WO2010102551A1 - 应用于mp组的报文处理方法及装置 - Google Patents

应用于mp组的报文处理方法及装置 Download PDF

Info

Publication number
WO2010102551A1
WO2010102551A1 PCT/CN2010/070922 CN2010070922W WO2010102551A1 WO 2010102551 A1 WO2010102551 A1 WO 2010102551A1 CN 2010070922 W CN2010070922 W CN 2010070922W WO 2010102551 A1 WO2010102551 A1 WO 2010102551A1
Authority
WO
WIPO (PCT)
Prior art keywords
packet
fragmented
header
received
stored
Prior art date
Application number
PCT/CN2010/070922
Other languages
English (en)
French (fr)
Inventor
康泱
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2010102551A1 publication Critical patent/WO2010102551A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00—Packet switching elements
    • H04L49/90—Buffering arrangements
    • H04L49/9084—Reactions to storage capacity overflow
    • H04L49/9089—Reactions to storage capacity overflow replacing packets in a storage arrangement, e.g. pushout
    • H04L49/9094—Arrangements for simultaneous transmit and receive, e.g. simultaneous reading/writing from/to the storage element
    • 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
    • H04L47/43—Assembling or disassembling of packets, e.g. segmentation and reassembly [SAR]

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to a packet processing method and apparatus applied to an MP group. Background technique
  • the Point-to-Point Protocol has a multi-link protocol (MP) that is extended in use.
  • MP multi-link protocol
  • the principle is to bundle multiple physically independent PPP links. It behaves as a logical link, and these physical channels are called an MP group.
  • the data packets transmitted on the MP group will be fragmented first.
  • the B flag in the header fragment is set to 1.
  • the E flag in the tail fragment is set to 1.
  • the fragment packets in the same MP group will be followed.
  • the sequence number is incremented, and then the fragments are sent to each sub-link in sequence.
  • the actual delays of different links may be different.
  • the fragments sent by sequence numbers are delayed by different links. When they reach the receiving end, they cannot arrive in the order of sending, so they may be out of order. phenomenon.
  • the first fragment received will not be the header fragment. If the fragment is discarded, the first packet will not be received correctly. To ensure that the first packet is received correctly, you need to save the first or more non-header fragments, which in turn will result in a variety of exceptions.
  • the prior art simplifies the process of processing a slice after the MP is started, that is, if the first slice is received, if it is not a header slice, the slice is discarded, and the sequence number of the slice is used as a start.
  • Comparison Value for the next received fragment, if the sequence number is smaller than the initial comparison value, the fragments are discarded, and the initial comparison value remains unchanged; if the received fragment number is larger than the initial comparison value, The initial comparison value is updated to the sequence number of the currently received fragment, and the fragment corresponding to the initial comparison value is saved, and the previously saved fragment is discarded. According to the above comparison update process, if a fragment corresponding to the initial comparison value larger than the previous initial comparison value is detected as a header fragment, the header fragment is found, the startup process is ended, and the normal packet operation state is entered.
  • the prior art described above has at least the following disadvantages: when the header fragment of the data packet in which the first fragment is located arrives at the receiving end only for a time when the delay of the first fragment is small due to the delay of the link existence. Because the sequence number of the header fragment is smaller than the initial comparison value, the header fragment can only be discarded, so that the first data packet cannot be grouped, thereby affecting the correct rate of the MP group packet.
  • the embodiment of the invention provides a packet processing method and device applied to an MP group, so as to improve the correct rate of the MP group packet.
  • a packet processing method is applied to a multi-link protocol MP group, including: determining whether the received first fragmented packet is a header fragmentation packet; And storing the first fragmented packet, and receiving the subsequent fragmented packet; and receiving the fragmented packet as a header fragmentation packet, and confirming that the header fragmentation packet is working as a group packet
  • the packet fragmentation message enters the group working state according to the header fragmentation packet.
  • a message processing apparatus for a multi-link protocol MP group, including: a receiving module, configured to receive a fragmented packet sent by the MP group, and a determining module, configured to determine whether the first fragmented packet received by the receiving module is a header fragmentation packet, and a search processing module, configured to: When the determining module determines that the first fragmented packet is not a header fragmented packet, the first fragmented packet is stored, and the subsequent fragmented packet received by the receiving module is received.
  • the fragmented packet is entered according to the fragmented packet.
  • the working status of the group package may cause the first fragment packet to arrive later due to the link delay in the first packet. In this case, the embodiment can be used. Ensure that the entire packet is not discarded, improve the normal startup process of the MP group, and improve the correct rate of the MP group packet.
  • FIG. 1 is a schematic flowchart of a first embodiment of a packet processing method applied to a multi-link protocol MP group according to the present invention
  • FIG. 2 is a schematic diagram of a sequence number space according to an embodiment of the present invention
  • Figure 6 is a block diagram showing the structure of an embodiment of a text processing system applied to a multi-link protocol MP group of the present invention. detailed description
  • the embodiment of the present invention adopts a packet processing method that considers a multi-link protocol (MP) group under various conditions, so as to implement a method for starting an MP group under various conditions and completing the group work of the text. And provides automatic reset function.
  • MP multi-link protocol
  • Step 101 A receiving end receives a packet transmitted on an MP group.
  • the fragment packet has a message sequence number.
  • the packets are sent to multiple chains of the MP group according to the sequence number of the packets.
  • the transmission is performed on the road, and the message sequence number is the minimum sequence number starting from the header fragment packet of the group, and the maximum fragment number is the end fragment packet message.
  • Step 102 The receiving end determines whether the received first fragmented packet is a header fragmentation packet, and if yes, step 103 is performed; otherwise, step 104 is performed; the actual delay of different links may be different.
  • the sender is in accordance with the message number
  • the sequence is sent in ascending order, but may arrive out of order at the receiving end. That is, the first fragmented packet received by the receiving end may not be a fragmented packet. In this case, the receiving end needs to be started in the MP group. The judgment is made; the determination of the header fragmentation message can be determined by the position 1 of the B flag in the header fragment.
  • Step 103 The storage header fragmentation packet is sequentially stored, and the subsequent received fragmentation packet is sequentially stored, and the group packet working state is entered;
  • the receiving end can receive the normal fragmented packet based on the header fragmented packet when receiving the other fragmented packet sent by the MP group. If the other packet is received by the B flag of the fragmented packet, the other packet is also stored, and the packet header packet belongs to the first packet fragment. After the packet is completed, the packet to which it belongs is grouped according to the other fragmented packet.
  • Step 104 The first fragmented packet is stored, and the subsequent fragmented packet is received.
  • the header fragmentation packet is confirmed to be in the working state of the packet.
  • the packet fragmentation message enters the group working state.
  • the receiving end stores the first fragmented packet first, and then receives the subsequent fragmented packet.
  • the fragmented packet is searched for in the fragmented packet and the stored fragmented packet, so that the header packet that is subsequently found can enter the normal working state of the packet.
  • the normal working status of the packet is simply described as follows: the head pointer of the fragmented packet indicates the packet sequence number of the fragmented packet, and the packet number of the subsequent received fragment is increased by one.
  • the slice message that is, the head pointer is also incremented by one, so that the data packets are sequentially grouped in the order of the message sequence number.
  • the packet is fragmented and the packet is discarded, the abnormal packet is discarded, or the fragment packet is received.
  • the operation of the fragmented message can be preset, and will not be described in detail here.
  • the text processing method applied to the multi-link protocol MP group may cause the first packet to be caused by the link delay before the MP group is started, before entering the normal group working state.
  • the fragmented packet arrives later. In this case, the entire packet can be discarded without being discarded.
  • the normal startup process of the MP group is improved, and the correct rate of the MP group packet is improved.
  • the method for processing the multi-link protocol MP group described in the embodiment of the present invention firstly sets the normal discarding condition of the fragmented packet, and considers the following two conditions, but is not limited to the following two conditions. Normal operation of dropping fragmented packets:
  • the fragmented packet must be grouped according to the sequence of the message sequence, and then sent to the subsequent processing, that is, if the previous message is received, the receiving end is still waiting for it.
  • the subsequent packets cannot be used for grouping.
  • the subsequent packet packet will be seriously blocked.
  • the receiving end needs a timeout processing mechanism, that is, it needs to set a time. Threshold.
  • the fragmented packet related to the fragmented packet must be discarded. This time threshold can be set with reference to the maximum link delay that can be tolerated.
  • the MP slice header format can refer to RFC-1990, which stipulates that the MP slice can adopt two types of space allocation methods, that is, long serial number and short serial number mode, wherein the long serial number is represented by 24 bits, the short serial number is represented by 12 bits, and the transmitting end is used.
  • the sequence number of the fragmented message is encoded, that is, the increment is 1 from 0. If the maximum value of the serial number space is counted, the number is returned to 0 and the number is restarted.
  • sequence number of the fragmented packets sent in chronological order is incremented by one.
  • cur- sn indicates the fragmented packet of the largest packet sequence that has been discarded. If the storage space of the fragmented packet is exhausted or the waiting time of the fragmented packet exceeds the set threshold, then the ratio is cur-sn.
  • the fragmented message with the small message sequence number has been discarded.
  • the cur- sn is the current packet serial number comparison value.
  • the next received fragmented text of the receiving end is the next-sn number.
  • next-sn can be regarded as the message sequence number of the fragmented message sent by the cur-sn, and can also be regarded as the fragment before the cur-sn.
  • the message number of the message Probably, if next-sn is in the first part of the window, then next-sn can be considered as the fragment number after cur-sn.
  • next-sn is in the second part of the window, then next-sn can be considered as cur. — sn
  • the previous fragment number The fragmented packet corresponding to the packet sequence number before the cur- sn is considered to be an abnormal fragmented packet, and is discarded.
  • the fragmented packet corresponding to the packet sequence after the cur-s is considered to be a normal segment.
  • a piece of message, which is received and stored. Therefore, the window range of the abnormal fragmentation message can be set to the second partial window.
  • the size of the window of the abnormal fragmented message should be set to (cur-sn, cur-sn+min(S/2, SNmax)), that is, The minimum value of S/2 and SNmax.
  • the number of the packet sequence number of each fragmented packet is unique.
  • the sequence number of the fragmented packet may be repeated only when the sequence number is full and the sequence number of the packet is reversed.
  • the number of fragmented packets after a packet is fragmented is smaller than the number of serial numbers owned by the serial number space. Therefore, the sequence number of the fragmented packet of the same packet is duplicated. Will appear.
  • the receiving end receives the fragmented packet with the repeated sequence number, it cannot distinguish the sequence in which the sending end sends the fragmented packet, and then discards the fragmented packet that is received later or overwrites the previously received fragmented packet. This will inevitably lead to group package errors.
  • the fragmented packet whose sequence number is repeated is also an abnormal fragmented packet.
  • the packet sequence number may be the same, resulting in an abnormal fragmentation message.
  • the normal discarding condition of the fragmented packet is not limited to the present invention. Different normal discarding conditions may be specifically defined according to different situations.
  • the definition of the abnormal packet may also be defined by general probability statistics. The manner in which the message sequence number belongs to the abnormal message is not limited to the present invention.
  • FIG. 3 is a schematic diagram of a state machine of a second embodiment of a packet processing method applied to a multi-link protocol MP group according to the present invention.
  • the state machine can clearly describe the startup of the MP group under various conditions. As shown in Figure 3, the various states of the state machine are first described briefly:
  • EMPTY state In the Pl, P2, P3, and NORMAL states, when the statistics of consecutive abnormal fragmented packets reach a threshold, all fragmented packets are discarded and returned to the EMPTY state, that is, the automatic reset function.
  • P1 state In the EMPTY state, if the first fragmented packet is received as a non-header fragmented packet, it is transferred to the P1 state.
  • P2 state When the normal discarding condition is met, a number of fragmented packets are discarded, and then transferred to the P2 state.
  • P3 state In the P1 state, when the header fragmentation packet is received, the packet sequence number of the fragmented packet is larger than The smallest packet sequence number of the received fragmented packet, or the header fragmentation packet is received in the P2 state, and the header fragmentation packet sequence number is within the normal receiving window range, and the packet sequence number of the fragmentation packet is greater than After receiving the smallest packet sequence number of the fragmented packet, it enters the P3 state;
  • the state of the normal packet is the normal packet state.
  • the first fragmented packet is received, that is, the first fragmented packet is received; or in the P1, P2, and P3 states, the header fragmented packet is received and the header is received.
  • the packet sequence number of the fragment packet is within the range of the normal receiving window, and the packet sequence number of the fragment fragment packet is smaller than the smallest packet sequence number of the currently received fragment packet; or in the P3 state, the normal discard condition is met.
  • the fragmented message is discarded. Until a header fragmentation message is encountered, it enters the NORMAL state.
  • the conditions for the jump between states are as shown by the arrows in Figure 3, including:
  • Condition 1.1 The fragmented message is received as a header fragmentation message, and the header fragmentation message is stored, and the state is entered into the NORMAL state.
  • Condition 1.2 The fragmented packet is received as a non-header fragmented packet, and the fragmented packet is stored and enters the P1 state.
  • Condition 2.1 If the fragmented packet is received as a normal fragmented packet, and the non-header fragmented packet is received, the normal fragmented packet is stored and returned to the P1 state; or the fragmented packet is received. If the packet is abnormal, the abnormal fragment packet is discarded and the P1 state is returned.
  • Condition 2.2 After receiving the fragmented packet as the header fragmentation packet, and the packet fragment number of the header fragmentation packet is smaller than the smallest packet sequence number of the stored fragmented packet, the header fragmentation packet is stored and enters NORMAL. status.
  • Condition 2.3 After receiving the fragmented packet as the header fragmentation packet, and the packet fragment number of the header fragmentation packet is greater than the smallest packet sequence number of the stored fragmentation packet, the header fragmentation packet is stored, and the packet is entered into P3. status.
  • Condition 2.4 When the normal discarding condition is met, the fragmented packet corresponding to the current minimum packet sequence number is discarded. Until the normal discarding condition is not met, the P2 state is entered.
  • Condition 2.5 When the number of consecutive abnormal fragmented packets reaches a threshold, the device discards all the fragmented packets from the current minimum packet number and returns to the EMPTY state. .
  • Condition 3.1 When the normal discarding condition is met, the fragmented packet corresponding to the current minimum packet sequence number is discarded, and the normal discarding condition is not satisfied, and the packet is returned to the P2 state; or the normal fragmented packet is received. If the packet is not fragmented, the fragmented packet is stored and returned. P2 state; or receiving an abnormal fragmented packet, discarding the fragmented packet and returning to the P2 state.
  • Condition 3.3 When the number of consecutively received abnormal fragmented packets reaches a threshold, the fragmented packets corresponding to the smallest packet sequence currently stored are discarded, and all fragmented packets are discarded and returned to the EMPTY state.
  • Condition 4.2 When a normal fragmented packet is received and is a header fragmented packet, and the packet sequence number of the header fragmentation packet is smaller than the smallest packet sequence number of the stored fragmented packet, the header is stored. The packet enters the NORMAL state; or when the condition for dropping the fragmented packet is satisfied, the fragmented packet corresponding to the smallest packet sequence currently stored is discarded, and is discarded until a stored header fragment is received. Text, enter the NORMAL state.
  • Condition 4.3 When the number of consecutively received abnormal fragmented packets reaches a threshold, the fragmented packets corresponding to the smallest packet sequence currently stored are discarded, and all fragmented packets are discarded. Go back to the EMPTY state.
  • Condition 5.1 Receive a normal fragment packet in the normal packet, complete the packet processing, and return to the NORMAL state; or receive the abnormal fragment packet, discard the fragmented packet and return to NORMAL. status.
  • Condition 5.2 When the number of consecutively received abnormal fragmented packets reaches a threshold, the packet is discarded from the smallest packet sequence currently stored, and all fragmented packets are discarded and returned to the EMPTY state.
  • the normal start of the MP group that is, five conditions from EMPTY to NORMAL can be realized; there are also four conditions that return EMPTY from PI, P2, P3, and NORMAL, that is, the automatic reset function is implemented.
  • the state machine will be described in detail below.
  • the MP group has just started.
  • the receiving end receives the first fragmented packet, it is the header fragment packet of a group of packets.
  • the header packet is stored and enters the normal group working state.
  • the header fragment packet can be used to find the subsequent fragmented packets of the group of packets.
  • the MP group has just started.
  • the receiving end receives the first fragmented packet, it is not the first fragmented packet.
  • the first fragmented packet is stored and enters the P1 state. In the P1 state, the fragmented packet continues to be received. If the received fragmented packet is a normal fragmented packet and is not a fragmented fragmented packet, the packet is discarded. If the received fragmented packet is an abnormal fragmented packet, the abnormal fragmented packet is discarded; If the packet number of the header packet is smaller than the smallest packet sequence in the fragmented packet, the fragment fragment packet is stored and enters the normal packet working state. For example, a group of messages is divided into 8 pieces, and the message numbers of the fragmented messages are 0 ⁇ 7.
  • the receiving end stores the fragmented packets with the message sequence number 1, 3, and 7 in sequence.
  • the B flag is passed. The judgment may be that the header fragment packet is obtained at this time, and the packet sequence number of the header fragment packet is smaller than the minimum packet sequence number 1 of the previously received fragment packet, and the header fragment packet is stored. After the normal packet working status is entered, the subsequent packets are received and then grouped according to the sequence number.
  • the MP group has just started.
  • the receiving end receives the first fragmented packet and is not the header fragmented packet, it stores the first fragmented packet and enters the P1 state.
  • the fragmented packet continues to be received. If the received fragmented packet is a normal fragmented packet and is not a fragmented fragment, the packet is returned to the P1 state. If the received fragmented packet is abnormal. The fragmented packet discards the abnormal fragmented packet and returns to the P1 state. When it is determined that the stored fragmented packet meets the normal discarding condition, the limited storage space of the receiving end may be full or waiting for fragmentation. If the time of the packet is too long, the packet is discarded from the fragmented packet with the smallest packet number. The P2 state is entered until the normal discard condition is no longer satisfied.
  • the fragmented packet continues to be received. If the fragmented packet is in the normal state, the fragmented packet corresponding to the smallest packet sequence that has been stored is discarded. If the normal discarding condition is met, the device returns to the P2 state; or if the received fragmented packet is a normal fragmented packet and is not a fragmented fragment, the fragmented packet is stored and returned to the P2 state; or received. If the fragmented packet is an abnormal fragmented packet, the abnormal fragmented packet is discarded and returned to the P2 state.
  • the header fragment packet is received, if the packet sequence number of the header fragment packet is within the normal fragment packet window and the packet sequence number is smaller than the smallest packet in the stored fragment packet. If the serial number is used, the header packet is stored and enters the normal group working state.
  • the MP group has just started.
  • the receiving end receives the first fragmented packet and is not the header fragmented packet, it stores the first fragmented packet and enters the P1 state.
  • the fragmented packet continues to be received. If the received fragmented packet is a normal fragmented packet and is not a fragmented fragment, the packet is returned to the P1 state. If the received fragmented packet is abnormal. The fragmented packet discards the abnormal fragmented packet and returns to the P1 state. When the stored fragmented packet does not meet the normal discarding condition, the header fragmentation packet is received when the header fragmentation packet is received. If the packet number of the packet is within the normal fragmented packet window and the packet sequence number is greater than the smallest packet sequence number in the stored fragmented packet, the header fragmentation packet is stored in the P3 state.
  • the receiver may receive a fragmented packet with the sequence number of 8, that is, the latter group of packets.
  • the header fragment stores the header fragmented message into the P3 state.
  • the P3 state continues to receive the fragmented packet.
  • the packet sequence number of the fragmented packet is greater than the smallest packet sequence number of the stored fragmented packet, the packet is stored.
  • the packet returns to the P3 state. If an abnormal fragment packet is received, the abnormal fragment packet is discarded and the P3 state is returned.
  • the packet number of the other fragmented packet is smaller than the smallest packet sequence in the stored fragmented packet, the other fragmented packet is stored and entered.
  • the normal packet working state or when the received and stored fragmented packet meets the normal discarding condition again, the fragmented packet with the smallest packet sequence number in the stored fragmented packet is discarded until the storage is completed.
  • the fragmented packet with the smallest packet sequence number in the fragmented packet is the previously stored header fragmented packet, and then enters the normal packet working state. After entering the P3 state, the device continues to receive the fragmented packet, and when it receives the header fragment packet with the packet sequence number 0, it enters the normal group packet working state; or the subsequent sequence number has not been received. If the packet is a fragmented packet of 0, and the packet is in the normal state, the fragmented packet with the smallest packet number is discarded from the fragmented packet. When the fragmented packet with the smallest sequence number is the header fragment packet with the sequence number of 8 previously stored, it can also enter the normal group working state.
  • the MP group has just started.
  • the receiving end receives the first fragmented packet and is not the header fragmented packet, it stores the first fragmented packet and enters the P1 state.
  • the fragmented packet continues to be received. If the received fragmented packet is a normal fragmented packet and is not a fragmented fragment, the packet is returned to the P1 state. If the received fragmented packet is abnormal. The fragmented packet discards the abnormal fragmented packet and returns to the P1 state. When it is determined that the stored fragmented packet meets the normal discarding condition, the limited storage space of the receiving end may be full or waiting for fragmentation. If the time of the packet is too long, the packet is discarded from the fragmented packet with the smallest packet number. The P2 state is entered until the normal discard condition is no longer satisfied.
  • the fragmented packet continues to be received. If the fragmented packet is in the normal state, the fragmented packet corresponding to the smallest packet sequence that has been stored is discarded. If the normal discarding condition is met, the device returns to the P2 state; or if the received fragmented packet is a normal fragmented packet and is not a fragmented fragment, the fragmented packet is stored and returned to the P2 state; or received. If the fragmented packet is an abnormal fragmented packet, the abnormal fragmented packet is discarded and returned to the P2 state.
  • the header fragmentation packet is received, if the packet sequence number of the header fragmentation packet is within the normal fragmentation packet window and the packet sequence number is greater than the stored fragmentation packet. Minimum message order No., the header packet is stored in the P3 state.
  • the P3 state continues to receive the fragmented packet.
  • the packet sequence number of the fragmented packet is greater than the smallest packet sequence number of the stored fragmented packet, the packet is stored.
  • the packet returns to the P3 state. If an abnormal fragment packet is received, the abnormal fragment packet is discarded and the P3 state is returned.
  • the other fragmented packet is stored and entered.
  • the normal packet working state or when the received and stored fragmented packet meets the normal discarding condition again, the fragmented packet with the smallest packet sequence number in the stored fragmented packet is discarded until the storage is completed.
  • the fragmented packet with the smallest packet sequence number in the fragmented packet is the previously stored header fragmented packet, and then enters the normal packet working state.
  • the text processing method applied to the multi-link protocol MP group provided in this embodiment may be that the first packet may arrive later in the packet due to the link delay after the MP group is just started. In this case, the entire packet can be discarded.
  • the receiver can ensure that the receiver does not discard the normal fragmented packet. A large number of abnormal fragmented packets appear in the received fragmented packets, which also ensures timely discovery and availability.
  • the function of automatic reset ensures that the MP group can start normally.
  • the embodiment of the invention improves the normal startup process of the MP group, and improves the correct rate of the MP group package.
  • the apparatus includes: a receiving module 41, a determining module 42, a storage processing module 43, and a lookup processing module 44.
  • the receiving module 41 is configured to receive the fragmented packet sent by the MP group, and the determining module 42 is configured to determine whether the first fragmented packet received by the receiving module 41 is a header fragmented packet, when determining the first
  • the notification storage processing module 43 sequentially stores the subsequent fragmentation packet received by the receiving module 41, and when determining that the first fragmentation packet is not a header fragmentation packet,
  • the notification search processing module 44 stores the first fragmented packet and stores the subsequent fragmented packet received by the receiving module 41.
  • the storage processing module 43 is configured to determine, in the determining module 42, that the first fragmented packet is the header.
  • the search processing module 44 is configured to determine the first one in the determining module 42.
  • the first fragmented packet is stored, and the subsequent fragmented packet is received by the receiving module 41, and the received fragmented packet is a header fragmented packet.
  • the packet packet enters the packet working manner according to the fragmented packet.
  • Each of the fragmented packets has a packet sequence number. After the packet is fragmented into fragments, the packet is sent to the MP group for transmission according to the sequence number of the packet, and the header fragment packet is used. The start and end fragmentation messages are ended.
  • the packet processing apparatus applied to the multi-link protocol MP group provided in this embodiment may be placed in the device at the receiving end of the packet, or connected to the receiving end as a separate device; the MP group normally starts each setting device.
  • the specific functions of the module are as described in the foregoing embodiment of the normal startup method of the MP group, and details are not described herein again.
  • the file processing device applied to the multi-link protocol MP group provided in this embodiment may arrive at the first packet after the MP group is started, and the header packet may arrive later. In this case, the entire packet is not discarded, and the normal startup process of the MP group is improved, and the correct rate of the MP group packet is improved.
  • FIG. 5 is a schematic structural diagram of a second embodiment of a message processing apparatus applied to a multi-link protocol MP group according to the present invention.
  • the apparatus includes the receiving module 41, the judging module 42, the storage processing module 43, and the lookup processing module 44 described in the above embodiments.
  • the search processing module 44 may specifically include: the first storage unit 441 is configured to: when the determining module 42 determines that the first fragmented packet is not a header fragmentation packet, store the first fragmented packet, and receive the receiving module 41. And receiving the subsequent fragmented packet; the first processing unit 442, configured to: when it is confirmed that the fragmented packet received by the first storage unit is a header fragmentation packet, and confirm that the header fragmentation packet is When the packet header packet is required to enter the working state of the packet, the packet fragment enters the packet working state according to the header fragment packet.
  • the first processing unit 442 confirms that the subsequent fragmented packet received by the first storage unit 441 is the header fragmentation packet according to the B flag position 1, and if the packet sequence number of the header fragmentation packet is smaller than the stored fraction The smallest packet sequence number in the fragment packet, notifying the first storage unit 441 to store the header fragment packet, and entering a normal group packet working state; if the packet fragment number of the header fragment packet is larger than the stored The smallest packet sequence number in the fragmented packet is sent to the first storage unit 441 to store the header fragment packet and continue to receive the subsequent fragment packet; when the subsequent fragment received by the first storage unit 441 is confirmed When the message is the other fragmented packet, if the message sequence number of the other fragmented packet is smaller than the smallest packet sequence number in the stored fragmented packet, the first storage unit 441 is notified to store the other packet. A fragmented packet continues to receive subsequent fragmented packets and enters the group working state.
  • the lookup processing module 44 may further include: a discarding processing unit 443, configured to be in the first storage unit When the fragmented packet of the 441 packet satisfies the normal discarding condition, a certain number of stored fragmented packets are discarded from the fragmented packet with the smallest packet sequence number.
  • the packet sequence number is the smallest from the stored fragmented packet. The fragmented packet begins to discard a certain number of fragmented packets, and then continues to store the fragmented packet and finds the header fragmented packet.
  • the first processing unit 442 finds that the packet sequence number is greater than the stored one. After the fragmented packet of the smallest packet sequence number in the fragmented packet, the fragmentation packet that the first storage unit 441 continues to store satisfies the normal discarding condition, and then the packet is received from the stored fragmented packet. The fragmented packet with the smallest sequence number is discarded. The fragmented packet that has been discarded until the smallest fragment of the packet is discarded is the stored fragment fragment packet.
  • the storage processing module 43 may specifically include: the second storage unit 431 is configured to: when the determining module 42 determines that the first fragmented packet is a header fragmentation packet, store the header fragmentation packet, and sequentially store the receiving module 41. The subsequent fragmented packet is received; the second processing unit 432 is configured to perform normal grouping work on the fragmented packet stored by the second storage unit 431.
  • the packet header and the packet data in the packet are stored as a whole, and the packet header and the packet data in the packet are separately stored, that is, respectively. Stores the packet header and packet data of the packet.
  • the search processing module 44 may further include: an exception processing module 444, configured to: when detecting that the fragmented packet stored by the first storage unit 441 is an abnormal fragmented packet, discarding the fragmented packet; When the number of the abnormal fragmented packets reaches a threshold, the fragmented packet stored in the storage unit 441 discards all the fragmented packets stored in the first storage unit 441, and restarts the receiving module 41 to receive the transmitted points on the MP group. Piece of message.
  • an exception processing module 444 configured to: when detecting that the fragmented packet stored by the first storage unit 441 is an abnormal fragmented packet, discarding the fragmented packet; When the number of the abnormal fragmented packets reaches a threshold, the fragmented packet stored in the storage unit 441 discards all the fragmented packets stored in the first storage unit 441, and restarts the receiving module 41 to receive the transmitted points on the MP group. Piece of message.
  • the file processing device applied to the multi-link protocol MP group provided in this embodiment may arrive at the first packet after the MP group is started, and the header packet may arrive later. In this case, the entire packet can be discarded.
  • the receiver can ensure that the receiver does not discard the normal fragmented packet.
  • a large number of abnormal fragmentation messages appear in the received fragmented packets, which also ensures timely discovery and provides automatic reset function, thus ensuring that the MP group can be started normally.
  • the embodiment of the present invention improves the normal startup process of the MP group, and improves the correct rate of the MP group package.
  • FIG. 6 is a schematic structural diagram of an embodiment of a message processing system applied to a multi-link protocol MP group according to the present invention.
  • the system includes: a sending end 61, configured to perform packet processing on the packet, and send the fragmented fragment 4 to the MP group for transmission, where the MP group is bundled with multiple physically independent PPP links.
  • the following logical link the receiving end 62 is configured to receive the fragmented packet transmitted on the MP group; and the MP group normal startup setting device 63 as shown in FIG. 4 or FIG. 5, which can be placed at the receiving end of the packet 62
  • the existing device is connected to the receiving end 62 as a separate device for processing the fragmented text received by the receiving end 62.
  • each device in the packet processing system applied to the multi-link protocol MP group provided in this embodiment is as described in the foregoing embodiment of the packet processing method applied to the multi-link protocol MP group. This will not be repeated here.
  • the embodiment of the present invention improves the normal startup process of the MP group, and improves the correct rate of the MP group package.
  • the machine may be in a storage medium, and when executed, the program may include the flow of an embodiment of the methods as described above.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

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

Abstract

本发明实施例涉及一种应用于多链路协议MP组的报文处理方法及装置。本发明实施例的应用于多链路协议MP组的报文处理方法包括：判断接收到的第一个分片报文是否为头分片报文； 若不是，存储第一个分片报文，并接收后续的分片报文；当接收到的分片报文为头分片报文，并确认头分片报文为进入组包工作状态所需的头分片报文时，根据头分片报文进入组包工作状态。本发明实施例在MP组刚刚启动后，对于第一个报文可能由于链路延时而导致其头分片报文稍晚到达，在这种情况下，通过本实施例可以保证整个报文不会被丢弃，对MP组正常启动的过程进行完善，提高MP组组包的正确率。

Description

应用于 MP组的报文处理方法及装置
本申请要求于 2009年 3月 9日提交中国专利局、申请号为 200910127218.2、 发明名称为"应用于 MP组的报文处理方法及装置"的中国专利申请的优先权， 其全部内容通过引用结合在本申请中。 技术领域
本发明涉及通信技术领域，特别涉及一种应用于 MP组的报文处理方法及 装置。 背景技术
点对点协议（ Point-to-Point Protocol, PPP )在使用中产生了一种扩展特性 的多链路协议 ( Multilink Protocol, MP ), 其原理是将多个物理上独立的 PPP 链路通过捆绑， 对外表现为一条逻辑链路， 这几个物理通道称为一个 MP组。 在 MP组上传送的数据包， 会先被分片， 头分片中的 B标志位置 1， 尾分片中 的 E标志位置 1，同一个 MP组中的所有的分片数据包，会按照递增顺序编号， 然后按序号先后将分片发送到各个子链路上传送。 一个 MP组内， 由于绑定的 不同链路的实际时延可能存在不同，按序号先后发出的分片经过不同链路的延 时， 到达接收端时不能按照发送先后到达， 因此可能产生乱序现象。 而在 MP 组刚刚启动工作过程中， 由于乱序现象的存在，接收到的第一个分片往往不会 是头分片， 如果丢弃该分片， 则第一个包必然不能正确接收， 如要保证第一个 包能正确接收， 则需要保存第一个甚至更多的非头分片，这又将带来^ ί艮多种异 常情况。 现有技术对 MP刚刚启动后对分片进行处理的过程进行了简化，即收到第 一个分片如果不是头分片， 则将此分片丢弃， 并将此分片的序号作为起始比较 值； 对接下来收到的分片， 如果序号比起始比较值小， 则丢弃这些分片， 且起 始比较值保持不变； 如果接收到的分片序号比起始比较值大， 则将起始比较值 更新为当前接收到的分片的序号， 并保存该起始比较值对应的分片，将之前保 存的分片丢弃。 当按照上述比较更新过程，如果检测到比前一起始比较值大的 起始比较值对应的分片为一个头分片， 则找到头分片， 结束启动过程， 进入正 常的组包工作状态。 以上描述的现有技术至少存在如下缺点：当由于链路存在的延时而导致第 一个分片所在数据包的头分片仅仅比第一个分片延迟很小的一个时间到达接 收端时， 由于该头分片的序号比起始比较值小， 因此该头分片也只能被丢弃， 这样第一个数据包就不能组包， 从而影响 MP组组包的正确率。
发明内容
本发明实施例提供一种应用于 MP组的报文处理方法和装置， 以提高 MP 组组包的正确率。 根据本发明实施例的一方面，提供了一种应用于多链路协议 MP组的报文 处理方法， 包括： 判断接收到的第一个分片报文是否为头分片报文； 若不是， 存储所述第一个分片报文， 并接收后续的分片报文； 当接收到的分片报文为头分片报文，并确认所述头分片报文为进入组包工 作状态所需的头分片报文时， 根据所述头分片报文进入所述组包工作状态。 根据本发明实施例的另一方面，提供了一种应用于多链路协议 MP组的报 文处理装置， 包括： 接收模块， 用于接收 MP组上传送的分片报文； 判断模块，用于判断所述接收模块接收到的第一个分片报文是否为头分片 报文； 查找处理模块，用于在所述判断模块判断所述第一个分片报文不是头分片 报文时，存储所述第一个分片报文， 并接收所述接收模块接收到的后续的分片 报文，及当接收到的分片报文为头分片报文， 并确认所述头分片报文为进入组 包工作状态所需的头分片报文时， 根据所述分片报文进入所述组包工作状态。 本发明实施例的技术方案， 在 MP组刚刚启动后， 对于第一个报文可能由 于链路延时而导致其头分片报文稍晚到达，在这种情况下，通过本实施例可以 保证整个报文不会被丢弃， 对 MP组正常启动的过程进行完善， 提高 MP组组 包的正确率。 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施 例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地， 下面描述 中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付 出创造性劳动性的前提下， 还可以根据这些附图获得其他的附图。 图 1为本发明的应用于多链路协议 MP组的报文处理方法的第一实施例的 流程示意图； 图 2为本发明实施例的序列号空间示意图； 图 3为本发明的应用于多链路协议 MP组的报文处理方法的第二实施例的 状态机示意图； 图 4为本发明的应用于多链路协议 MP组的报文处理装置的第一实施例的 结构示意图； 图 5为本发明的应用于多链路协议 MP组的报文处理装置的第二实施例的 结构示意图；
图 6为本发明的应用于多链路协议 MP组的 文处理系统的实施例的结构 示意图。 具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清 楚、 完整地描述， 显然， 所描述的实施例仅仅是本发明一部分实施例， 而不是 全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造 性劳动前提下所获得的所有其他实施例， 都属于本发明保护的范围。 本发明实施例采用了一种考虑多种条件下的多链路协议 ( MP )组的报文 处理方法， 以实现多种条件下均可以正常启动 MP组， 并完成 文的组包工作 的方法， 并且提供了自动复位的功能。 图 1为本发明的应用于多链路协议 MP组的报文处理方法的第一实施例的 流程示意图， 如图 1所示， 包括如下步骤： 步骤 101、 接收端接收 MP组上传送的分片报文； 其中，各个分片报文具有报文序号，在一组报文被分片为各个分片报文后， 将按照报文序号顺序递增的方式被发送到 MP组的多条链路上进行传送，且报 文序号为以该组的头分片报文为起始即最小序号、尾分片报文为结束即最大序 号。 步骤 102、接收端判断接收到的第一个分片报文是否为头分片报文，若是， 则执行步骤 103; 否则， 执行步骤 104; 由于绑定的不同链路实际时延可能存在不同，虽然发送端是按照报文序号 的递增顺序发送的，但到达接收端时可能就会产生乱序， 即接收端接收到的第 一个分片报文可能并非头分片报文，此时就需要接收端在 MP组刚刚启动时进 行判断； 对于头分片报文的确定可以通过头分片中的 B标志位置 1来判断。
步骤 103、 存储头分片报文并依次存储后续接收的分片报文， 进入组包工 作状态；
若最先到达接收端的分片报文即为头分片报文，则接收端再接收到 MP组 上传送的其他分片报文时， 就可以基于该头分片报文进行正常的组包工作； 若 在正常的组包工作中， 通过分片报文的 B标志位判断又接收到另一头分片报 文时， 同样将其存储， 待第一个头分片报文所属数据包组包完成后， 再根据该 另一头分片报文对其所属的数据包进行组包。
步骤 104、 存储第一个分片报文， 并接收后续的分片报文， 当接收到的分 片报文为头分片报文，并确认头分片报文为进入组包工作状态所需的头分片报 文时， 根据头分片报文进入组包工作状态。
若由于链路实际时延， 导致接收端接收到的第一个分片报文非头分片报 文， 则接收端会先将该第一分片报文存储， 然后再从后续接收到的分片报文和 已存储的分片报文中查找头分片报文，从而可以基于后续查找到的头分片报文 进入正常的组包工作状态。
其中正常的组包工作状态简单描述为：指向分片报文的头指针指示头分片 报文的报文序号，并在后续接收到的分片报文中查找报文序号递增加一的分片 报文， 即此时头指针也递增加一， 从而依次将数据包按报文序号的顺序组包。 在正常的组包工作状态下，也会出现针对头分片报文一直组包不成功而超时并 正常丢弃分片报文、丢弃异常报文、或是又接收到另一头分片报文的处理等情 况， 此时对分片报文的操作可预先设定， 在此不再详述。
本实施例提供的应用于多链路协议 MP组的 文处理方法，在 MP组刚刚 启动后，进入正常的组包工作状态前，对于第一个报文可能由于链路延时而导 致其头分片报文稍晚到达，在这种情况下，通过本实施例可以保证整个报文不 会被丢弃， 对 MP组正常启动的过程进行完善， 提高 MP组组包的正确率。
在具体介绍本发明实施例描述的应用于多链路协议 MP组的 ^艮文处理方 首先设定分片报文的正常丢弃条件，如考虑以下两个，但不限于以下两个 条件带来的正常丢弃分片报文的操作：
( 1 )发送端将报文分片后， 不同的分片报文经过不同的链路的延迟， 接 收端接收的先后顺序和各分片报文的报文序号的先后顺序不一致，这就导致了 进行一定容量的分片报文的存储。 然而任何接收端的存储空间都不是无限大 的， 因此在某些情况下， 当接收的乱序的分片报文超出接收端的存储容量时， 将对分片报文进行正常丢弃处理；
( 2 )对接收端来说， 必须按照报文序号的顺序对分片报文进行组包， 然 后再送往后续进行相应处理， 即如果对于前面的一个报文，接收端还在等待其 所有的分片报文都到达的时候，即便后面的一个报文的所有的分片报文都已经 到达， 后面的报文也不能进行组包的工作。 那么， 当前面的一个报文的各个分 片报文之间等待的时间过长， 就会对后面的报文组包造成严重的阻塞。特别是 当链路出现异常，导致前面的报文中的部分分片报文丢失， 则丢失的分片报文 将永远等不到。 因此， 接收端就需要一个超时处理机制， 即需要设定一个时间 门限， 当等待分片报文超过了此时间按门限的值， 则对与等待的分片报文相关 的分片报文必须进行丢弃处理。这个时间门限可以参考可以容忍的最大链路延 时来设定。
另外， 还需要对异常分片报文进行界定：
( 1 )序号不重复的异常分片。 MP分片头格式可参考 RFC- 1990, 其中规 定了 MP分片可以采用两种序号空间分配方式， 即长序号和短序号方式， 其中 长序号用 24比特表示，短序号用 12比特表示，发送端根据时间先后顺序为分 片报文编报文序号， 即从 0开始递增加 1， 如计数到序号空间的最大值， 则回 到 0重新开始编号。 图 2为本发明实施例的序列号空间示意图， 以 12比特标 识的短序号为例， 最大序号空间为 212 = 4096， 序号空间为一个封闭的循环。
如图 2所示， sn=0为序号空间的起点， 发送分片报文时， 按时间先后发 送的分片报文的报文序号递增加一，循环一圏后， 当 sn=4095时，接下来的 sn 将变回 0。假设 cur— sn表示当前已丢弃的最大报文序号的分片报文， 如果存储 分片报文的存储空间耗尽或者分片报文的等待时间超出设定的阔值， 则比 cur— sn报文序号小的分片报文已经被丢弃， cur— sn即为当前的报文序号比较值； 接收端下一个接收的分片 文的 4艮文序号为 next— sn。 此时可以把整个序号空 间分为两个部分， 第一部分空间的范围为 [cur— sn， cur— sn+2048]， 第二部分即 图中阴影部分范围为 [0， cur— sn]以及 [cur— sn+2048， 4095]。 从图中可以分析， 由于序号空间是循环的闭环结构， 因此 next— sn可以认为是 cur— sn后发的分片 报文的报文序号， 同时也可以认为是在 cur— sn之前的分片报文的报文序号。 从概率上来说，如果 next— sn落在第一部分窗口，则可以认为 next— sn为 cur— sn 之后的分片序号，如果 next— sn落在第二部分窗口，则可以认为 next— sn为 cur— sn 之前的分片序号。 对于 cur— sn之前的报文序号对应的分片报文， 认为是异常 分片报文， 进行丢弃处理； 而对于 cur— sn之后的报文序号对应的分片报文， 认为是正常的分片报文，将其接收下来并存储。 因此该异常分片报文的窗口范 围可以设定为第二部分窗口。
另外，该异常分片报文的窗口范围的设定还需要考虑 MP组的带宽和可容 忍链路延时指标， 如果带宽为 B(bit/s)， 链路延时指标为 T(s)， 最小分片报文 的长度为 N个字节， 则在 T时间内可收到的最大分片报文的数量 SNmax = B χ Τ / (Ν χ 8)个。因此，可以认为接收的分片报文的数量如果超过了 SNmax个， 则必有异常分片存在。此种情况下，异常分片报文的窗口应该设置为（cur— sn， cur— sn+SNmax λ
综合上述两方面考虑， 如果分片报文的序号空间为 S， 则异常分片报文的 窗口的大小应该设置为（cur— sn， cur— sn+min(S/2， SNmax) ), 即 S/2和 SNmax 的最小值。
( 2 )序号重复的异常分片。 由于发送端对每个分片报文的报文序号进行 编号都是唯一的， 只有当序号空间计满， 报文序号发生翻转时， 分片报文的报 文序号才可能重复。 而一般情况下， 一个报文被分片后的分片报文的数量会小 于序号空间所拥有的序号的数量， 因此， 同一个报文的分片报文的报文序号重 复的情况是不会出现的。 当接收端接收到报文序号重复的分片报文时， 无法分 辨出发送端发送分片报文的先后顺序，则需要丢弃后来接收到的分片报文或者 覆盖之前接收到的分片报文，这必然会导致组包出错。 因此该报文序号重复的 分片报文也为异常分片报文。 另外， 当链路传输中产生误码时， 也可能会导致 报文序号相同， 从而产生异常分片报文。 需要指出的是， 上述介绍的分片报文的正常丢弃条件并非对本发明的限 定，根据不同的情况可以具体限定不同的正常丢弃条件； 异常报文的界定也可 以通过一般的概率统计来界定具有哪些报文序号的属于异常报文，本发明实施 例中描述的方式并非对本发明的限定。
图 3为本发明的应用于多链路协议 MP组的报文处理方法的第二实施例的 状态机示意图。 通过该状态机可以清楚的描述各种条件下 MP组启动的情况。 如图 3所示， 首先对该状态机的各种状态进行简单说明：
EMPTY状态： 初始化状态， 在 Pl、 P2、 P3、 NORMAL状态下， 当连续 异常分片报文统计达到阔值时， 则丢弃所有分片报文， 回到该 EMPTY状态， 即自动复位功能；
P1状态： 在 EMPTY状态下， 如接收到第一个分片报文为非头分片报文， 即转入 P1状态；
P2状态： 当满足正常丢弃条件， 若干分片报文被丢弃， 则转入 P2状态； P3状态： 在 P1状态下， 当接收到头分片报文且该分片报文的报文序号大 于已接收到的分片报文的最小报文序号， 或者在 P2状态下接收到头分片报文 且该头分片报文序号在正常接收窗口范围内且该分片报文的报文序号大于已 接收到分片报文的最小报文序号， 则进入 P3状态；
NORMAL状态： 即正常组包状态， 在 EMPTY状态下， 接收到第一个分 片报文即头分片报文； 或者在 Pl、 P2、 P3状态下， 收到头分片报文且该头分 片报文的报文序号在正常接收窗口范围内，且该头分片报文的报文序号小于当 前已接收的分片报文的最小报文序号；或者在 P3状态下， 满足正常丢弃条件， 进行分片报文丢弃， 一直到遇见一个头分片报文， 则进入 NORMAL状态。 各个状态间跳转的条件如图 3中的箭头所示， 包括：
( 1 )条件 1.1 : 接收到分片报文为头分片报文， 存储该头分片报文， 进入 NORMAL状态。
条件 1.2: 接收到分片报文为非头分片报文， 存储该分片报文， 进 入 P1状态。
( 2 )条件 2.1 : 接收到分片报文为正常的分片报文， 且非头分片报文， 则 存储该正常的分片报文， 回到 P1 状态； 或者接收到分片报文为异 常分片报文， 则丢弃该异常分片报文， 回到 P1状态。
条件 2.2: 接收到分片报文为头分片报文， 且头分片报文的报文序 号小于已存储分片报文的最小报文序号， 则存储该头分片报文， 进 入 NORMAL状态。
条件 2.3: 接收到分片报文为头分片报文， 且头分片报文的报文序 号大于已存储分片报文的最小报文序号， 则存储该头分片报文， 进 入 P3状态。
条件 2.4: 当满足正常丢弃条件， 则从当前最小报文序号对应的分 片报文开始丢弃， 一直到正常丢弃条件不满足， 进入 P2状态。 条件 2.5: 当连续异常分片报文的数量统计达到阔值， 则从当前最 小报文序号对应的分片报文开始丢弃， 一直将存储的所有分片报文 都丢弃完毕， 回到 EMPTY状态。
( 3 )条件 3.1 : 当满足正常丢弃条件， 则从当前最小报文序号对应的分片 报文开始丢弃， 一直到正常丢弃条件不满足， 回到 P2状态； 或者 接收到正常分片报文， 且非头分片报文， 则存储该分片报文， 回到 P2状态； 或者接收到异常分片报文， 则丢弃该分片报文， 回到 P2 状态。
条件 3.2: 当接收到头分片报文， 且头分片报文的报文序号在正常 接收窗口之内， 且小于已存储分片报文的最小报文序号， 则存储该 头分片报文， 进入 NORMAL状态。
条件 3.3: 当连续接收异常分片报文的数量统计达到阔值， 则从当 前存储的最小报文序号对应的分片报文开始， 丢弃所有分片报文， 回到 EMPTY状态。
条件 3.4: 当接收到正常的分片报文， 且为头分片报文， 且该分片 报文的报文序号大于已存储的分片报文的最小报文序号， 则存储该 头分片报文， 进入 P3状态。
件 4.1 : 当接收到正常的分片报文， 且该分片报文的报文序号大于 已存储的分片报文的最小报文序号， 则存储该分片报文， 回到 P3 状态； 或者接收到异常分片报文， 则丢弃异常分片报文， 回到 P3 状态。
条件 4.2: 当接收到正常的分片报文， 且为头分片报文， 且该头分 片报文的报文序号小于已存储分片报文的最小报文序号， 则存储该 头分片报文， 进入 NORMAL状态； 或者当正常丢弃分片报文的条 件满足， 则从当前存储的最小报文序号对应的分片报文开始丢弃， 一直丢弃到遇见一个已存储的头分片报文， 进入 NORMAL状态。 条件 4.3: 当连续接收异常分片报文的数量统计达到阔值， 则从当 前存储的最小报文序号对应的分片报文开始， 丢弃所有分片报文， 回到 EMPTY状态。
( 5 )条件 5.1 : 在正常组包中接收到正常的分片报文， 完成组包处理， 回 到 NORMAL状态； 或者接收到异常分片报文， 则丢弃该分片报文， 回到 NORMAL状态。
条件 5.2: 当连续接收异常分片报文的数量统计达到阔值， 则从当 前存储的最小报文序号分片开始，丢弃所有分片报文，回到 EMPTY 状态。
该 MP组的正常启动，即从 EMPTY到 NORMAL共有 5种条件可以实现； 另外还有 4种条件使得从 PI、 P2、 P3和 NORMAL返回 EMPTY, 即自动复 位功能的实现。 下面将 ^居该状态机加以详细介绍。
第一种 EMPTY到 NORMAL的情况（ 1.1 )
MP组刚刚启动， 接收端接收到第一个分片报文即为一组报文的头分片报 文时， 存储该头分片报文并进入正常的组包工作状态， 即有了该头分片报文， 就可以找到该组报文的后续的分片报文进行组包的工作。
第二种 EMPTY到 NORMAL的情况（ 12—>22 )
MP组刚刚启动， 接收端接收到第一个分片报文不是头分片报文， 则存储 该第一个分片报文， 进入 P1状态； 在 P1状态下， 继续接收分片报文， 如果接 收到的均为正常的分片报文且非头分片报文， 则存储，如果接收到的分片报文 为异常分片报文， 则丢弃该异常分片报文； 当接收到头分片报文时， 若该头分 片报文的报文序号小于已存储的分片报文中的最小报文序号，则存储该头分片 报文并进入正常的组包工作状态。 例如： 一组报文被分为 8片， 分片报文的报 文序号分别为 0~7， 如果接收到的分片报文的报文序号依次为 1、 3、 7、 0、 2、 5、 6、 4， 那么接收端在接收到报文序号为 1、 3、 7的分片报文时依次存储， 当接收到报文序号为 0的分片报文时， 通过对 B标志位的判断可知此时获取 的为头分片报文，且该头分片报文的报文序号小于之前接收到的分片报文的最 小报文序号 1， 则存储该头分片报文并进入正常的组包工作状态， 即接收到后 续报文后按照序号顺序组包。
第三种 EMPTY到 NORMAL的情况（ 1.2— >2.4— >3.2 )
MP组刚刚启动， 接收端接收到第一个分片报文不是头分片报文， 则存储 该第一个分片报文， 进入 P1状态。
在 P1状态下， 继续接收分片报文， 如果接收到的均为正常的分片报文且 非头分片报文， 则存储并回到 P1状态； 如果接收到的分片报文为异常分片报 文， 则丢弃该异常分片报文并回到 P1状态； 当判断存储的分片报文满足正常 的丢弃条件时，即可能接收端有限的存储空间已满或是等待头分片报文的时间 过长， 则从当前已存储的最小报文序号对应的分片报文开始丢弃， 一直到不再 满足正常的丢弃条件， 进入 P2状态。
在 P2状态下， 继续接收分片报文， 如果存储的分片报文又满足正常的丢 弃条件时， 则从当前已存储的最小报文序号对应的分片报文开始丢弃， 一直到 不再满足正常的丢弃条件为止， 回到 P2状态； 或者若接收到的均为正常的分 片报文且非头分片报文， 则存储该分片报文， 回到 P2状态； 或者接收到的分 片报文为异常分片报文， 则丢弃该异常分片报文， 回到 P2状态。
当后续再接收中，接收到头分片报文， 若该头分片报文的报文序号在正常 的分片报文窗口之内且报文序号小于已存储的分片报文中的最小报文序号，则 存储该头分片报文并进入正常的组包工作状态。 第四种 EMPTY到 NORMAL的情况（ 1.2—>2.3—>4.2 )
MP组刚刚启动， 接收端接收到第一个分片报文不是头分片报文， 则存储 该第一个分片报文， 进入 P1状态。
在 P1状态下， 继续接收分片报文， 如果接收到的均为正常的分片报文且 非头分片报文， 则存储并回到 P1状态； 如果接收到的分片报文为异常分片报 文， 则丢弃该异常分片报文并回到 P1状态； 当存储的分片报文还没有满足正 常的丢弃条件时， 就接收到头分片报文时，若该头分片报文的报文序号在正常 的分片报文窗口之内且报文序号大于已存储的分片报文中的最小报文序号，则 存储该头分片报文进入 P3状态。 如第二种 EMPTY到 NORMAL的情况下的 具体举例，当在分片报文的报文序号为 0~7的一组报文后还有一组分片报文的 报文序号为 8~15的一组报文， 则在接收端依次接收到报文序号为 1、 3、 7的 分片报文后，可能会接收到报文序号为 8的分片报文， 即后一组报文的头分片 存储该头分片报文进入 P3状态。
在 P3状态下， 继续接收分片报文， 当接收到正常的分片报文， 且该分片 报文的报文序号大于已存储的分片报文的最小报文序号， 则存储该分片报文， 回到 P3状态；或者接收到异常分片报文，则丢弃异常分片报文，回到 P3状态。 当又接收到另一头分片报文时，若该另一头分片报文的报文序号小于已存储的 分片报文中的最小报文序号，则存储该另一头分片报文并进入正常的组包工作 状态； 或者当接收并存储的分片报文再次满足正常的丢弃条件时，会从已存储 的分片报文中报文序号最小的分片报文开始丢弃，直到存储的分片报文中报文 序号最小的分片报文为之前存储的头分片报文， 则进入正常的组包工作状态。 在进入到 P3状态下， 继续接收分片报文， 当后续又接收到报文序号为 0的头 分片报文时， 则进入正常的组包工作状态； 或者后续一直没有接收到报文序号 为 0的头分片报文，且又满足正常的丢弃条件时， 则从已存储的分片报文中报 文序号最小的分片报文开始丢弃，直到存储的分片报文中报文序号最小的分片 报文为之前存储的报文序号为 8的头分片报文时，也可以进入正常的组包工作 状态。
第五种 EMPTY到 NORMAL的情况（ 1.2— >2.4— >3.4— >4.2 )
MP组刚刚启动， 接收端接收到第一个分片报文不是头分片报文， 则存储 该第一个分片报文， 进入 P1状态。
在 P1状态下， 继续接收分片报文， 如果接收到的均为正常的分片报文且 非头分片报文， 则存储并回到 P1状态； 如果接收到的分片报文为异常分片报 文， 则丢弃该异常分片报文并回到 P1状态； 当判断存储的分片报文满足正常 的丢弃条件时，即可能接收端有限的存储空间已满或是等待头分片报文的时间 过长， 则从当前已存储的最小报文序号对应的分片报文开始丢弃， 一直到不再 满足正常的丢弃条件， 进入 P2状态。
在 P2状态下， 继续接收分片报文， 如果存储的分片报文又满足正常的丢 弃条件时， 则从当前已存储的最小报文序号对应的分片报文开始丢弃， 一直到 不再满足正常的丢弃条件为止， 回到 P2状态； 或者若接收到的均为正常的分 片报文且非头分片报文， 则存储该分片报文， 回到 P2状态； 或者接收到的分 片报文为异常分片报文， 则丢弃该异常分片报文， 回到 P2状态。
当后续再接收中，若接收到头分片报文时，若该头分片报文的报文序号在 正常的分片报文窗口之内且报文序号大于已存储的分片报文中的最小报文序 号， 则存储该头分片报文进入 P3状态。
在 P3状态下， 继续接收分片报文， 当接收到正常的分片报文， 且该分片 报文的报文序号大于已存储的分片报文的最小报文序号， 则存储该分片报文， 回到 P3状态；或者接收到异常分片报文，则丢弃异常分片报文，回到 P3状态。 当又接收到另一头分片报文时，若该另一头分片报文的报文序号小于已存储的 分片报文中的最小报文序号，则存储该另一头分片报文并进入正常的组包工作 状态； 或者当接收并存储的分片报文再次满足正常的丢弃条件时，会从已存储 的分片报文中报文序号最小的分片报文开始丢弃，直到存储的分片报文中报文 序号最小的分片报文为之前存储的头分片报文， 则进入正常的组包工作状态。
另外还有 4种条件 ( 2.5、 3.3、 4.3和 5.2 )使得从 PI、 P2、 P3和 NORMAL 返回 EMPTY, 即自动复位功能的实现。 在如图 3所示的条件 2.5、 3.3、 4.3和 5.2下， 若对应的 Pl、 P2、 P3和 NORMAL状态中接收到的分片报文连续为异 常分片报文，则统计异常分片报文的数量，若达到阔值，则丢弃所有分片报文， MP组重新启动。 其中配置异常分片报文的数量的阔值可以参考发送端将一个 报文分片的最大分片数量来设定， 最大分片数量的计算公式可以为： 最大重组 接收包长（MRRU ) /最小分片包长。 则该阔值可设为最大分片数量的 1~2倍 大小。
本实施例提供的应用于多链路协议 MP组的 文处理方法，在 MP组刚刚 启动后，对于第一个报文可能由于链路延时而导致其头分片报文稍晚到达，在 这种情况下，通过本实施例可以保证整个报文不会被丢弃； 另外当接收到异常 分片报文时， 能够保证接收端不会大量丢弃正常分片报文的情况； 如在接收端 接收到的分片报文中出现大量的异常分片报文，也能够保证及时发现并可提供 自动复位的功能， 从而保证了 MP组能够正常的启动。 本发明实施例对 MP组 正常启动的过程进行了完善， 提高了 MP组组包的正确率。
图 4为本发明的应用于多链路协议 MP组的报文处理装置的第一实施例的 结构示意图。 如图 4所示， 该装置包括： 接收模块 41、 判断模块 42、 存储处 理模块 43和查找处理模块 44。 其中， 接收模块 41用于接收 MP组上传送的 分片报文；判断模块 42用于判断接收模块 41接收到的第一个分片报文是否为 头分片报文， 当判断第一个分片报文是头分片报文时， 通知存储处理模块 43 依次存储接收模块 41接收到的后续的分片报文， 及当判断第一个分片报文不 是头分片报文时， 通知查找处理模块 44存储该第一个分片报文， 并存储接收 模块 41接收到的后续的分片报文； 存储处理模块 43用于在判断模块 42判断 第一个分片报文是头分片报文时， 存储该头分片报文， 并依次存储接收模块 41接收到的后续的分片报文， 进入组包工作状态； 查找处理模块 44用于在判 断模块 42判断第一个分片报文不是头分片报文时， 存储第一个分片报文， 并 接收所述接收模块 41后续的分片报文，及当接收到的分片报文为头分片报文， 并确认所述头分片报文为进入组包工作状态所需的头分片报文时，根据所述分 片报文进入所述组包工作状。其中各个分片报文具有报文序号，在报文被分片 为各个分片报文后， 将按照报文序号顺序递增的方式发送到 MP组上进行传 送， 且以头分片报文为起始、 尾分片报文为结束。
本实施例提供的应用于多链路协议 MP组的报文处理装置可置于报文的 接收端的设备中， 或作为单独的设备与接收端相连接； 该 MP组正常启动设置 装置中的各个模块的具体功能的实现如上述 MP组正常启动方法实施例中的 具体描述， 在此不再赘述。 本实施例提供的应用于多链路协议 MP组的 文处理装置，在 MP组刚刚 启动后，对于第一个报文可能由于链路延时而导致其头分片报文稍晚到达，在 这种情况下， 通过本实施例可以保证整个报文不会被丢弃， 对 MP组正常启动 的过程进行完善， 提高 MP组组包的正确率。
图 5为本发明的应用于多链路协议 MP组的报文处理装置的第二实施例的 结构示意图。 如图 5所示， 该装置包括上述实施例中描述的接收模块 41、 判 断模块 42、 存储处理模块 43和查找处理模块 44。
其中查找处理模块 44具体可以包括： 第一存储单元 441用于在判断模块 42 判断第一个分片报文不是头分片报文时， 存储第一个分片报文， 并接收接 收模块 41接收到的后续的分片报文； 第一处理单元 442， 用于当确认所述第 一存储单元接收到的分片报文为头分片报文，并确认所述头分片报文为进入组 包工作状态所需的头分片报文时， 根据所述头分片报文进入所述组包工作状 态。 第一处理单元 442根据 B标志位置 1确认第一存储单元 441接收的后续 的分片报文为该头分片报文时，并若该头分片报文的报文序号小于已存储的分 片报文中的最小报文序号， 通知所述第一存储单元 441存储所述头分片报文， 并进入正常的组包工作状态；若该头分片报文的报文序号大于已存储的分片报 文中的最小报文序号，通知所述第一存储单元 441存储所述头分片报文并继续 接收后续分片报文；当确认第一存储单元 441接收的后续的分片报文为该另一 头分片报文时，若另一头分片报文的报文序号小于已存储的分片报文中的最小 报文序号，通知所述第一存储单元 441存储所述另一头分片报文并继续接收后 续分片报文， 进入组包工作状态。
查找处理模块 44还可以包括： 丢弃处理单元 443， 用于在第一存储单元 441存储的分片报文满足正常的丢弃条件时， 从报文序号最小的分片报文开始 丢弃一定数量的已存储的分片报文。 其中一种情况为， 在第一处理单元 442 查找到头分片报文之前， 当第一存储单元 441 存储的分片报文满足丢弃条件 时，从已存储的分片报文中报文序号最小的分片报文开始丢弃一定数量的分片 报文， 然后继续存储分片报文并查找头分片报文； 另一种情况为， 当第一处理 单元 442 查找到了报文序号大于已存储的分片报文中的最小报文序号的头分 片报文之后， 第一存储单元 441继续存储的分片报文满足正常的丢弃条件时， 则从已存储的分片报文中报文序号最小的分片报文开始丢弃分片报文，一直丢 弃到存储的分片报文中报文序号最小的分片报文即为该存储的头分片报文为 止。
其中存储处理模块 43具体可以包括： 第二存储单元 431用于在判断模块 42 判断第一个分片报文是头分片报文时， 存储该头分片报文， 并依次存储接 收模块 41接收到的后续的分片报文； 第二处理单元 432用于对第二存储单元 431存储的分片报文进行正常的组包工作。 在本实施例中， 存储分片报文时， 可以将报文中的报文头和报文数据作为一个整体进行存储，可以对报文中报文 头和报文数据进行分开存储， 即分别存储报文的报文头和报文数据。
该查找处理模块 44还可以包括： 异常处理模块 444， 用于当检测第一存 储单元 441存储的分片报文为异常分片报文时，丢弃该分片报文； 还用于当检 测第一存储单元 441存储的分片报文为异常分片报文的数量达到阔值时，丢弃 第一存储单元 441存储的所有分片报文， 并重新启动接收模块 41接收 MP组 上传送的分片报文。
本实施例提供的应用于多链路协议 MP组的报文处理装置中的各个模块 的具体功能的实现如上述 MP组正常启动方法实施例中的具体描述，在此不再 赘述。
本实施例提供的应用于多链路协议 MP组的 文处理装置，在 MP组刚刚 启动后，对于第一个报文可能由于链路延时而导致其头分片报文稍晚到达，在 这种情况下，通过本实施例可以保证整个报文不会被丢弃； 另外当接收到异常 分片报文时， 能够保证接收端不会大量丢弃正常分片报文的情况； 如在接收端 接收到的分片报文中出现大量的异常分片报文，也能够保证及时发现并可提供 自动复位的功能， 从而保证了 MP组能够正常的启动。 本发明实施例对 MP组 正常启动的过程进行了完善， 提高了 MP组组包的正确率。
图 6为本发明的应用于多链路协议 MP组的报文处理系统的实施例的结构 示意图。 该系统包括： 发送端 61， 用于将报文进行分片处理， 并将分片后的 分片 4艮文发送到 MP组上传送，其中 MP组为多个物理上独立的 PPP链路捆绑 后的逻辑链路； 接收端 62， 用于接收 MP组上传送的分片报文； 以及如图 4 或图 5所示的 MP组正常启动设置装置 63， 可置于报文的接收端 62的现有设 备中，或作为单独的设备与接收端 62相连接， 用于对接收端 62接收到的分片 文进行处理。
本实施例提供的应用于多链路协议 MP组的报文处理系统中的各个装置 的具体功能的实现如上述应用于多链路协议 MP组的报文处理方法实施例中 的具体描述， 在此不再赘述。 本发明实施例对 MP组正常启动的过程进行了完 善， 提高了 MP组组包的正确率。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程， 是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算 机可获取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。 其中，所述的存储介质可为磁碟、光盘、只读存储记忆体（Read-Only Memory, ROM )或随才 Λ ^储记忆体 ( Random Access Memory, RAM )等。
最后应说明的是： 以上实施例仅用以说明本发明的技术方案， 而非对其限 制； 尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员 应当理解： 其依然可以对前述各实施例所记载的技术方案进行修改，或者对其 中部分技术特征进行等同替换； 而这些修改或者替换， 并不使相应技术方案的 本质脱离本发明各实施例技术方案的精神和范围。

Claims

权 利 要 求
1、 一种应用于多链路协议 MP组的 文处理方法， 其特征在于， 包括： 判断接收到的第一个分片报文是否为头分片报文；
若不是， 存储所述第一个分片报文， 并接收后续的分片报文；
当接收到的分片报文为头分片报文，并确认所述头分片报文为进入组包工 作状态所需的头分片报文时， 根据所述头分片报文进入所述组包工作状态。
2、 根据权利要求 1所述的方法， 其特征在于， 所述当接收到的分片报文 为头分片报文，并确认所述头分片报文为进入组包工作状态所需的头分片报文 时， 根据所述头分片报文进入所述组包工作状态的步骤包括：
当接收到头分片报文时，若所述头分片报文的报文序号小于已存储的分片 报文中的最小报文序号， 则存储所述头分片报文并进入组包工作状态。
3、 根据权利要求 1所述的方法， 其特征在于， 所述当接收到的分片报文 为头分片报文，并确认所述头分片报文为进入组包工作状态所需的头分片报文 时， 根据所述头分片报文进入所述组包工作状态的步骤包括：
当后续接收到头分片报文时，若所述头分片报文的报文序号大于已存储的 分片报文中的最小报文序号， 则存储所述头分片报文并继续接收分片报文； 当接收到另一头分片报文时，若所述另一头分片报文的报文序号小于已存 储的分片报文中的最小报文序号，则存储所述另一头分片报文并进入组包工作 状态。
4、 根据权利要求 1所述的方法， 其特征在于， 所述当接收到的分片报文 为头分片报文，并确认所述头分片报文为进入组包工作状态所需的头分片报文 时， 根据所述头分片报文进入所述组包工作状态的步骤包括： 当后续接收到头分片报文时，若所述头分片报文的报文序号大于已存储的 分片报文中的最小报文序号， 则存储所述头分片报文并继续接收分片报文； 当接收并存储的分片报文满足丢弃条件时，从已存储的分片报文中报文序 号最小的分片报文开始丢弃，直到存储的分片报文中报文序号最小的分片报文 为存储的所述头分片报文， 则进入组包工作状态。
5、 根据权利要求 3或 4所述的方法， 其特征在于， 所述存储所述头分片 报文并继续接收分片报文的步骤包括：
判断存储的分片报文是否满足丢弃条件；
若满足， 丢弃已存储的分片报文中报文序号最小的分片报文。
6、 根据权利要求 1所述的方法， 其特征在于， 所述接收后续的分片报文 的步骤还包括：
若接收到的分片报文为异常分片报文， 丢弃所述分片报文； 或
若接收到的分片报文为异常分片报文的数量达到阔值，丢弃所有已存储的 分片报文以重新接收分片报文。
7、 一种应用于多链路协议 MP组的报文处理装置， 其特征在于， 包括： 接收模块， 用于接收 MP组上传送的分片报文；
判断模块，用于判断所述接收模块接收到的第一个分片报文是否为头分片 报文；
查找处理模块，用于在所述判断模块判断所述第一个分片报文不是头分片 报文时，存储所述第一个分片报文， 并接收所述接收模块接收到的后续的分片 报文，及当接收到的分片报文为头分片报文， 并确认所述头分片报文为进入组 包工作状态所需的头分片报文时， 根据所述分片报文进入所述组包工作状态。
8、 根据权利要求 7所述的装置， 其特征在于， 所述查找处理模块包括： 第一存储单元，用于在所述判断模块判断第一个分片报文不是头分片报文 时， 存储第一个分片报文， 并接收所述接收模块接收到的后续的分片报文； 第一处理单元，用于当确认所述第一存储单元接收到的分片报文为头分片 报文， 并确认所述头分片报文为进入组包工作状态所需的头分片报文时，根据 所述头分片报文进入所述组包工作状态。
9、 根据权利要求 8所述的装置， 其特征在于， 所述第一处理单元还用于 确认所述头分片报文的报文序号小于已存储的分片报文中的最小报文序号时， 通知所述第一存储单元存储所述头分片报文，及根据所述头分片报文进入组包 工作状态。
10、根据权利要求 8所述的装置， 其特征在于， 所述第一处理单元还用于 依次查找存储的分片报文，通知所述第一存储单元存储所述头分片报文并继续 接收后续分片报文；当确认所述第一存储单元接收到的另一分片报文为另一头 分片报文时，若所述另一头分片报文的报文序号小于已存储的分片报文中的最 小报文序号，通知所述第一存储单元存储所述另一头分片报文，根据所述另一 头分片报文进入组包工作状态。
11、根据权利要求 8所述的装置，其特征在于，所述查找处理模块还包括： 丢弃处理单元，用于当所述第一存储单元存储的分片报文满足正常的丢弃 条件时， 从报文序号最小的分片报文开始丢弃已存储的分片报文。
12、 根据权利要求 8至 11任一权利要求所述的装置， 其特征在于， 还包 括： 异常处理模块，用于当所述第一存储单元存储的分片报文为异常分片报文 时，丢弃所述分片报文，或用于当所述第一存储单元存储的分片报文为异常分 片报文的数量达到阔值时， 丢弃所有存储的分片报文； 所述接收模块还用于当所述异常处理模块丢弃所有存储的分片报文时，重 新接收 MP组上传送的分片 文。
PCT/CN2010/070922 2009-03-09 2010-03-09 应用于mp组的报文处理方法及装置 WO2010102551A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910127218A CN101510886B (zh) 2009-03-09 2009-03-09 应用于mp组的报文处理方法及装置
CN200910127218.2 2009-03-09

Publications (1)

Publication Number Publication Date
WO2010102551A1 true WO2010102551A1 (zh) 2010-09-16

Family

ID=41003149

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/070922 WO2010102551A1 (zh) 2009-03-09 2010-03-09 应用于mp组的报文处理方法及装置

Country Status (2)

Country Link
CN (1) CN101510886B (zh)
WO (1) WO2010102551A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9714292B2 (en) 2012-04-05 2017-07-25 Hoffmann-La Roche Inc. Bispecific antibodies against human TWEAK and human IL17 and uses thereof
CN112312204A (zh) * 2020-09-30 2021-02-02 新华三大数据技术有限公司 一种视频流数据分片的组包方法及装置

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101510886B (zh) * 2009-03-09 2012-09-05 华为技术有限公司 应用于mp组的报文处理方法及装置
CN107864135A (zh) * 2017-11-07 2018-03-30 山东网智物联网科技有限公司 物联网通信方法、装置及物联网通信的实现装置
CN112787881A (zh) * 2019-11-11 2021-05-11 中兴通讯股份有限公司 通信链路检测方法、通信装置、存储介质
CN112383418B (zh) * 2020-11-02 2023-07-14 北京左江科技股份有限公司 一种基于fpga的以太网报文高速可靠传输的设计方法
CN113053380B (zh) * 2021-03-29 2023-12-01 海信电子科技(武汉)有限公司 服务器及语音识别方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002045328A1 (en) * 2000-11-30 2002-06-06 Netrake Corporation Method for reordering and reassembling data packets
CN1509025A (zh) * 2002-12-18 2004-06-30 ��Ϊ�������޹�˾ 一种实现报文分片组包的方法
US20040218592A1 (en) * 2003-05-04 2004-11-04 Eyal Nagar Method and apparatus for fast contention-free, buffer management in a multi-lane communication system
CN1852100A (zh) * 2005-07-22 2006-10-25 华为技术有限公司 一种接收端重组数据包的方法
CN101510886A (zh) * 2009-03-09 2009-08-19 华为技术有限公司 应用于mp组的报文处理方法及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002045328A1 (en) * 2000-11-30 2002-06-06 Netrake Corporation Method for reordering and reassembling data packets
CN1509025A (zh) * 2002-12-18 2004-06-30 ��Ϊ�������޹�˾ 一种实现报文分片组包的方法
US20040218592A1 (en) * 2003-05-04 2004-11-04 Eyal Nagar Method and apparatus for fast contention-free, buffer management in a multi-lane communication system
CN1852100A (zh) * 2005-07-22 2006-10-25 华为技术有限公司 一种接收端重组数据包的方法
CN101510886A (zh) * 2009-03-09 2009-08-19 华为技术有限公司 应用于mp组的报文处理方法及装置

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9714292B2 (en) 2012-04-05 2017-07-25 Hoffmann-La Roche Inc. Bispecific antibodies against human TWEAK and human IL17 and uses thereof
CN112312204A (zh) * 2020-09-30 2021-02-02 新华三大数据技术有限公司 一种视频流数据分片的组包方法及装置
CN112312204B (zh) * 2020-09-30 2022-05-24 新华三大数据技术有限公司 一种视频流数据分片的组包方法及装置

Also Published As

Publication number Publication date
CN101510886B (zh) 2012-09-05
CN101510886A (zh) 2009-08-19

Similar Documents

Publication Publication Date Title
WO2010102551A1 (zh) 应用于mp组的报文处理方法及装置
CN110661723B (zh) 一种数据传输方法、计算设备、网络设备及数据传输系统
US10263894B2 (en) Method and apparatus for network congestion control based on transmission rate gradients
WO2017101503A1 (zh) 一种数据传送方法、发送节点、接收节点及数据传送系统
WO2019056899A1 (zh) Oam消息传输方法、传输设备及存储介质
CN101119183A (zh) 重传控制方法及传输设备
EP1536589B1 (en) Transmitter device for controlling data transmission
WO2009026854A1 (fr) Procédé de commande d'envoi de données et dispositif de transmission de données
JP2000115232A (ja) デ―タ・パケット送信方法および装置
US7653060B2 (en) System and method for implementing ASI over long distances
US9197373B2 (en) Method, apparatus, and system for retransmitting data packet in quick path interconnect system
EP1987637B1 (en) System and method for transferring data packets through a communication system
US9876727B2 (en) Physical-layer signaling of flow control updates
EP1798913B1 (en) Transport control method in wireless communication system
US20170134299A1 (en) Method and apparatus for controlling message over heterogeneous network
CN112332951A (zh) 一种以太网帧抢占方法、装置、设备及存储介质
CN113765626A (zh) 一种移动通信系统的数据传输方法和装置
WO2006052053A1 (en) Apparatus for detecting data looping phenomenon and method used by the same
Hu et al. FDA: A novel base station flow control scheme for TCP over heterogeneous networks
CN112565105A (zh) 一种降低时间敏感帧转发时延的方法、装置及设备
US20070253446A1 (en) Method of reducing buffer usage by detecting missing fragments and idle links for multilink protocols and devices incorporating same
EP1774746B1 (en) Data unit receiver and sender control method
JP3587080B2 (ja) パケットバッファ管理装置およびパケットバッファ管理方法
CN113315601B (zh) 多点协助的数据传输方法、装置、存储介质及电子设备
JP2008199432A (ja) データ転送装置及びヘルスチェックデータ処理方法

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: 10750347

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 10750347

Country of ref document: EP

Kind code of ref document: A1