CN1218540C - Control method for exact multicast forwarding - Google Patents

Abstract

The present invention provides an IP control method for exact multicast forwarding, which comprises the steps that 1. establishing a multicast route between a multicast source and an orderer, and a multicast forwarding list by a multicast module; 2. testing whether an IP pack is legal when multicast data is sent in, if true, going to the next step; 3. searching a three-layer list according to a source address and a multicast address, if obtaining a multicast forwarding table index, going to the next step; 4. judging whether a multicast mark in the three-layer list is effective, if true, going to the next step; 5. searching a multicast forwarding list and testing whether a source port is matched with VLAN, if true, going to the next step; 6. searching the multicast forwarding table, and if the source port and a target port are in the same VLAN, performing the second layer forwarding; if the source port and the target port are in different VLAN, performing the third layer forwarding. The adoption of the technical scheme of the present invention realizes the accurate forwarding of multicast data by all the ports of an exchanger on the basis of a standard protocol.

Description

Control Method of Multicast Precise Forwarding

Technical field:

The invention relates to a method for controlling precise multicast forwarding of a communication network, in particular to the control method for precise forwarding of broadband Ethernet multicast, and belongs to the field of broadband network communication.

Background technique:

There are two ways of traditional IP communication, the first is unicast, and the second is broadcast. If the information is to be sent to multiple hosts in the network but not all hosts, either the broadcast method is used, or the source host sends IP packets to multiple target hosts in the network in unicast mode. When broadcasting is implemented, it will not only send information to unnecessary hosts, but also waste bandwidth, and may cause serious broadcast storms due to routing loops. When implemented in unicast mode, a large amount of bandwidth will be wasted due to repeated sending of IP packets, and the load on the server will also be increased. Therefore, the traditional unicast and broadcast communication methods cannot effectively solve the problem of single-point transmission and multi-point reception.

The basic idea of IP multicast is that the source host only sends one copy of data. The destination address in this data is the address of the multicast group, and all receivers in the multicast group can receive the same data copy, and only the target host in the multicast group can receive it, and other hosts in the data network cannot receive it. arrive.

How to control the data to be sent only to the required host, and other hosts cannot receive it, has become an urgent problem to be solved. Early multicast was mainly used on routers. In recent years, with the development of technology, switches have appeared. Switches include almost all functions of routers, and realize hardware forwarding with faster speed. This brings the following new problems for the application of multicast in switches:

A router port is a Layer 3 interface. The switch divides ports according to VLAN (virtual local area network), and VLAN can be equipped with multiple ports. A VLAN can be either a Layer 3 VLAN or a Layer 2 VLAN, and a port can be both a Layer 3 port and a Layer 2 port. Routers do not have this problem. For switches, there is an essential difference between data layer 2 forwarding and data layer 3 forwarding. Multicast data forwarding must distinguish this.

In a switch, a layer-3 VLAN is a layer-3 interface. This interface can have multiple ports. Traditional multicast can only be accurate to the Layer 3 interface, but cannot be accurate to each port of the Layer 3 interface. Multicast data can only be broadcast in VLAN, which wastes network bandwidth and increases network load.

When the source port and the egress port of the multicast are in the same VLAN, the source port and the egress port belong to the same layer-3 interface, and the traditional router and switch RPF (reverse path lookup) inspection cannot pass the request at all.

Invention content:

The purpose of the present invention is to provide a method for controlling precise forwarding of IP multicast, and realize precise forwarding of multicast data by all ports of switches on the basis of standard protocols.

Technical scheme of the present invention is as follows:

A control method for multicast precise forwarding, the steps of which include:

1. The multicast module establishes a multicast route between the multicast source and the orderer, and establishes a multicast forwarding table;

2. When the multicast data arrives, first check whether the IP packet is legal; if it is illegal, discard the data packet or process it to the CPU; for legal IP packets, go to the next step;

3. Search the three-layer table according to the source address and multicast address. If the multicast forwarding table index cannot be found, you can choose to discard it or process it to the CPU; if you get the multicast forwarding table index, go to the next step;

4. Determine whether the multicast flag in the three-layer table is valid. If it is invalid, you can choose to discard it or give it to the CPU for processing; if it is valid, go to the next step;

5. Find the multicast forwarding table, check whether the source port and VLAN match, if not match, you can choose to discard or process it to the CPU; if match, go to the next step;

6. Search the multicast forwarding table, if the source port and the destination port are in the same VLAN, perform Layer 2 forwarding without any processing on the packet; if the source port and destination port are in different VLANs, then perform Layer 3 forwarding.

The control method of described multicast accurate forwarding, in step (1), the mode of setting up multicast forwarding table is as follows: source port and purpose port are in same VLAN, set up two-layer multicast forwarding table; Source port and purpose port are in For different VLANs, establish a three-layer multicast forwarding table, and set the corresponding port working mode to precise forwarding mode;

Correspondingly, in step (6):

The method of Layer 2 forwarding is: the port mode is broadcast mode (the default value when not set), the Layer 2 outbound port is all the ports of the VLAN; the port mode is other values, and the Layer 2 outbound port is the set Layer forwarding port. The port mode only affects Layer 2 forwarding, not Layer 3 forwarding.

The method of performing layer-3 forwarding is: the layer-3 outbound port is the set layer-3 forwarding port.

The port actually forwarded is the sum of the Layer 2 forwarding port and the Layer 3 forwarding port.

In the method, considering the problem of multicast traffic, the switch supports multicast traffic sharing. If the outgoing port is a trunk port, traffic sharing is performed according to a specific search algorithm.

Advantages and positive effects of the present invention: adopting the technical scheme of the present invention, the practical application of IP multicast on the switch is realized, and the efficient data transmission from point to multipoint in the IP network is realized, which can effectively save network bandwidth and reduce network bandwidth. The load is widely used in real-time data transmission, network TV, multimedia conferencing, data copying, games and simulation.

Description of drawings:

FIG. 1 is a flow chart of a specific embodiment of a method for controlling multicast precise forwarding according to the present invention.

Detailed ways:

According to the technical solution of the present invention, an embodiment of the present invention is given below, and Fig. 1 shows a flow chart of the method of the present invention. Specific steps are as follows:

1. Configure the media server server and start the multicast service.

2. Enable the multicast function on the switch and configure the corresponding interface.

3. The orderer starts to order and sends an IGMP join message.

4. The interface management module obtains the port number, VLAN ID, and slot number of the broadcaster, and notifies the multicast routing module.

5. The multicast module establishes a multicast route between the multicast source and the orderer. The source port and the destination port are in the same VLAN, and a Layer 2 multicast forwarding table is established; the source port and the destination port are in different VLANs, and a Layer 3 multicast is established. forwarding table, and set the corresponding port working mode to precise forwarding mode. Data is sent from the multicast source along the RPT tree.

6. When the multicast data arrives, first check whether the IP packet is legal: (1) Ip packet; (2) no option option; (3) IPV4 packet; (4) IP checksum is correct; (5) TTL value is correct ; (6) The destination address belongs to the class D address. If there is a mismatch, it is illegal to discard the data packet, or you can choose to send it to the CPU for processing. A specific processing method is: record into the system log, and give an alarm.

7. According to the multicast address or the source address plus the multicast address, search the three-layer table to obtain the index of the multicast forwarding table. If not found, the data packet is discarded, or sent to the CPU for processing.

8. Determine whether the multicast flag in the Layer 3 table is valid. If it is invalid, you can choose to discard it or give it to the CPU for processing; if it is valid, go to the next step.

9. Search the multicast forwarding table, and check whether the source port and VLAN match. If they do not match, you can choose to discard them or give them to the CPU for processing.

10. Search the multicast forwarding table, check whether it needs to broadcast in the entire VLAN according to the current configuration, whether it needs to perform Layer 2 forwarding, and obtain the corresponding outbound port bitmap.

If the source port and the destination port are in the same VLAN, layer 2 forwarding is performed without any processing on the packet; if the source port and destination port are in different VLANs, layer 3 forwarding is required, the TTL value is reduced by 1, and the source MAC and destination MAC, and recalculate the checksum.

The method for Layer 2 forwarding is: the port mode is broadcast mode, and the Layer 2 outgoing port is the port owned by the VLAN; the port mode is other values, and the Layer 2 outgoing port is the set Layer 2 forwarding port.

The method of performing layer-3 forwarding is: the layer-3 outbound port is the set layer-3 forwarding port.

The port actually forwarded is the sum of the Layer 2 forwarding port and the Layer 3 forwarding port.

11. If the forwarding port is a Trunk port, calculate the port according to a certain algorithm, such as source MAC, source IP, destination MAC, and destination IP.

12. The data comes out from the port, arrives at the requester, and starts playing.

13. The orderer wants to stop ordering, sends an IGMP leave message, and the multicast module updates the multicast route.

14. Update the outgoing interface of the multicast forwarding table, and set the corresponding port to empty.

15. When the multicast data arrives, check that the corresponding port is empty, and then the data will not be forwarded.

16. On-demand stop.

Claims (3)

1. the control method of group broadcast precise relay, its step comprises:

(1) the multicast module between multicast source and program request person, set up multicast path by, set up multicast forward table in the following manner: source port and destination interface are set up Layer 2 Multicast and are transmitted at same VLAN; Source port and destination interface are set up three layers of multicast forward table at different VLAN, and it is the precise relay pattern that the corresponding ports mode of operation is set simultaneously;

(2) multicast packet arrives, and checks earlier whether the IP bag is legal; If have illegally, then abandon this packet or handle to CPU; IP bag for legal enters next step;

(3) according to source address and multicast address, search three layers of table, if can not find the multicast forwarding table index, then select to abandon or handle to CPU; If obtain the multicast forwarding table index, then enter next step;

(4) judge whether multicast flags is effective in three layers of table, if invalid then select to abandon or to handle to CPU; If effectively then enter next step;

(5), check whether source port and VLAN mate, if do not match then select to abandon or handle to CPU according to multicast forward table index search multicast forward table; If the coupling would enter next step;

(6) search multicast forward table, if source port and destination interface are in same VLAN, then carry out two layers of forwarding as follows, bag is left intact: if port mode is a broadcast mode then two layers of outbound port are all ports of this VLAN, if port mode is for other value then two layers of outbound port are transmitted ports for two layers of setting; If source port and destination interface at different VLAN, then carry out three layers of forwarding as follows: three layers of outbound port are transmitted ports for three layers of setting, and the actual port of transmitting is two layers and transmits ports and three layers of summation of transmitting ports.

2. the control method of group broadcast precise relay as claimed in claim 1 is characterized in that, the described actual port of transmitting is if the Trunk port then according to searching algorithm, calculates port and carries out traffic sharing.

3. the control method of group broadcast precise relay as claimed in claim 2 is characterized in that, described algorithm is selected from source MAC, source IP, purpose MAC or purpose IP algorithm.

Images