A method and apparatus for shortening congestion of a practical general multicast network

Abstract

The present disclosure provides a method and device for shortening congestion of a practical general multicast network. The method comprises: each switching device sending a flow head packet count announcement message and a flow tail packet count announcement message to a convergence point in the PGM network; wherein the flow head packet count announcement message and the flow tail packet count announcement message respectively carry a count value of a first part of original data packets and a count value of a second part of original data packets of each service flow of a specified multicast group; the convergence point calculates an end-to-end packet loss rate in the PGM network based on the flow head packet count announcement message or the flow tail packet count announcement message of each switching device; the convergence point sends a flow control announcement message of the highest end-to-end packet loss rate to a multicast source; and the multicast source reduces a sending rate of the specified multicast group based on the flow control announcement message.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the communication technology, in particular to a method and device for shortening congestion of a pragmatic general multicast network. BACKGROUND

[0002] PGM (Pragmatic General Multicast) is a reliable multicast transmission protocol, which is suitable for multicast data with ordered or unordered and no duplication from multiple multicast senders to multiple receivers.

[0003] The multicast sender sends ODATA (Original Data) packets in sequence according to a local policy to define a sending window, and the multicast receiver compares the sequence number of the currently received ODATA with the sequence number of the previously received ODATA. If there is a gap between the sequence numbers, it is considered that a packet is lost, and a unicast NAK (Non-Acknowledgment) packet is sent on the reverse path of the distribution tree from the multicast sender after waiting for a random time to confirm that no matching NAK or matching NCF is received.

[0004] The multicast receiver calculates the packet loss rate of the multicast group to which the ODATA belongs, and sends the calculated packet loss rate to the multicast sender through the NAK packet. The three-layer switching device receives the NAK packet and writes the local link packet loss rate to the multicast sender. The multicast sender receives the NAK packet and reduces the sending rate of the original data packet based on the highest packet loss rate. Since the flow control mechanism is initiated by the NAK packet sent by the multicast receiver, the congestion in the PGM has occurred for a period of time after the multicast sender receives the flow control request. The multicast sender still sends the original data packet at the original sending rate, which not only prolongs the congestion, but also causes the sequence number of the lost multicast data packet in the NAK packet to be outside the current sending window, so that the multicast sender no longer sends the repair packet of the original data packet, and the receiver cannot repair the multicast packet. SUMMARY

[0005] The present disclosure aims to provide a method and device for shortening congestion of a pragmatic general multicast network, and avoiding the receiver from being unable to repair the multicast packet.

[0006] To achieve the above object, the disclosure provides a method for shortening congestion of a practical general multicast network, which comprises: each switching device sending a flow head section packet count announcement packet and a flow tail section packet count announcement packet to a convergence point in the PGM network; wherein the flow head section packet count announcement packet and the flow tail section packet count announcement packet respectively carry a count value of a first part of original data packets of each service flow of a specified multicast group and a count value of a second part of original data packets; the convergence point calculates an end-to-end packet loss rate in the PGM network based on the flow head section packet count announcement packet or the flow tail section packet count announcement packet of each switching device; the convergence point sends a flow control announcement packet with the highest end-to-end packet loss rate to a multicast source; and the multicast source reduces the sending rate of the specified multicast group based on the flow control announcement packet.

[0007] To achieve the above object, the disclosure provides a device for shortening congestion of a practical general multicast network, which comprises a processor, a machine readable storage medium, a memory, a switching chip and a network interface; the processor executes the following operations by running machine executable instructions recorded in the machine readable storage medium: receiving a flow head section packet count announcement packet and a flow tail section packet count announcement packet from each switching device in the PGM network; wherein the flow head section packet count announcement packet and the flow tail section packet count announcement packet respectively carry a count value of a first part of original data packets of each service flow of a specified multicast group and a count value of a second part of original data packets; calculating an end-to-end packet loss rate in the PGM network based on the flow head section packet count announcement packet or the flow tail section packet count announcement packet from each switching device; and sending a flow control announcement packet with the highest end-to-end packet loss rate to a multicast source.

[0008] To achieve the above object, the disclosure further provides a device for shortening congestion of a practical general multicast network, which comprises a processor, a machine readable storage medium, a memory, a switching chip and a network interface; the processor executes the following operations by running machine executable instructions recorded in the machine readable storage medium: establishing a pair of head section statistical flow table entries and a tail section statistical flow table for a specified multicast group; counting the number of packets of a first part of original data packets of a service flow of the specified multicast group based on the head section statistical flow table entries; and counting the number of packets of a second part of original data packets of the service flow of the specified multicast group based on the tail section statistical flow table entries.

[0009] The disclosure has the beneficial effect that the convergence point announces the packet loss rate of a specified multicast group of the PGM network to the multicast source, and the multicast source timely reduces the sending rate of original data packets of the specified multicast group, which not only quickly shortens the congestion in the PGM network, but also avoids the sending window of the original data packets from sliding to the next sending window, so that the multicast source can send repair packets to the multicast receiver according to the lost multicast data packet sequence number of the NAK packet of the multicast receiver. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A schematic diagram of a method embodiment for shortening congestion of a practical general multicast network according to the present disclosure is shown in FIG. 1.

[0011] Figure 2 A schematic diagram of a practical general multicast network according to the present disclosure is shown in FIG. 2.

[0012] Figure 3 A schematic diagram of a practical general multicast network according to the present disclosure is shown in FIG. 3.

[0013] Figures 4A-4B A schematic diagram of a practical general multicast network according to the present disclosure is shown in FIG. 4.

[0014] Figure 5 A schematic diagram of a device embodiment for shortening congestion of a practical general multicast network according to the present disclosure is shown in FIG. 5.

[0015] Figure 6 A schematic diagram of a device embodiment for shortening congestion of a practical general multicast network according to the present disclosure is shown in FIG. 6. DETAILED DESCRIPTION

[0016] A number of instances will be described in detail with reference to the accompanying drawings. In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure. Well-known methods, procedures, components, and circuits have not been described in detail so as not to unnecessarily obscure aspects of the examples.

[0017] In the use of terms, the term "include" means "including, but not limited to"; the term "contain" means "including, but not limited to"; the terms "above", "within", and "below" include the number; the terms "greater than", "less than" mean not including the number. The term "based on" means at least based on part of it.

[0018] Figure 1 A method embodiment for shortening congestion of a practical general multicast network according to the present disclosure is shown in FIG. 1, which includes the following steps:

[0019] Step 101, each switching device sends a flow head message count announcement message and a flow tail message count announcement message to a convergence point in the PGM network;

[0020] In which, the flow head message count announcement message and the flow tail message count announcement message respectively carry the count value of the first part of the original data message of each service flow of the designated multicast group, the count value of the second part of the original data message;

[0021] Step 102, the convergence point calculates the end-to-end packet loss rate in the PGM network based on the flow head message count announcement message or the flow tail message count announcement message of each switching device;

[0022] Step 103, the rendezvous point sends a flow control notification message with the highest end-to-end packet loss rate to the multicast source;

[0023] Step 104, the multicast source reduces the sending rate of the specified multicast group based on the flow control notification message.

[0024] The advantage of the present disclosure is that the rendezvous point notifies the multicast source of the packet loss rate of the specified multicast group in the PGM network, and the multicast source timely reduces the sending rate of the original data packet of the specified multicast group, which not only quickly shortens the congestion in the PGM network, but also avoids the sending window of the original data packet sliding to the next sending window, so that the multicast source can send repair packets to the multicast receiver according to the sequence number of the lost multicast data packet in the NAK message of the multicast receiver.

[0025] Figure 2 A schematic diagram of a general multicast network is provided for the embodiments of the present disclosure. Switch A is configured as a rendezvous point (RP) or is elected as the RP via switch A-switch I according to the PIM-SM (Protocol Independent Multicast-Sparse Mode) protocol.

[0026] The multicast source s1 defines the sending window according to the local policy and sends the in-sequence ODATA data packets of the multicast group.

[0027] The protocol number of the PGM protocol is 113; different types of PGM protocol messages are used to realize different functions. In the PGM protocol, the type field value of the data packet is 4.

[0028] Therefore, switch A-switch F first identifies the PGM protocol message according to the protocol number; then, according to the type 4 of the PGM protocol message, it identifies that the ODATA packet is received, rather than other types of control packets or acknowledgement packets.

[0029] Figure 2 In the PIM-SM protocol, the receivers R1 and R2 can initiate an active prune message, and the packets from the multicast source S1 are sent to the multicast receivers R1 and R2 via switch B, switch A, switch C, switch D, and switch E on the multicast tree.

[0030] Switch B-Switch E, according to the received ODATA first message, establishes a pair of first segment statistical flow table items and tail segment statistical flow table for the multicast group; in the first segment statistical flow table item, the source IP address is the IP address of the multicast source s1, the destination IP address is the IP address of the multicast group, the identification value of the IP header is 0-32767; the action item is the packet quantity counting; in the tail segment statistical flow table item, the source IP address is the IP address of the multicast source s1, the destination IP address is the IP address of the multicast group, the identification value of the IP header is 32768-65535; the action item is the packet quantity counting.

[0031] In Switch B-Switch E, the first segment statistical flow table item is used for recording the first part original data packet quantity of the identification value 0-32767 in each service flow of the multicast group, and the tail segment statistical flow table item records the original data packet quantity of the identification value 32768-65535 in each service flow.

[0032] The original data packet received by Switch B-Switch E is matched to the first segment statistical flow table item according to the source IP address, the destination IP address and the identification value of the IP header, and the packet quantity of the service flow identification value 0-32767 is refreshed.

[0033] The original data packet received by Switch B-Switch E is matched to the tail segment statistical flow table item according to the source IP address, the destination IP address and the identification value of the IP header, a first segment packet counting announcement message is generated, and the packet quantity recorded in the first segment statistical flow table item is sent to the convergence point switch A; and then the packet quantity of the service flow identification value 32767-65535.

[0034] The original data packet received by Switch B-Switch E is matched to the first segment statistical flow table item again according to the source IP address, the destination IP address and the identification value of the IP header, and a tail segment packet counting announcement message is generated.

[0035] The multicast source sends the original data packet of the next hop service flow after sending 65536 original data packets of one service flow of the multicast group, and the identification value of the IP of the original data packet of the next hop service flow is flipped to 0. Therefore, Switch B-Switch I re-counts the next original data packet.

[0036] In the present disclosure, each switching device counts the number of first part original data packets of each service flow based on the flow head segment statistics flow table item, and counts each service flow of the multicast group into two parts, so as to prevent the switching device from being unable to distinguish the out-of-order original data packet and the original data packet with the identification value rollover. For example, the identification values of two original data packets received by the switching device are both 1, the original data packet received after the completion of the head segment statistics flow table item matching occurs an out-of-order fault; the original data packet manually pressed after the completion of the tail segment statistics flow table item matching is the original data packet of the next service flow, and the head segment statistics flow table item matching should be performed.

[0037] Figure 3 A schematic diagram for shortening congestion of a practical general multicast network is provided in the present disclosure;

[0038] The switching device B, the switching device C (not shown in the figure), the switching device D (not shown in the figure) and the switching device E respectively send a flow head segment packet count announcement packet 301 to the switching device A as the RP.

[0039] The switching device A identifies the sending switching device based on the source IP address of each flow head segment packet count announcement packet 301, calculates the number of received packets of the switching device B, the switching device C, the switching device D and the switching device E respectively according to the number of packets written by each flow head segment packet count announcement packet 301, and calculates the end-to-end packet loss rate between two switching devices on the multicast tree according to the packet loss rate of each switching device.

[0040] Figure 3 In the present disclosure, the switching device A identifies the maximum end-to-end packet loss rate in the calculated end-to-end packet loss rate, and sends a flow control announcement packet 302 to the multicast source S1; the source IP address is the IP address of the switching device A, the destination IP address is the IP address of the multicast source S1, and the type field is a custom specified value 0X20.

[0041] The multicast source S1 receives the flow control announcement packet 302, reads the maximum end-to-end packet loss rate, and reduces the sending rate of the original data packet of the multicast group.

[0042] Therefore, in the present disclosure, the three-layer switching device of the PGM network accelerates the feedback of the packet loss rate, and shortens the congestion time in the PGM network.

[0043] The switching device B, the switching device C (not shown in the figure), the switching device D (not shown in the figure) and the switching device E respectively send a flow tail segment packet count announcement packet 303 to the switching device A as the RP.

[0044] Switch A identifies the sending switch based on the source IP address of each flow tail segment packet count announcement packet 303, and calculates the number of packets received by switch B, switch C, switch D, and switch E according to the number of packets written by each flow tail segment packet count announcement packet 303, and calculates the end-to-end packet loss rate between two switches on the multicast tree according to the packet loss rate of each switch.

[0045] Figure 3 In the end-to-end packet loss rate calculated by switch A, the largest end-to-end packet loss rate is identified; a flow control announcement packet 304 is sent to the multicast source S1; the source IP address is the IP address of switch A, the destination IP address is the IP address of the multicast source S1; the type (type) field is 0X20.

[0046] If the congestion of the PGM network has been improved through the previous traffic control processing, the multicast source S1 receives the flow control announcement packet 304, determines that the largest end-to-end packet loss rate is reduced, and then the multicast source S1 increases the sending rate of the original data packets of the multicast group.

[0047] If the congestion of the PGM network is not improved enough through the previous traffic control processing, the multicast source S1 receives the flow control announcement packet 304, determines that the largest end-to-end packet loss rate continues to increase, and then the multicast source S1 further reduces the sending rate of the original data packets of the multicast group to speed up the shortening of the congestion time in the PGM network.

[0048] Figures 4A-4B A schematic diagram for shortening the congestion of a practical general multicast network is provided for another embodiment of the present disclosure.

[0049] Figure 4A In the end-to-end packet loss rate calculated by switch A, the largest end-to-end packet loss rate is identified; a flow control announcement packet 304 is sent to the multicast source S1; the source IP address is the IP address of switch A, the destination IP address is the IP address of the multicast source S1; the type (type) field is 0X20.

[0050] Switch A receives the NAK packet 401, strips off the packet loss rate of the multicast receiver R1 carried by the NAK packet 401, and continues to send to the multicast source S1 according to the destination IP address of the NAK packet 401.

[0051] The multicast source S1 receives the NAK packet 401, and if it is determined that the sequence number of the lost multicast data packet in the NAK packet 401 is within the current sending window, the multicast source S1 sends a repair packet 403 to the multicast receiver R1.

[0052] Figure 4AIn the figure, the multicast receiver R2 sends the NAK message 402 with the packet loss rate in the reverse direction on the multicast tree; the switch E directly connected to the multicast receiver R2 receives the NAK message 401 and sends it to the switch C (not shown in the figure) in the reverse direction on the multicast tree; and the switch C finally sends it to the switch A.

[0053] The switch A receives the NAK message 402, strips the packet loss rate of the multicast receiver R1 carried by the NAK message 402, and continues to send it to the multicast source S1 according to the destination IP address of the NAK message 402.

[0054] The multicast source S1 receives the NAK message 402, and if it is determined that the sequence number of the lost multicast data message in the NAK message 402 is within the current sending window, the multicast source S1 sends the repair message 404 to the multicast receiver R2.

[0055] Figure 4B In the figure, before the switch A sends the NAK message 401 with the packet loss rate stripped to the multicast source S1, the switch A identifies the sending end as the multicast receiver R1 according to the source IP address of the NAK message 401, and identifies the directly connected switch of the multicast receiver R1 on the multicast tree as the switch D.

[0056] The switch A calculates the packet loss rate of the directly connected switch D based on the count value of the flow head message count announcement message 301 and the count value of the flow tail message count announcement message 303 sent by the directly connected switch A of the multicast receiver R1.

[0057] The switch A compares the packet loss rate of the switch D with the packet loss rate in the NAK message 401, and determines that they are the same, indicating that the multicast receiver R1 does not further send the packet loss, and then sends the NAK message with the packet loss rate stripped according to the source IP address of the NAK message 401. Figure 4A as shown in the figure.

[0058] Figure 4B In the figure, before the switch A sends the NAK message 402 with the packet loss rate stripped to the multicast source S1, the switch A identifies the sending end as the multicast receiver R2 according to the source IP address of the NAK message 402, and identifies the directly connected switch of the multicast receiver R2 on the multicast tree as the switch E.

[0059] The switch A calculates the packet loss rate of the directly connected switch E based on the count value of the flow head message count announcement message 301 and the count value of the flow tail message count announcement message 303 sent by the directly connected switch E of the multicast receiver R2.

[0060] The switch A compares the packet loss rate of the switch D with the packet loss rate in the NAK message 401, and determines that they are the same, indicating that the multicast receiver R2 further sends the packet loss, and then sends the NAK message with the packet loss rate stripped according to the source IP address of the NAK message 401. Figure 4AThe sending switch A strips the NAK packet 402 of the packet loss rate. Then, the sending switch A generates a flow control notification packet 501; the flow control notification packet 501 carries the packet loss rate of the multicast receiver R2.

[0061] The multicast source S1 receives the flow control notification packet 501, and further reduces the sending rate of the original data packets of the multicast group.

[0062] Since the multicast receivers feed back the packet loss rate at a slower speed than the switches on the multicast tree, the packet loss rate of the PGM network is timely notified to the multicast source by the only one switch as the RP, and the congestion of the PGM network is shortened. The original multicast data packets are finally forwarded to the multicast receivers through the network, the packet loss rate fed back by the multicast receivers is more accurate, and the multicast source continues to adjust the sending rate of the original multicast data packets according to the packet loss rate fed back by the multicast receivers, so that the number of lost packets is minimized.

[0063] Figure 5 A schematic diagram of an equipment embodiment for shortening the congestion of a practical general multicast network is provided for the embodiments of the present disclosure; the equipment 50 can be applied to a switching device as a convergence point in a PGM network. The equipment 50 includes a processor 51, a machine readable storage medium 52, a switching chip 53, and a network interface 54.

[0064] The processor 51 performs the following operations by running the machine executable instructions recorded in the machine readable storage medium 52: receiving a flow head packet count notification packet and a flow tail packet count notification packet from each switching device in the PGM network; wherein the flow head packet count notification packet and the flow tail packet count notification packet respectively carry a count value of a first part of original data packets and a count value of a second part of original data packets of each service flow of a specified multicast group; calculating an end-to-end packet loss rate in the PGM network based on the flow head packet count notification packet or the flow tail packet count notification packet from each switching device; and sending a flow control notification packet of the highest end-to-end packet loss rate to a multicast source.

[0065] Before the processor 51 performs the operation of sending the flow control notification packet of the highest end-to-end packet loss rate to the multicast source by running the machine executable instructions of the machine readable storage medium 52, the processor 51 further performs the following operations: receiving a negative acknowledgement packet; stripping the packet loss rate of the negative acknowledgement packet; and sending the negative acknowledgement packet with the stripped packet loss rate to the multicast source.

[0066] The processor 51, by running the machine executable instructions recorded by the machine readable storage medium 52, performs the following operations before stripping the loss rate of the negative acknowledgement message: identifying the multicast receiver sending the NAK message; identifying the direct connection switching device of the multicast receiver; calculating the loss rate of the direct connection switching device based on the count value of the first part of the original data message and the count value of the second part of the original data message announced by the direct connection switching device; determining that the loss rate of the direct connection switching device is less than the loss rate carried by the negative acknowledgement message, and identifying the loss rate carried by the negative acknowledgement message as the highest end-to-end loss rate.

[0067] The processor 51, by running the machine executable instructions recorded by the machine readable storage medium 52, performs the following operations: identifying the multicast receiver sending the NAK message; identifying the direct connection switching device of the multicast receiver; calculating the loss rate of the direct connection switching device based on the count value of the first part of the original data message and the count value of the second part of the original data message announced by the direct connection switching device; determining that the loss rate of the direct connection switching device is equal to the loss rate carried by the negative acknowledgement message, and identifying the highest end-to-end loss rate in the calculation of the end-to-end loss rate in the PGM network.

[0068] Figure 6 The device 60 provided by the embodiment of the present disclosure can be applied to the switching device other than the sink in the PGM network. The device 60 comprises a processor 61, a machine readable storage medium 62, a switching chip 63, and a network interface 64.

[0069] The processor 61, by running the machine executable instructions recorded by the machine readable storage medium 62, performs the following operations: establishing a pair of head segment statistical flow table items and tail segment statistical flow table for a specified multicast group; counting the number of the first part of the original data message of the service flow of the specified multicast group based on the head segment statistical flow table item; and counting the number of the second part of the original data message of the service flow of the specified multicast group based on the tail segment statistical flow table item.

[0070] The processor 61, by running the machine executable instructions recorded by the machine readable storage medium 62, performs the following operations: sending a flow head segment message count announcement message to the sink; the flow head segment message count announcement message carrying the number of the first part of the original data message of the service flow; clearing the number of the first part of the original data message of the service flow; sending a flow tail segment message count announcement message to the sink, and clearing the number of the second part of the original data message of the service flow.

[0071] The processor 61, by running the machine executable instructions recorded by the machine readable storage medium 62, performs the following operations: receiving a negative acknowledgement message; and sending a negative acknowledgement message to the sink.

[0072] In this disclosure, a machine-readable storage medium can be any electronic, magnetic, optical, or other physical storage device that stores or contains data, such as executable instructions, data structures, or the like. For example, any machine-readable storage medium herein can be any type of random access memory (RAM), volatile memory, non-volatile memory, flash memory, storage drive (such as a hard drive), solid-state drive, any type of storage optical disc (such as a compact disc, DVD, etc.), and the like, or a combination thereof. Moreover, any machine-readable storage medium herein can be a non-transitory machine-readable storage medium.

[0073] The above description is merely illustrative of the application, and is not intended to limit the scope of the application. As such, modifications and variations of the disclosed concepts should be considered within the scope of the applications.

Claims

1. A method for shortening congestion of a practical general multicast network, characterized by, The method comprises, Each exchange device sends a flow head section packet count announcement message and a flow tail section packet count announcement message to a convergence point in the PGM network; wherein the flow head section packet count announcement message and the flow tail section packet count announcement message respectively carry a count value of a first part of original data packets of each service flow of a designated multicast group and a count value of a second part of original data packets; The convergence point calculates an end-to-end packet loss rate in the PGM network based on the flow head section packet count announcement message or the flow tail section packet count announcement message of each exchange device; The convergence point sends a flow control announcement message of the highest end-to-end packet loss rate to a multicast source; The multicast source determines to read the highest end-to-end packet loss rate increase based on the flow control announcement message, and reduces the sending rate of the designated multicast group.

2. The method of claim 1, wherein, Before the convergence point sends the flow control announcement message of the highest end-to-end packet loss rate to the multicast source, the method further comprises, Each exchange device receives a negative acknowledgement message from any multicast receiver and sends the negative acknowledgement message to the convergence point; The convergence point strips the packet loss rate of the negative acknowledgement message; The convergence point sends the negative acknowledgement message with the stripped packet loss rate to the multicast source.

3. The method of claim 2, wherein, Before the convergence point strips the packet loss rate of the negative acknowledgement message, the method further comprises, The convergence point identifies the multicast receiver that sends the negative acknowledgement message; The convergence point identifies a directly connected exchange device of the multicast receiver; The convergence point calculates the packet loss rate of the directly connected exchange device based on the count value of the first part of original data packets and the count value of the second part of original data packets announced by the directly connected exchange device; The convergence point determines that the packet loss rate of the directly connected exchange device is less than the packet loss rate carried by the negative acknowledgement message, and identifies the packet loss rate carried by the negative acknowledgement message as the highest end-to-end packet loss rate.

4. The method of claim 3, wherein, The method further comprises that the convergence point identifies the multicast receiver that sends the negative acknowledgement message; The convergence point identifies a directly connected exchange device of the multicast receiver; The convergence point calculates the packet loss rate of the directly connected exchange device based on the count value of the first part of original data packets and the count value of the second part of original data packets announced by the directly connected exchange device; The convergence point determines that the packet loss rate of the directly connected exchange device is equal to the packet loss rate carried by the negative acknowledgement message, and identifies the highest end-to-end packet loss rate in calculating the end-to-end packet loss rate in the PGM network.

5. The method of claim 1, wherein, Each exchange device establishes a pair of head section statistical flow table items and a tail section statistical flow table for the designated multicast group; Each exchange device counts the number of packets of the first part of original data packets of each service flow based on the head section statistical flow table items; Each exchange device counts the number of packets of the second part of original data packets of each service flow based on the tail section statistical flow table items; Each exchange device clears the count of the number of packets of the first part of original data packets of each service flow after sending the flow head section packet count announcement message; Each exchange device clears the count of the number of packets of the second part of original data packets of each service flow after sending the flow tail section packet count announcement message.

6. An apparatus for shortening congestion of a practical general multicast network, characterized by, The device comprises a processor, a machine readable storage medium, a memory, a switching chip, a network interface; the processor executes the following operations by running the machine executable instructions recorded by the machine readable storage medium, receiving the flow head packet count announcement packet and the flow tail packet count announcement packet from each switching device in the PGM network; wherein the flow head packet count announcement packet and the flow tail packet count announcement packet respectively carry the count value of the first part of original data packets and the count value of the second part of original data packets of each service flow of the specified multicast group; calculating the end-to-end packet loss rate in the PGM network based on the flow head packet count announcement packet or the flow tail packet count announcement packet from each switching device; sending the flow control announcement packet with the highest end-to-end packet loss rate to the multicast source.

7. The apparatus of claim 6, wherein, Before the processor executes the operation of sending the flow control announcement packet with the highest end-to-end packet loss rate to the multicast source by running the machine executable instructions recorded by the machine readable storage medium, it also executes the following operations, receiving the negative acknowledgement packet; stripping the packet loss rate of the negative acknowledgement packet; sending the negative acknowledgement packet with the stripped packet loss rate to the multicast source.

8. The apparatus of claim 7, wherein, Before the processor executes the operation of stripping the packet loss rate of the negative acknowledgement packet, it also executes the following operations, identifying the multicast receiver sending the negative acknowledgement packet; identifying the directly connected switching device of the multicast receiver; calculating the packet loss rate of the directly connected switching device based on the count value of the first part of original data packets and the count value of the second part of original data packets announced by the directly connected switching device; determining that the packet loss rate of the directly connected switching device is less than the packet loss rate carried by the negative acknowledgement packet, and identifying the packet loss rate carried by the negative acknowledgement packet as the highest end-to-end packet loss rate.

9. The apparatus of claim 8, wherein, The processor also executes the following operations by running the machine executable instructions recorded by the machine readable storage medium, identifying the multicast receiver sending the negative acknowledgement packet; identifying the directly connected switching device of the multicast receiver; calculating the packet loss rate of the directly connected switching device based on the count value of the first part of original data packets and the count value of the second part of original data packets announced by the directly connected switching device; determining that the packet loss rate of the directly connected switching device is equal to the packet loss rate carried by the negative acknowledgement packet, and identifying the packet loss rate carried by the negative acknowledgement packet as the highest end-to-end packet loss rate.

10. An apparatus for shortening congestion in a practical general multicast network, characterized by, The device comprises a processor, a machine readable storage medium, a memory, a switching chip, a network interface; the processor executes the following operations by running the machine executable instructions recorded by the machine readable storage medium, establishing a pair of head statistical flow table items and tail statistical flow table for a specified multicast group; counting the number of packets of the first part of original data packets of the service flow of the specified multicast group based on the head statistical flow table item; counting the number of packets of the second part of original data packets of the service flow of the specified multicast group based on the tail statistical flow table item; The flow head packet count announcement message and the flow tail packet count announcement message are sent to the sink point in the PGM network, wherein the flow head packet count announcement message and the flow tail packet count announcement message respectively carry the packet number of the first part of original data packets and the packet number of the second part of original data packets of each service flow of the designated multicast group.

11. The apparatus of claim 10, wherein, The processor, by running the machine-executable instructions recorded by the machine readable storage medium, further performs the following operations, The packet number of the first part of original data packets of the service flow is cleared. The packet number of the second part of original data packets of the service flow is cleared.

12. The apparatus of claim 10, wherein, The processor, by running the machine-executable instructions recorded by the machine readable storage medium, further performs the following operations, A negative acknowledgement message is received. The negative acknowledgement message is sent to the sink point.

Images