.
  • the congestion processing module 02 is further configured to: when the EOP packet fragment arrives, correct and update the average queue length of the packet according to the difference between the actual packet length and the configured packet length in the fragmentation parameter. .
  • the congestion processing module 02 determines the information. Whether the SOP packet fragment is included in the packet; if yes, the congestion processing module 02 calculates the queue length by using the configured packet length; after determining the queue length, the congestion processing module 02 passes the queue length, the queue calculation weight, and the average queue length. Calculate the average queue length corresponding to this message. After the average queue length corresponding to the current packet is obtained, the congestion processing module 02 combines the average queue upper limit length threshold and the average queue lower limit length threshold to discard the current packet.
  • the congestion processing module 02 if the average queue length is greater than the average queue upper limit length threshold, the congestion processing module 02 allows the packet to be discarded; if the average queue length is less than the average queue lower length threshold, the congestion processing module 02 allows the packet to be discarded at this time; If the average queue length is less than the average queue upper limit length threshold and greater than the average queue lower limit length threshold, the congestion processing module 02 starts to randomly discard the incoming message.
  • the congestion processing module 02 allows the WRED algorithm to be used for discarding decisions. That is, the discard probability is calculated according to the WRED algorithm, and the obtained discard probability is compared with the generated random number, if the generated random number is small. The congestion processing module 02 allows the packet to be discarded.
  • the congestion processing module 02 allows the packet to be discarded. All message fragments of the entire message are passed or discarded according to the calculation result of the SOP fragment.
  • the length of the packet is the length of the configuration, not the actual length of the packet. If the SOP message fragment is judged to be discarded, the subsequent fragments of the message are discarded according to the judgment result of the SOP fragment; if passed, the subsequent fragments of the message are all passed by referring to the judgment result of the SOP fragment.
  • the congestion processing module 02 corrects and updates the current packet by using the difference between the actual packet length and the configured packet length.
  • the length of the queue is obtained by the slot interval between the EOP packet fragment and the SOP packet fragment.
  • the congestion processing module 02 calculates the average queue length corresponding to the current packet according to the actual packet length, the queue calculation weight, and the average queue length of the packet fragment.
  • the SOP packet fragmentation, the congestion processing module 02 uses the configured packet length for subsequent calculation; when the SOP packet is fragmented, the congestion processing module 02 determines whether the packet is discarded; when receiving the EOP fragment, the congestion Processing module 02 corrects the queue length.
  • the invention obtains relevant congestion parameters required for the packet congestion processing according to the received packet fragmentation and related fragmentation parameters, and obtains the relevant congestion parameters according to the relevant parameters when the packet fragmentation is only the SOP packet fragmentation.
  • the average queue length and the length threshold of the packet, and processing the packet according to the average queue length and the length threshold which has the beneficial effect of effectively solving the congestion of the communication network packet, and improves the operation rate of processing the congestion packet. Greatly saved internal cache space.
  • FIG. 8 is a schematic structural diagram of an embodiment of a congestion processing module in a packet congestion processing apparatus according to the present invention.
  • This embodiment further describes only the congestion processing module 02, and other modules involved in the packet congestion processing apparatus of the present invention are provided. For a detailed description of the related embodiments, details are not described herein again.
  • the congestion processing module 02 specifically includes: a parameter acquisition submodule 021, configured to: obtain an average of the packet according to the fragmentation parameter and the congestion parameter.
  • the queue length is obtained according to the packet queue number in the fragmentation parameter, and the upper limit length threshold and the lower limit length of the average queue length are obtained according to the queue priority and the policy group parameter in the WRED policy group.
  • the parameter acquisition sub-module 021 calculates the queue length by using the configured packet length. After determining the queue length, the queue length, the queue calculation weight, and the packet fragmentation average are determined according to the queue length. The queue length, the parameter acquisition sub-module 021 calculates the average queue length corresponding to the current message. In a preferred embodiment, the parameter acquisition sub-module 021 processes the message congestion by using the WRED algorithm. For the working principle of the WRED, refer to the detailed description of the embodiment of FIG. 4, and details are not described herein again.
  • the parameter obtaining sub-module 021 obtains a WRED policy group according to the received packet queue number, and obtains a policy group parameter according to the policy group, where the policy group parameter includes The queue calculation weight and the queue length threshold; the parameter acquisition sub-module 021 obtains parameters required for the current packet congestion processing according to the WRED policy group and the packet priority, such as the average queue upper limit length threshold and the average queue. Lower limit length threshold, discard curve slope parameter, etc.
  • the message decision sub-module 022 is configured to: determine a relationship between the average queue length and the upper limit length threshold and the lower limit length threshold; if the average queue length is greater than the upper limit length threshold, allow the SOP message to be Fragment discarding; if the average queue length is less than the lower limit length threshold, allowing the SOP packet to be fragmented; if the average queue length is not less than the lower limit length threshold and not greater than the upper limit length threshold, The SOP packet fragment is randomly discarded according to the WRED policy group; all packet fragments of the packet are passed or discarded according to the processing result of the SOP packet fragmentation.
  • the message decision sub-module 022 determines the relationship between the average queue length and the upper limit length threshold and the lower limit length threshold: if the average queue length is greater than the upper limit length threshold, the message decision sub-module 022 allows the SOP to be If the average queue length is less than the lower limit length threshold, the message decision sub-module 022 allows the SOP message to be fragmented; if the average queue length is not less than the lower limit length threshold and The packet decision sub-module 022 randomly discards the SOP packet fragment according to the WRED policy group.
  • the message decision sub-module 022 randomly discards the SOP message fragment according to the WRED policy group, and specifically includes: the message decision sub-module 022 generates a random number, and calculates and acquires according to the WRED policy group. a discarding probability of the fragmentation of the SOP packet; if the random number is smaller than the discarding probability, the message judging sub-module 022 allows the SOP packet to be fragmented; if the random number is not less than the discarding Probability, then the message decision sub-module 022 allows the SOP message to be fragmented.
  • the packet decision sub-module 022 processes all the packets in the packet according to the processing result of the SOP packet fragmentation; if the SOP packet fragmentation is discarded, the subsequent fragments of the packet are The judgment result of the reference SOP fragment is all discarded; if the SOP message fragmentation judgment is passed, the subsequent fragments of the message are all passed by referring to the judgment result of the SOP fragment.
  • the method for obtaining the average queue length of the SOP packet fragments according to the fragmentation parameter and the congestion parameter, and processing the packet according to the average queue length has the beneficial effect of efficiently solving the communication network packet, thereby saving the effect. Internal cache space.
  • FIG. 9 is a schematic structural diagram of an embodiment of a packet decision sub-module in a packet congestion processing apparatus according to the present invention; this embodiment only further describes the packet decision sub-module 022, and the packet congestion processing apparatus of the present invention is involved.
  • the packet congestion processing apparatus of the present invention is involved.
  • FIG. 9 please refer to the detailed description of the related embodiments, and details are not described herein again. As shown in FIG.
  • the message decision sub-module 022 specifically includes: a length threshold value determining unit 221, configured to determine the relationship between the average queue length and the upper limit length threshold and the lower limit length threshold
  • the first result unit 222 is configured to allow the SOP packet to be fragmented when the average queue length is greater than the upper limit length threshold
  • the second result unit 223 is configured to be smaller than the average queue length
  • the third result unit 224 is configured to, when the average queue length is not less than the lower limit length threshold and not greater than the upper limit length threshold, according to the The WRED policy group randomly discards the SOP packet fragmentation
  • the fourth result unit 225 is configured to allow all the packets of the packet to be allowed to pass or discard according to the processing result of the SOP packet fragmentation; Referring to FIG.
  • FIG. 10 is a schematic structural diagram of a third result unit in a message congestion processing apparatus according to the present invention; as shown in FIG. 10, the third result unit 224 is shown in FIG.
  • the method includes: a random parameter obtaining sub-unit 241, configured to generate a random number, and calculating and acquiring a discarding probability of the SOP packet fragment according to the WRED policy group; randomly performing random discarding of the SOP packet fragmentation
  • the parameter acquisition sub-unit 241 first generates a random number R, and calculates the discarding probability Pb according to the WRED algorithm .
  • the WRED algorithm refer to the detailed description of the embodiment of FIG. 4, and details are not described herein again.
  • the random parameter comparison subunit 242 is configured to determine whether the random number is smaller than the discard probability;
  • the result obtaining sub-unit 243 is configured to allow the SOP message fragment to be discarded when the random number is smaller than the discard probability; and to allow the SOP message when the random number is not less than the discard probability Fragmentation passed.
  • the random parameter R and the discard probability Pb are arbitrated by the random parameter comparison sub-unit 242, and the result obtaining sub-unit 243 randomly discards the SOP message fragment; specifically, if the random parameter comparison sub-unit 242 determines that the random number R is smaller than the discard probability Pb If yes, the result acquisition sub-unit 243 allows the SOP message fragment to be discarded.
  • the result acquisition sub-unit 243 allows the SOP message fragment to not be discarded. If the result acquisition sub-unit 243 allows the SOP message fragment to be discarded, the message is allowed to be discarded; if the result acquisition sub-unit 243 allows the SOP message not to be discarded, the message is allowed to be discarded.
  • the random discarding of the packet according to the relationship between the random number and the discarding probability has the beneficial effect of further improving the congestion processing efficiency.
  • FIG. 11 is a schematic structural diagram of a second embodiment of a message congestion processing apparatus according to the present invention.
  • the difference between the present embodiment and the first embodiment of the packet congestion processing apparatus of the present invention is that only the congestion output module 03 is added.
  • This embodiment only describes the congestion output module 03, and the packet congestion processing apparatus of the present invention relates to For other modules, please refer to the detailed description of the related embodiments, and details are not described herein again.
  • the packet congestion processing apparatus of the present invention further includes: a congestion output module 03, configured to discard the packet that is allowed to be discarded according to the processing result of the packet, and transmit the packet that is allowed to pass. The average queue length of the updated message.
  • the congestion output module 03 is responsible for receiving the packet processed by the congestion processing module 02, discarding the packet to be discarded, and sending the normal packet to the lower module, waiting for the processing of the lower module. At the same time, the congestion output module 03 transmits the updated average queue length of the message to the relevant module, and is set to update the average queue length stored in the RAM.
  • the method for processing the packet according to the packet processing result and transmitting the average queue length of the updated packet according to the packet processing result further improves the operation rate of processing the congestion packet and saves the internal cache space. Beneficial effect.

Landscapes

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

Abstract

本发明公开一种报文拥塞处理方法，包括以下步骤：接收报文的报文分片及分片参数，根据所述分片参数获取所述报文拥塞处理所需的拥塞参数；在所述报文分片为SOP报文分片时，根据所述分片参数和拥塞参数，获取所述报文的平均队列长度及长度阈值，对所述报文进行处理。本发明还公开一种报文拥塞处理装置。本发明通过接收的报文分片为SOP报文分片时，根据获取的分片参数和拥塞参数获取报文的平均队列长度及长度阈值，根据平均队列长度及长度阈值的关系对报文进行处理的方法，具有高效解决通信网络报文拥塞的有益效果，提高了处理拥塞报文的操作速率，极大地节省了内部缓存空间。

Description

报文拥塞处理方法及装置 技术领域 本发明涉及网络通信技术领域， 尤其涉及一种报文拥塞处理方法及装置。 背景技术 在网络通信过程中， 若不限制用户发送的数据流量， 则大量用户不断突发的数据 会造成网络拥塞； 过度的拥塞会对网络资源造成极大的危害， 必须采取措施加以解除。 拥塞避免是一种流控机制， 可以监视网络资源的使用情况， 通过调整网络的流量解除 网络过载， 比如网络拥塞时主动丢弃报文。 设备在丢弃报文时， 通过特定的丢包策略 和远端流控机制的有效组合， 使网络的吞吐量和利用效率最大化的同时， 使报文丢弃 和延迟最小化。 现有技术中， 网络处理器中主要采用加权随机早期检测 （Weighted Random Early Detection, 简称为 WRED) 算法来实现报文的拥塞处理。 该方法主要采用平均队列长 度和设置的平均队列上限、 下限比较来确定丢弃的概率， 而平均队列长度是队列长度 的函数， 反映了队列的变化趋势。 在对报文进行丢弃判断时， 首先需要获取相同队列 号报文的队列长度， 从而通过相关算法得到平均队列长度， 之后通过平均队列长度和 设置的平均队列上限、 下限比较来确定丢弃的概率。 采用这种方法对报文进行丢弃处 理需要耗费较多的资源和较长的等待时间。 因为这种方法计算报文对应队列长度需要 预先获取当前报文的报文长度， 而报文是按照报文分片形式进入网络处理器的。 为了 得到当前报文的长度， 需采用随机存储器 （Random Access Memory, 简称为 RAM) 资源对报文分片进行存储， 直到整个报文全部被接收以获取报文长度， 才能进行拥塞 处理。 当前网络处理器提供的接口数目越来越多， 各个接口的报文分片在大多数情况 下是相互交织的， 这种方法需要消耗与接口数目成正比的 RAM资源。 当接口数目超 过一定的范围时， 这种 RAM资源的消耗是不可接受的 发明内容 本发明的主要目的是提供一种报文拥塞处理方法及装置， 旨在高效地解决通信网 络报文的拥塞问题。 本发明公开了一种报文拥塞处理方法， 包括以下步骤： 接收报文的报文分片及分片参数， 根据所述分片参数获取所述报文拥塞处理所需 的拥塞参数； 在所述报文分片为报文起始（SOP)报文分片时， 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理。 优选地， 所述根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度及长 度阈值， 对所述报文进行处理的步骤之后还包括步骤： 当报文结束 （EOP) 报文分片到达时， 根据实际报文长度与所述分片参数中的配 置报文长度之差， 校正并更新所述报文的平均队列长度。 优选地， 所述根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度及长 度阈值， 对所述报文进行处理的步骤具体包括： 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 根据所述分片参数中的报文队列号， 获取加权随机早期检测 （WRED) 策略组， 根据所述 WRED策略组中的队列优先级及策略组参数，获取所述平均队列长度的上限 长度阈值和下限长度阈值； 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值， 则允许将所述 SOP报文分片丢弃； 若所述平均队列长度小于所述下限长度阈值， 则允许所述 SOP报文分片通过； 若所述平均队列长度不小于所述下限长度阈值且不大于所述上限长度阈值， 则根 据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 根据所述 SOP报文分片的处理结果，将所述报文的所有报文分片允许通过或丢弃。 优选地， 所述若所述平均队列长度不小于所述下限长度阈值且不大于所述上限长 度阈值， 则根据所述 WRED策略组， 随机丢弃所述 SOP报文分片的步骤具体包括： 产生一个随机数， 根据所述 WRED策略组， 计算并获取所述 SOP报文分片的丢 弃概率； 判断所述随机数是否小于所述丢弃概率; 若是， 则允许将所述 SOP报文分片丢弃； 若否， 则允许所述 SOP报文分片通过。 优选地， 所述当 EOP报文分片到达时， 根据实际报文长度与所述分片参数中的配 置报文长度之差， 校正并更新所述报文的队列长度的步骤之后还包括步骤： 根据所述报文的处理结果， 丢弃允许丢弃的所述报文， 传送允许通过的所述报文 和更新后的所述报文的平均队列长度。
本发明还公开一种报文拥塞处理装置包括： 拥塞输入模块， 设置为接收报文的报文分片及分片参数， 根据所述分片参数获取 所述报文拥塞处理所需的拥塞参数； 拥塞处理模块， 设置为在所述报文分片为 SOP报文分片时， 根据所述分片参数和 拥塞参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理。 优选地， 所述拥塞处理模块还设置为， 当 EOP报文分片到达时， 根据实际报文长 度与所述分片参数中的配置报文长度之差， 校正并更新所述报文的平均队列长度。 优选地， 所述拥塞处理模块具体包括： 参数获取子模块， 设置为- 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 根据所述分片参数中的报文队列号， 获取 WRED策略组， 根据所述 WRED策略 组中的队列优先级及策略组参数， 获取所述平均队列长度的上限长度阈值和下限长度 阈值； 报文判决子模块， 设置为- 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值， 则允许将所述 SOP报文分片丢弃； 若所述平均队列长度小于所述下限长度阈值， 则允许所述 SOP报文分片通过; 若所述平均队列长度不小于所述下限长度阈值且不大于所述上限长度阈值， 则根 据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 所述报文的所有报文分片将根据所述 SOP报文分片的处理结果通过或丢弃。 优选地， 所述报文判决子模块具体包括： 长度阈值判断单元， 设置为判断所述平均队列长度与所述上限长度阈值和下限长 度阈值的关系； 第一结果单元， 设置为在所述平均队列长度大于所述上限长度阈值时， 允许将所 述 SOP报文分片丢弃； 第二结果单元， 设置为在所述平均队列长度小于所述下限长度阈值时， 允许所述 SOP报文分片通过； 第三结果单元， 设置为在所述平均队列长度不小于所述下限长度阈值且不大于所 述上限长度阈值时， 根据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 第四结果单元， 设置为根据所述 SOP报文分片的处理结果， 将所述报文的所有报 文分片允许通过或丢弃； 所述第三结果单元具体包括： 随机参数获取子单元， 设置为产生一个随机数， 根据所述 WRED策略组， 计算并 获取所述 SOP报文分片的丢弃概率； 随机参数比较子单元， 设置为判断所述随机数是否小于所述丢弃概率； 结果获取子单元， 设置为在所述随机数小于所述丢弃概率时， 允许将所述 SOP报 文分片丢弃； 在所述随机数不小于所述丢弃概率时， 允许所述 SOP报文分片通过。 优选地， 所述报文拥塞处理装置还包括： 拥塞输出模块， 设置为根据所述报文的处理结果， 丢弃允许丢弃的所述报文， 传 送允许通过的所述报文和更新后的所述报文的平均队列长度。 本发明通过根据接收的报文分片及分片参数， 获取所述报文拥塞处理所需的拥塞 参数， 在所述报文分片为 SOP报文分片时， 根据分片参数和拥塞参数获取报文的平均 队列长度及长度阈值， 对所述报文进行处理的方法， 具有高效解决通信网络报文的有 益效果， 提高了处理拥塞报文的操作速率， 极大地节省了内部缓存空间。 附图说明 图 1是本发明报文拥塞处理方法第一实施例流程示意图； 图 2是本发明报文拥塞处理方法第二实施例流程示意图； 图 3是本发明报文拥塞处理方法中根据分片参数和拥塞参数， 获取报文的平均队 列长度， 对报文进行处理一实施例流程示意图； 图 4是本发明报文拥塞处理方法中 WRED工作原理示意图； 图 5是本发明报文拥塞处理方法中若报文平均队列长度不小于下限长度阈值且不 大于上限长度阈值， 则根据 WRED策略组， 随机丢弃 SOP报文分片一实施例流程示 意图； 图 6是本发明报文拥塞处理方法第三实施例流程示意图； 图 Ί是本发明报文拥塞处理装置第一实施例结构示意图； 图 8是本发明报文拥塞处理装置中拥塞处理模块一实施例结构示意图； 图 9是本发明报文拥塞处理装置中拥报文判决子模块一实施例结构示意图； 图 10是本发明报文拥塞处理装置中第三结果单元一实施例结构示意图； 图 11是本发明报文拥塞处理装置第二实施例结构示意图。
本发明目的的实现、 功能特点及优点将结合实施例， 参照附图做进一步说明。 具体实施方式 以下结合说明书附图及具体实施例进一步说明本发明的技术方案。 应当理解， 此 处所描述的具体实施例仅仅用以解释本发明， 并不用于限定本发明。 参照图 1， 图 1是本发明报文拥塞处理方法第一实施例流程示意图； 如图 1所示， 本发明报文拥塞处理方法包括以下步骤： 步骤 S01、 接收报文的报文分片及分片参数， 根据所述分片参数获取所述报文拥 塞处理所需的拥塞参数； 在一优选的实施例中， 接收上层相关模块发送的报文及报文分片， 所述报文分片 包括报文起始（Start of Packet, 简称为 SOP)报文分片、 报文结束（End of Packet, 简 称为 E0P) 报文分片以及其它有效报文分片， 所述相关分片参数包括报文有效信号、 报文配置长度等， 通过 SOP报文分片和 EOP报文分片之间时隙间隔获取所述报文的 真实长度； 同时根据报文有效信号到相对应的内部 RAM中获取当前报文拥塞处理所 需要的参数， 如平均队列上限长度阈值、 平均队列下限长度阈值、 队列计算权重、 平 均队列长度、 丢弃曲线斜率参数等； 当从内部 RAM中读取到当前报文拥塞处理所需 要的参数后， 将该系列参数、 报文配置长度以及通过对报文所有分片进行计数得到的 报文真实长度， 将报文真实长度、 报文分片、 报文有效信号、 报文配置长度和处理后 的相关分片信息等信息进行传送。 步骤 S02、 判断所述报文分片是否为 SOP报文分片； 若是， 则执行步骤 S03; 步骤 S03、 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度及长度 阈值， 对所述报文进行处理。 在一优选的实施例中， 接收所述到报文真实长度、 报文分片、 报文有效信号、 报 文配置长度和处理后的相关分片信息等信息后，判断所述报文中是否包含 SOP报文分 片； 如果是则用配置报文长度计算队列长度。 确定队列长度后， 根据队列长度、 队列 计算权重以及平均队列长度， 通过计算得到本次报文对应的平均队列长度。 得到本次 报文对应的平均队列长度后， 结合平均队列上限长度阈值、 平均队列下限长度阈值， 对当前报文进行丢弃判断。 具体地， 如果平均队列长度大于平均队列上限长度阈值， 则此时报文将被丢弃； 如果平均队列长度小于平均队列下限长度阈值， 则此时报文不 会被丢弃； 如果平均队列长度小于平均队列上限长度阈值， 大于平均队列下限长度阈 值， 则开始随机丢弃到来的报文。在一优选的实施例中， 采用 WRED算法进行丢弃判 断。 即根据 WRED算法计算出丢弃概率， 并将得到的丢弃概率与产生的随机数进行比 较， 如果产生的随机数小于丢弃概率， 则此报文将会被丢弃， 如果产生的随机数大于 丢弃概率， 则此报文将不会被丢弃。整个报文的所有报文分片依据 SOP分片的计算结 果通过或丢弃。 在 SOP报文分片进行 WRED计算时， 报文长度用的是配置长度而非 报文实际的长度。 SOP报文分片判断丢弃， 则这个报文的后续分片都参照 SOP分片的 判断结果全部丢弃； 如果通过， 则这个报文的后续分片都参照 SOP分片的判断结果全 部通过。 本发明通过根据接收的报文分片及分片参数， 获取所述报文拥塞处理所需的拥塞 参数， 在所述报文分片仅为 SOP报文分片时， 根据分片参数和拥塞参数获取报文的平 均队列长度及长度阈值， 对所述报文进行处理的方法， 具有高效解决通信网络报文拥 塞的有益效果， 提高了处理拥塞报文的操作速率， 极大地节省了内部缓存空间。
参照图 2， 图 2是本发明报文拥塞处理方法第二实施例流程示意图； 本实施例与 本发明报文拥塞处理方法第一实施例的区别是， 仅增加了步骤 S04; 本实施例仅对步 骤 S04作具体描述， 本发明报文拥塞处理方法所涉及的其他步骤请参照相关实施例的 具体描述， 在此不再赘述。 如图 2所示， 本发明报文拥塞处理方法在步骤 S03、 根据所述分片参数和拥塞参 数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理的步骤之后还包 括步骤： 步骤 S04、 当 EOP报文分片到达时， 根据实际报文长度与所述分片参数中的配置 报文长度之差， 校正并更新所述报文的队列长度。 本实施例报文拥塞处理方法中， 如果报文分片既是 SOP分片又是 EOP分片， 则 采用实际报文长度与配置报文长度之差， 校正并更新本次报文的队列长度； 所述实际 报文长度通过 EOP报文分片与 SOP报文分片之间时隙间隔获取。 确定实际报文长度 后， 根据该实际报文长度、 队列计算权重及报文分片的平均队列长度， 计算得到本次 报文对应的平均队列长度； 如果报文分片仅为 SOP报文分片， 则采用配置报文长度进 行后续的计算； 在对 SOP报文分片时判断该报文是否丢弃， 当收到 EOP分片时校正 队列长度。 本实施例通过当 EOP 报文分片到达时根据实际报文长度与所述分片参数中的配 置报文长度之差、校正并更新所述报文的队列长度的方法， 具有提高处理 SOP拥塞报 文的操作速率的有益效果。
参照图 3， 图 3是本发明报文拥塞处理方法中根据分片参数和拥塞参数， 获取报 文的平均队列长度， 对报文进行处理一实施例流程示意图； 本实施例仅对步骤 S03作 具体描述，本发明报文拥塞处理方法所涉及的其他步骤请参照相关实施例的具体描述， 在此不再赘述。 如图 3所示， 本发明报文拥塞处理方法中， 步骤 S03、 根据所述分片参数和拥塞 参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理的步骤具体包 括- 步骤 Sll、 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 在所述报文分片仅为 SOP报文分片时， 采用配置报文长度计算队列长度； 确定队 列长度后， 根据该队列长度、 队列计算权重及报文分片的平均队列长度， 计算得到本 次报文对应的平均队列长度。 步骤 S12、 根据所述分片参数中的报文队列号， 获取加权随机早期检测 WRED策 略组， 根据所述 WRED策略组中的队列优先级及策略组参数， 获取所述平均队列长度 的上限长度阈值和下限长度阈值； 在一优选的实施例中， 采用 WRED算法对报文拥塞进行处理。 参照图 4， 图 4是 本发明报文拥塞处理方法中 WRED工作原理示意图； 如图 4所示， 根据接收到的报文 队列号获取 WRED策略组， 根据策略组获取策略组参数， 该策略组参数包括队列计算 权重、 队列长度阈值； 根据 WRED策略组和报文优先级， 到相对应的内部 RAM中获 取当前报文拥塞处理所需要的参数， 比如平均队列上限长度阈值、 平均队列下限长度 阈值、 丢弃曲线斜率参数等。 步骤 S13、 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值， 则执行步骤 S14; 若所述平均队列长度 小于所述下限长度阈值， 则执行步骤 S15; 若所述平均队列长度不小于所述下限长度 阈值且不大于所述上限长度阈值， 则执行步骤 S16; 步骤 S14、 允许将所述 SOP报文分片丢弃； 步骤 S15、 允许所述 SOP报文分片通过； 步骤 S16、 根据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系： 若所述平均 队列长度大于所述上限长度阈值， 则允许将所述 SOP报文分片丢弃； 若所述平均队列 长度小于所述下限长度阈值， 则允许所述 SOP报文分片通过； 若所述平均队列长度不 小于所述下限长度阈值且不大于所述上限长度阈值， 则根据所述 WRED策略组， 随机 丢弃所述 SOP报文分片。 在一优选的实施例中， 所述根据 WRED策略组， 随机丢弃所述 SOP报文分片具 体包括： 产生一个随机数， 根据 WRED策略组， 计算并获取所述 SOP报文分片的丢 弃概率； 若所述随机数小于所述丢弃概率， 则允许将所述 SOP报文分片丢弃； 若所述 随机数不小于所述丢弃概率， 则允许所述 SOP报文分片通过。 步骤 S17、 根据所述 SOP报文分片的处理结果， 将所述报文的所有报文分片允许 通过或丢弃。 对报文中的所有报文分片， 都根据所述 SOP报文分片的处理结果进行处理； 如果 SOP报文分片判断丢弃，则这个报文的后续分片都参照 SOP分片的判断结果全部丢弃; 如果 SOP报文分片判断通过， 则这个报文的后续分片都参照 SOP分片的判断结果全 部通过。 本实施例通过根据分片参数和拥塞参数， 获取 SOP报文分片的平均队列长度， 根 据平均队列长度对所述报文进行处理的方法，具有高效解决通信网络报文的有益效果， 节省了内部缓存空间。
参照图 5， 图 5是本发明报文拥塞处理方法中若报文平均队列长度不小于下限长 度阈值且不大于上限长度阈值， 则根据 WRED策略组， 随机丢弃 SOP报文分片一实 施例流程示意图； 本实施例仅对上述情况进行进一步的描述。 本发明报文拥塞处理方 法所涉及的其他步骤请参照相关实施例的具体描述， 在此不再赘述。 如图 5所示， 本发明报文拥塞处理方法中若报文平均队列长度不小于下限长度阈 值且不大于上限长度阈值， 则根据 WRED策略组， 随机丢弃 SOP报文分片的步骤具 体包括： 步骤 S21、产生一个随机数， 根据所述 WRED策略组， 计算并获取所述 SOP报文 分片的丢弃概率； 在进行随机丢弃 SOP报文分片前， 先产生一个随机数 R， 并根据 WRED算法计 算丢弃概率 Pb; 对所述 WRED算法的描述请参照图 4实施例的具体描述， 在此不再 赘述。 步骤 S22、 判断所述随机数是否小于所述丢弃概率； 若是， 则执行步骤 S23 ; 若 否， 则执行步骤 S24; 步骤 S23、 允许将所述 SOP报文分片丢弃； 步骤 S24、 允许所述 SOP报文分片通过。 通过对随机数 R和丢弃概率 Pb进行仲裁判断， 随机丢弃 SOP报文分片；具体地， 如果随机数 R小于丢弃概率 Pb成立， 则判决 SOP报文分片丢弃； 如果随机数 R小于 丢弃概率 Pb不成立， 则判决 SOP报文分片不丢弃。 若 SOP报文分片丢弃， 则所述报 文丢弃； 若所述 SOP报文不丢弃， 则所述报文不丢弃。 本实施例通过产生随机数并根据 WRED策略组计算获取丢弃概率，根据随机数与 丢弃概率的关系， 随机丢弃报文的方法， 具有更进一步提高拥塞处理效率的有益效果。
参照图 6， 图 6是本发明报文拥塞处理方法第三实施例流程示意图； 本实施例与 本发明报文拥塞处理方法第二实施例的区别是， 仅增加了步骤 S05 ; 本实施例仅对步 骤 S05作具体描述， 本发明报文拥塞处理方法所涉及的其他步骤请参照相关实施例的 具体描述， 在此不再赘述。 如图 6所示， 本发明报文拥塞处理方法在步骤 S04、 当 EOP报文分片到达时， 根 据实际报文长度与所述分片参数中的配置报文长度之差， 校正并更新所述报文的平均 队列长度的步骤之后还包括步骤： 步骤 S05、 根据所述报文的处理结果， 丢弃允许丢弃的所述报文， 传送允许通过 的所述报文和更新后的所述报文的队列长度。 根据报文是否丢弃的处理结果， 将更新后的所述报文的平均队列长度传送至相关 模块， 设置为更新 RAM 内存储的平均队列长度， 同时根据丢弃判断结果， 将需要丢 弃的报文直接丢弃， 将正常的报文送入下级模块， 等待下级模块的处理。 本实施例通过根据报文处理结果， 将所述报文进行处理并将更新后的报文的平均 队列长度传送的方法， 具有更进一步地提高处理拥塞报文的操作速率、 节省内部缓存 空间的有益效果。 参照图 7， 图 7是本发明报文拥塞处理装置第一实施例结构示意图； 如图 7所示， 本发明报文拥塞处理装置包括： 拥塞输入模块 01， 设置为接收报文的报文分片及分片参数， 根据所述分片参数获 取所述报文拥塞处理所需的拥塞参数； 在一优选的实施例中， 拥塞输入模块 01 接收上层相关模块发送的报文及报文分 片， 所述报文分片包括报文起始 （Start of Packet, 简称为 SOP) 报文分片、 报文结束 (End of Packet, 简称为 EOP) 报文分片以及其它有效报文分片， 所述相关分片参数 包括报文有效信号、 报文配置长度等， 通过 SOP报文分片和 EOP报文分片之间时隙 间隔获取所述报文的真实长度；同时拥塞输入模块 01根据报文有效信号到相对应的内 部 RAM中获取当前报文拥塞处理所需要的参数， 如平均队列上限长度阈值、 平均队 列下限长度阈值、 队列计算权重、 平均队列长度、 丢弃曲线斜率参数等； 当拥塞输入 模块 01从内部 RAM中读取到当前报文拥塞处理所需要的参数后， 将该系列参数、 报 文配置长度以及通过对报文所有分片进行计数得到的报文真实长度，将报文真实长度、 报文分片、 报文有效信号、 报文配置长度和处理后的相关分片信息等信息传送至拥塞 处理模块 02。 拥塞处理模块 02，设置为在所述报文分片为 SOP报文分片时，根据所述分片参数 和拥塞参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理。 所述拥塞处理模块 02还设置为， 当 EOP报文分片到达时， 根据实际报文长度与 所述分片参数中的配置报文长度之差， 校正并更新所述报文的平均队列长度。 在一优选的实施例中， 拥塞处理模块 02接收所述到报文真实长度、报文分片、报 文有效信号、 报文配置长度和处理后的相关分片信息等信息后， 判断所述报文中是否 包含 SOP报文分片； 如果是， 则拥塞处理模块 02用配置报文长度计算队列长度； 确 定队列长度后， 拥塞处理模块 02根据队列长度、 队列计算权重以及平均队列长度， 通 过计算得到本次报文对应的平均队列长度。 得到本次报文对应的平均队列长度后， 拥 塞处理模块 02结合平均队列上限长度阈值、平均队列下限长度阈值，对当前报文进行 丢弃判断。 具体地， 如果平均队列长度大于平均队列上限长度阈值， 则拥塞处理模块 02允许将此时报文丢弃； 如果平均队列长度小于平均队列下限长度阈值， 则拥塞处理 模块 02允许此时报文不被丢弃； 如果平均队列长度小于平均队列上限长度阈值， 大于 平均队列下限长度阈值， 则拥塞处理模块 02开始随机丢弃到来的报文。在一优选的实 施例中， 拥塞处理模块 02允许采用 WRED算法进行丢弃判断。 即根据 WRED算法计 算出丢弃概率， 并将得到的丢弃概率与产生的随机数进行比较， 如果产生的随机数小 于丢弃概率，则拥塞处理模块 02允许此报文被丢弃，如果产生的随机数大于丢弃概率， 则拥塞处理模块 02允许此报文不被丢弃。 整个报文的所有报文分片依据 SOP分片的 计算结果通过或丢弃。 在 SOP报文分片进行 WRED计算时， 报文长度用的是配置长 度而非报文实际的长度。 SOP报文分片判断丢弃， 则这个报文的后续分片都参照 SOP 分片的判断结果全部丢弃； 如果通过， 则这个报文的后续分片都参照 SOP分片的判断 结果全部通过。 本实施例报文拥塞处理方法中， 如果报文分片既是 SOP分片又是 EOP分片， 则 拥塞处理模块 02采用实际报文长度与配置报文长度之差，校正并更新本次报文的队列 长度； 所述实际报文长度通过 EOP报文分片与 SOP报文分片之间时隙间隔获取。 确 定实际报文长度后，拥塞处理模块 02根据该实际报文长度、队列计算权重及报文分片 的平均队列长度， 计算得到本次报文对应的平均队列长度； 如果报文分片仅为 SOP报 文分片， 则拥塞处理模块 02采用配置报文长度进行后续的计算； 在对 SOP报文分片 时， 拥塞处理模块 02判断该报文是否丢弃； 当收到 EOP分片时， 拥塞处理模块 02校 正队列长度。 本发明通过根据接收的报文分片及相关分片参数， 获取所述报文拥塞处理所需的 相关拥塞参数， 在所述报文分片仅为 SOP报文分片时， 根据相关参数获取报文的平均 队列长度及长度阈值， 并根据所述平均队列长度及长度阈值对所述报文进行处理， 具 有高效解决通信网络报文拥塞的有益效果， 提高了处理拥塞报文的操作速率， 极大地 节省了内部缓存空间。
参照图 8， 图 8是本发明报文拥塞处理装置中拥塞处理模块一实施例结构示意图； 本实施例仅对拥塞处理模块 02作进一步描述，本发明报文拥塞处理装置所涉及的其他 模块请参照相关实施例的具体描述， 在此不再赘述。 如图 8所述， 本发明报文拥塞处理装置中， 所述拥塞处理模块 02具体包括： 参数获取子模块 021， 设置为： 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 根据所述分片参数中的报文队列号， 获取 WRED策略组， 根据所述 WRED策略 组中的队列优先级及策略组参数， 获取所述平均队列长度的上限长度阈值和下限长度 阈值； 在所述报文分片仅为 SOP报文分片时，参数获取子模块 021采用配置报文长度计 算队列长度； 确定队列长度后， 根据该队列长度、 队列计算权重及报文分片的平均队 列长度， 参数获取子模块 021计算得到本次报文对应的平均队列长度。 在一优选的实 施例中， 参数获取子模块 021采用 WRED算法对报文拥塞进行处理。 WRED工作原 理请参照图 4实施例的具体描述， 在此不再赘述； 参数获取子模块 021根据接收到的 报文队列号获取 WRED策略组， 根据策略组获取策略组参数， 该策略组参数包括队列 计算权重、 队列长度阈值； 参数获取子模块 021根据 WRED策略组和报文优先级， 到 相对应的内部 RAM中获取当前报文拥塞处理所需要的参数， 比如平均队列上限长度 阈值、 平均队列下限长度阈值、 丢弃曲线斜率参数等。 报文判决子模块 022， 设置为- 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值， 则允许将所述 SOP报文分片丢弃； 若所述平均队列长度小于所述下限长度阈值， 则允许所述 SOP报文分片通过； 若所述 平均队列长度不小于所述下限长度阈值且不大于所述上限长度阈值， 则根据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 所述报文的所有报文分片将根据所述 SOP报文分片的处理结果通过或丢弃。 报文判决子模块 022判断所述平均队列长度与所述上限长度阈值和下限长度阈值 的关系： 若所述平均队列长度大于所述上限长度阈值， 则报文判决子模块 022允许将 所述 SOP报文分片丢弃； 若所述平均队列长度小于所述下限长度阈值， 则报文判决子 模块 022允许所述 SOP报文分片通过；若所述平均队列长度不小于所述下限长度阈值 且不大于所述上限长度阈值， 则报文判决子模块 022根据所述 WRED策略组， 随机丢 弃所述 SOP报文分片。 在一优选的实施例中， 报文判决子模块 022根据 WRED策略组， 随机丢弃所述 SOP报文分片具体包括： 报文判决子模块 022产生一个随机数， 根据 WRED策略组， 计算并获取所述 SOP报文分片的丢弃概率； 若所述随机数小于所述丢弃概率， 则报文 判决子模块 022允许将所述 SOP报文分片丢弃； 若所述随机数不小于所述丢弃概率， 则报文判决子模块 022允许所述 SOP报文分片通过。报文判决子模块 022对报文中的 所有报文分片， 都根据所述 SOP报文分片的处理结果进行处理； 如果 SOP报文分片 判断丢弃， 则这个报文的后续分片都参照 SOP分片的判断结果全部丢弃； 如果 SOP 报文分片判断通过， 则这个报文的后续分片都参照 SOP分片的判断结果全部通过。 本实施例通过根据分片参数和拥塞参数， 获取 SOP报文分片的平均队列长度， 根 据平均队列长度对所述报文进行处理的方法，具有高效解决通信网络报文的有益效果， 节省了内部缓存空间。
参照图 9， 图 9是本发明报文拥塞处理装置中报文判决子模块一实施例结构示意 图； 本实施例仅对报文判决子模块 022作进一步描述， 本发明报文拥塞处理装置所涉 及的其他模块请参照相关实施例的具体描述， 在此不再赘述。 如图 9所示， 本发明报文拥塞处理装置中， 报文判决子模块 022具体包括： 长度阈值判断单元 221， 设置为判断所述平均队列长度与所述上限长度阈值和下 限长度阈值的关系； 第一结果单元 222， 设置为在所述平均队列长度大于所述上限长度阈值时， 允许 将所述 SOP报文分片丢弃； 第二结果单元 223， 设置为在所述平均队列长度小于所述下限长度阈值时， 允许 所述 SOP报文分片通过； 第三结果单元 224， 设置为在所述平均队列长度不小于所述下限长度阈值且不大 于所述上限长度阈值时， 根据所述 WRED策略组， 随机丢弃所述 SOP报文分片； 第四结果单元 225， 设置为根据所述 SOP报文分片的处理结果， 将所述报文的所 有报文分片允许通过或丢弃； 参照图 10， 图 10是本发明报文拥塞处理装置中第三结果单元一实施例结构示意 图； 如图 10所示， 所述第三结果单元 224具体包括： 随机参数获取子单元 241， 设置为产生一个随机数， 根据所述 WRED策略组， 计 算并获取所述 SOP报文分片的丢弃概率； 在进行随机丢弃 SOP报文分片前，随机参数获取子单元 241先产生一个随机数 R， 并根据 WRED算法计算丢弃概率 Pb; 对所述 WRED算法的描述请参照图 4实施例的 具体描述， 在此不再赘述。 随机参数比较子单元 242， 设置为判断所述随机数是否小于所述丢弃概率； 结果获取子单元 243，设置为在所述随机数小于所述丢弃概率时，允许将所述 SOP 报文分片丢弃； 在所述随机数不小于所述丢弃概率时， 允许所述 SOP报文分片通过。 通过随机参数比较子单元 242对随机数 R和丢弃概率 Pb进行仲裁判断， 结果获 取子单元 243随机丢弃 SOP报文分片； 具体地， 如果随机参数比较子单元 242判断随 机数 R小于丢弃概率 Pb成立， 则结果获取子单元 243允许 SOP报文分片丢弃； 如果 随机参数比较子单元 242判断随机数 R小于丢弃概率 Pb不成立， 则结果获取子单元 243允许 SOP报文分片不丢弃。若结果获取子单元 243允许 SOP报文分片丢弃， 则允 许所述报文丢弃； 若结果获取子单元 243允许所述 SOP报文不丢弃， 则允许所述报文 不丢弃。 本实施例通过产生随机数并根据 WRED策略组计算获取丢弃概率，根据随机数与 丢弃概率的关系， 随机丢弃报文， 具有更进一步提高拥塞处理效率的有益效果。
参照图 11， 图 11 是本发明报文拥塞处理装置第二实施例结构示意图。 本实施例 与本发明报文拥塞处理装置第一实施例的区别是， 仅增加了拥塞输出模块 03， 本实施 例仅对拥塞输出模块 03作具体描述，本发明报文拥塞处理装置所涉及的其他模块请参 照相关实施例的具体描述， 在此不再赘述。 如图 11所示， 本发明报文拥塞处理装置还包括： 拥塞输出模块 03， 设置为根据所述报文的处理结果， 丢弃允许丢弃的所述报文， 传送允许通过的所述报文和更新后的所述报文的平均队列长度。 拥塞输出模块 03负责接收拥塞处理模块 02处理过后的报文，根据丢弃判断结果， 将需要丢弃的报文直接丢弃， 将正常的报文送入下级模块， 等待下级模块的处理。 同 时，拥塞输出模块 03将更新后的所述报文的平均队列长度传送至相关模块， 设置为更 新 RAM内存储的平均队列长度。 本实施例通过根据报文处理结果， 将所述报文进行处理并将更新后的报文的平均 队列长度传送的方法， 具有更进一步地提高处理拥塞报文的操作速率、 节省内部缓存 空间的有益效果。 以上所述仅为本发明的优选实施例， 并非因此限制其专利范围， 凡是利用本发明 说明书及附图内容所作的等效结构或等效流程变换， 直接或间接运用在其他相关的技 术领域， 均同理包括在本发明的专利保护范围内。

Claims

权 利 要 求 书 、 一种报文拥塞处理方法， 包括以下步骤：
接收报文的报文分片及分片参数， 根据所述分片参数获取所述报文拥塞处 理所需的拥塞参数；
在所述报文分片为报文起始 SOP报文分片时，根据所述分片参数和拥塞参 数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理。 、 如权利要求 1所述的报文拥塞处理方法， 其中， 所述根据所述分片参数和拥塞 参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理的步骤 之后还包括步骤：
当报文结束 EOP报文分片到达时，根据实际报文长度与所述分片参数中的 配置报文长度之差， 校正并更新所述报文的平均队列长度。 、 如权利要求 1或 2所述的报文拥塞处理方法， 其中， 所述根据所述分片参数和 拥塞参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行处理的 步骤具体包括：
根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 根据所述分片参数中的报文队列号， 获取加权随机早期检测 WRED 策略 组，根据所述 WRED策略组中的队列优先级及策略组参数， 获取所述平均队列 长度的上限长度阈值和下限长度阈值；
判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值，则允许将所述 SOP报文分片 丢弃；
若所述平均队列长度小于所述下限长度阈值，则允许所述 SOP报文分片通 过；
若所述平均队列长度不小于所述下限长度阈值且不大于所述上限长度阈 值， 则根据所述 WRED策略组， 随机丢弃所述 SOP报文分片；
根据所述 SOP报文分片的处理结果，将所述报文的所有报文分片允许通过 或丢弃。 、 如权利要求 3所述的报文拥塞处理方法， 其中， 所述若所述平均队列长度不小 于所述下限长度阈值且不大于所述上限长度阈值， 则根据所述 WRED策略组， 随机丢弃所述 SOP报文分片的步骤具体包括：
产生一个随机数， 根据所述 WRED策略组， 计算并获取所述 SOP报文分 片的丢弃概率；
判断所述随机数是否小于所述丢弃概率；
若是， 则允许将所述 SOP报文分片丢弃；
若否， 则允许所述 SOP报文分片通过。 、 如权利要求 1或 2所述的报文拥塞处理方法， 其中， 所述当 EOP报文分片到达 时， 根据实际报文长度与所述分片参数中的配置报文长度之差， 校正并更新所 述报文的队列长度的步骤之后还包括步骤：
根据所述报文的处理结果， 丢弃允许丢弃的所述报文， 传送允许通过的所 述报文和更新后的所述报文的平均队列长度。 、 一种报文拥塞处理装置， 包括：
拥塞输入模块， 设置为接收报文的报文分片及分片参数， 根据所述分片参 数获取所述报文拥塞处理所需的拥塞参数；
拥塞处理模块， 设置为在所述报文分片为 SOP报文分片时， 根据所述分片 参数和拥塞参数， 获取所述报文的平均队列长度及长度阈值， 对所述报文进行 处理。 、 如权利要求 6所述的报文拥塞处理装置， 其中， 所述拥塞处理模块还设置为， 当 EOP报文分片到达时，根据实际报文长度与所述分片参数中的配置报文长度 之差， 校正并更新所述报文的平均队列长度。 、 如权利要求 6或 7所述的报文拥塞处理装置， 其中， 所述拥塞处理模块具体包 括- 参数获取子模块， 设置为- 根据所述分片参数和拥塞参数， 获取所述报文的平均队列长度； 根据所述分片参数中的报文队列号，获取 WRED策略组，根据所述 WRED 策略组中的队列优先级及策略组参数， 获取所述平均队列长度的上限长度阈值 和下限长度阈值； 报文判决子模块， 设置为- 判断所述平均队列长度与所述上限长度阈值和下限长度阈值的关系； 若所述平均队列长度大于所述上限长度阈值，则允许将所述 SOP报文分片 丢弃；
若所述平均队列长度小于所述下限长度阈值，则允许所述 SOP报文分片通 过；
若所述平均队列长度不小于所述下限长度阈值且不大于所述上限长度阈 值， 则根据所述 WRED策略组， 随机丢弃所述 SOP报文分片；
所述报文的所有报文分片将根据所述 SOP 报文分片的处理结果通过或丢 弃。 如权利要求 8所述的报文拥塞处理装置，其中，所述报文判决子模块具体包括: 长度阈值判断单元， 设置为判断所述平均队列长度与所述上限长度阈值和 下限长度阈值的关系；
第一结果单元， 设置为在所述平均队列长度大于所述上限长度阈值时， 允 许将所述 SOP报文分片丢弃；
第二结果单元， 设置为在所述平均队列长度小于所述下限长度阈值时， 允 许所述 SOP报文分片通过；
第三结果单元， 设置为在所述平均队列长度不小于所述下限长度阈值且不 大于所述上限长度阈值时， 根据所述 WRED策略组， 随机丢弃所述 SOP报文 分片；
第四结果单元， 设置为根据所述 SOP报文分片的处理结果， 将所述报文的 所有报文分片允许通过或丢弃；
所述第三结果单元具体包括- 随机参数获取子单元， 设置为产生一个随机数， 根据所述 WRED策略组， 计算并获取所述 SOP报文分片的丢弃概率；
随机参数比较子单元， 设置为判断所述随机数是否小于所述丢弃概率； 结果获取子单元， 设置为在所述随机数小于所述丢弃概率时， 允许将所述 SOP报文分片丢弃； 在所述随机数不小于所述丢弃概率时， 允许所述 SOP报文 分片通过。 如权利要求 6或 7所述的报文拥塞处理装置， 其中， 还包括：
拥塞输出模块， 设置为根据所述报文的处理结果， 丢弃允许丢弃的所述报 文， 传送允许通过的所述报文和更新后的所述报文的平均队列长度。
PCT/CN2013/080824 2012-08-16 2013-08-05 报文拥塞处理方法及装置 WO2014026554A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US14/422,176 US9992116B2 (en) 2012-08-16 2013-08-05 Method and device for processing packet congestion
EP13829109.1A EP2887591A4 (en) 2012-08-16 2013-08-05 METHOD AND DEVICE FOR PROCESSING A PACKAGE CHARGE

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210292183.XA CN102833159B (zh) 2012-08-16 2012-08-16 报文拥塞处理方法及装置
CN201210292183.X 2012-08-16

Publications (1)

Publication Number Publication Date
WO2014026554A1 true WO2014026554A1 (zh) 2014-02-20

Family

ID=47336132

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/080824 WO2014026554A1 (zh) 2012-08-16 2013-08-05 报文拥塞处理方法及装置

Country Status (4)

Country Link
US (1) US9992116B2 (zh)
EP (1) EP2887591A4 (zh)
CN (1) CN102833159B (zh)
WO (1) WO2014026554A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102833159B (zh) * 2012-08-16 2015-10-21 中兴通讯股份有限公司 报文拥塞处理方法及装置
WO2014201635A1 (zh) * 2013-06-19 2014-12-24 华为技术有限公司 一种消息处理的方法及基站
CN105721332A (zh) * 2014-12-03 2016-06-29 深圳市中兴微电子技术有限公司 一种基于改进的wred的拥塞控制方法和装置
CN106330761B (zh) * 2015-06-30 2020-09-15 中兴通讯股份有限公司 基于队列时延的拥塞控制方法及装置
CN108632169A (zh) * 2017-03-21 2018-10-09 中兴通讯股份有限公司 一种分片的服务质量保证方法及现场可编程逻辑门阵列
US10594617B2 (en) * 2017-09-18 2020-03-17 Mellanox Technologies, Ltd. Probabilistic normalized congestion indication based on packet size
CN111385223B (zh) * 2018-12-28 2023-08-15 阿里巴巴集团控股有限公司 流控方法、设备、系统及存储介质
CN112764892A (zh) * 2019-10-21 2021-05-07 伊姆西Ip控股有限责任公司 管理进程的方法、设备和计算机程序产品

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101170492A (zh) * 2007-11-30 2008-04-30 中兴通讯股份有限公司 一种在网络拥塞时发送分片报文的方法
CN101958833A (zh) * 2010-09-20 2011-01-26 云南省科学技术情报研究院 一种基于red的网络拥塞控制算法
CN102833159A (zh) * 2012-08-16 2012-12-19 中兴通讯股份有限公司 报文拥塞处理方法及装置

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5764641A (en) * 1995-09-08 1998-06-09 Cisco Systems, Inc. Early and integrated tail packet discard system
JPH11298490A (ja) * 1998-04-14 1999-10-29 Nec Corp 選択的パケット廃棄制御装置
CA2279728A1 (en) * 1999-08-06 2001-02-06 Spacebridge Networks Corporation Soft, prioritized early packet discard (spepd) system
US6888824B1 (en) * 2000-10-20 2005-05-03 Cisco Technology, Inc. Random early detection (RED) algorithm using marked segments to detect congestion in a computer network
US6996062B1 (en) * 2001-02-28 2006-02-07 3Com Corporation Policy-based weighted random early detection method for avoiding congestion in internet traffic
AU2002357255A1 (en) * 2001-12-14 2003-06-30 Chip Engines Single cycle weighted random early detection circuit and method
US7245586B2 (en) * 2002-08-30 2007-07-17 Lucent Technologies Inc. Buffer management based on buffer sharing across ports and per-port minimum buffer guarantee
US20040100900A1 (en) * 2002-11-25 2004-05-27 Fulcrum Microsystems, Inc. Message transfer system
CN1327677C (zh) * 2003-11-21 2007-07-18 清华大学 基于ecn且带预测验证的拥塞控制方法
US20070070907A1 (en) 2005-09-29 2007-03-29 Alok Kumar Method and apparatus to implement a very efficient random early detection algorithm in the forwarding path
US8218546B2 (en) * 2005-11-10 2012-07-10 Broadcom Corporation Interleaved processing of dropped packets in a network device
US7953002B2 (en) * 2005-11-10 2011-05-31 Broadcom Corporation Buffer management and flow control mechanism including packet-based dynamic thresholding
TWI482460B (zh) * 2008-05-30 2015-04-21 Marvell Int Ltd 網路處理器單元及其相關方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101170492A (zh) * 2007-11-30 2008-04-30 中兴通讯股份有限公司 一种在网络拥塞时发送分片报文的方法
CN101958833A (zh) * 2010-09-20 2011-01-26 云南省科学技术情报研究院 一种基于red的网络拥塞控制算法
CN102833159A (zh) * 2012-08-16 2012-12-19 中兴通讯股份有限公司 报文拥塞处理方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP2887591A4 *

Also Published As

Publication number Publication date
US20150244622A1 (en) 2015-08-27
US9992116B2 (en) 2018-06-05
CN102833159A (zh) 2012-12-19
EP2887591A1 (en) 2015-06-24
EP2887591A4 (en) 2015-08-12
CN102833159B (zh) 2015-10-21

Similar Documents

Publication Publication Date Title
WO2014026554A1 (zh) 报文拥塞处理方法及装置
US9674102B2 (en) Methods and network device for oversubscription handling
US8824294B2 (en) Source-based congestion detection and control
US7447152B2 (en) Controlling traffic congestion
JP5157577B2 (ja) 到着パケット廃棄決定を行うパケットバッファ管理装置及びその廃棄決定方法
US20070183332A1 (en) System and method for backward congestion notification in network
US11870698B2 (en) Congestion control method and apparatus, communications network, and computer storage medium
KR20090004166A (ko) 통신 시스템에서 패킷 전송 속도 제한 장치 및 방법
WO2016086551A1 (zh) 一种基于改进的wred的拥塞控制方法和装置
US20090240804A1 (en) Method and apparatus for preventing igmp packet attack
WO2018113436A1 (zh) 一种网络拥塞处理方法及装置
CN106330742B (zh) 一种流量控制的方法及网络控制器
WO2010069234A1 (zh) 终端处于空闲模式下的业务终止方法、系统和设备
WO2017097201A1 (zh) 一种数据传输方法、发送装置及接收装置
WO2013016971A1 (zh) 一种分组交换网中数据包发送和接收的方法及装置
EP2040423A1 (en) Improved utilization of data links
CN112995048B (zh) 数据中心网络的阻塞控制与调度融合方法及终端设备
CN113315720B (zh) 一种数据流控制方法、系统及设备
JP2002026965A (ja) パケット交換機の送信キュー管理システム及び管理方法
US20060209687A1 (en) Communication rate control method and device
US9001657B2 (en) Mesh network node with multiple automatic repeat requesters
WO2015024400A1 (zh) 一种路由器的拥塞避免方法及装置
US20120063339A1 (en) Method and apparatus for transmitting packet in wireless network
JP5611171B2 (ja) パケット処理方法及びパケット処理装置
CN111431812A (zh) 一种报文转发控制方法及装置

Legal Events

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

Ref document number: 13829109

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 14422176

Country of ref document: US

WWE Wipo information: entry into national phase

Ref document number: 2013829109

Country of ref document: EP