Application of IP multicast technology in video

I. Introduction
With the development of the network, people have developed various services on the network platform, such as E-mail, TELNET, FTP, and WWW. These services are point-to-point data transmission; people prefer to hold video conferences, listen to live concerts, and watch live broadcasts on the Internet. These are data transmission from point to point or to point, and IP multicast is required) communication technology. At present, this technology has become a hot topic in foreign research groups and research institutions. Many network vendors have provided products that support IP multicast technology, and some network providers (ISPs) these advanced services are gradually provided, and many large companies that provide large-scale network applications and services are beginning to use multicast communication.

Ii. Generation of IP multicast technology
The concept of IP multicast was first introduced in Steve deering's doctoral thesis in 1988. In 1989, Steve deering expanded the standard IP network layer protocol and proposed the IP multicast specification; in March 1992, the first time I set up the multicast Backbone Network MBone, IETF successfully held a meeting on the multicast network, which attracted widespread attention. However, the first WWW browser appeared in 1990 and has grown to 1993 WWW sites by 100. Although multicast and WWW were in the same period, the development of multicast was far slower than that of WWW, the main reason is that the IP multicast communication mode requires a vro in a complex state, requiring the vro to provide the information status of each group and each source, with the increasing complexity of the Internet, further development of multicast is difficult. Later, some well-designed multicast routing protocols such as PIM-DM, PIM-SM, etc.), so that the multicast IP packet can be correctly and quickly sent to thousands of recipients, the Technology and Application of IP multicast began to develop rapidly.

Currently, IP Multicast can run on any Architecture Network, including the Internet, ATM, frame relay, SMDS, and satellites, and in many application fields, it can be used in video conferences, multimedia, news releases, and remote live broadcasts from space.

Iii. Concept of IP Multicast
IP multicast uses a protocol to transmit IP data packets from one source to multiple destinations. It copies the information and sends it to a group of addresses to all the recipients who want to receive it. IP multicast transfers IP data packets "Do your best" to a host set that constitutes a multicast group. Each member of the group can be distributed on an independent physical network. The relationship between members in an IP multicast group is dynamic, and the host can join and exit the group at any time. The Group's membership relationship determines whether the host receives multicast data packets sent to the group, A host that is not a member of a group can also send multicast data packets to the group.

Compared with unicast) and broadcast, multicast is very efficient, because any given link can be used once, which can save network bandwidth and resources.

In a unicast environment, the video server sends n information streams in sequence, which are received by users in the network. A total of nx1.5Mbit/s bandwidth is required; if the server is in 10 Mbit/s Ethernet ~ The bandwidth is fully occupied by the seven information flows. If a high-speed Ethernet network is used, it can only accommodate a maximum of 250 ~ 300 million Mbit/s video streams, so the capacity between the server and the host interface is a huge bottleneck.

In a multicast environment, regardless of the number of users in the network, a video stream sent by the server is replicated by the vrouters or switches in the network to produce n video streams at the same time, broadcast to each user, only 1.5 Mbit/s of bandwidth is required.

It can be seen that IP Multicast can effectively save network bandwidth and resources, manage network addition and control overhead, greatly reduce the load on the sending server, and send information with high performance.

In addition, the information transmitted by Multicast can reach the client at the same time, with a low latency, and the server on the network does not need to know the address of each client. All Receivers use one network multicast address to implement anonymous services. IP multicast is upgradeable and compatible with new IP addresses and services.

Iv. Features of IP multicast technology
The IP multicast technology has the following features.

1. Group address

In a multicast network, each multicast group has a unique multicast address Class D. Some IP multicast addresses are allocated by the Internet management organization, other multicast addresses are used by users as temporary addresses. multicast packets can be sent to the Group addresses that identify the target group. The sender does not have to know which group members are in the group and does not have to be a member of the group, there is no limit on the number and location of hosts in the group members. The host does not need to discuss with group members and senders. You can join or leave multicast groups at will. You do not need to know the location specified by the host when using the Group address, you can find any resource or server with this multicast address, search for the required information in the dynamically changing information provider, or publish the information to an optional user group of any size.

2. scalability

If the network speed is increased, the capacity of the wide area multicast network needs to be increased, later, multicast routing algorithms and protocols, such as PIM-DM, PIM-SM, and CBT, all support network scaling, the Group address and dynamic characteristics mentioned above are also the other aspects that adapt to scalability.

3. robustness

The routing protocols and algorithms used by IP multicast networks can adapt to the dynamic changes of network routes. They use software status refresh mechanisms to create route backups and other methods to maintain connections between group members, strengthen the robustness of the network.

4. Independence of Routing Algorithms

Multicast Routing Algorithms and protocols are independent of the protocols used by unicast routing. However, they rely on the existing unicast route table to dynamically generate multicast trees by adapting to network topology changes in the domain.

5. Flexibility of multicast Spanning Tree

The formation of the multicast spanning tree is related to the distribution of senders and receivers, network traffic conditions, and the dynamic nature of group members, the multicast spanning tree also reflects different Multicast Routing Algorithms and multicast applications. Flexible multicast Spanning Tree facilitates data packet transmission, and does not easily cause network congestion.

V. IP multicast technology
1. IP multicast address allocation

In a multicast network, a multicast group is specified as a Class D address. The value range of the multicast address is 224.0.0.0 to 239.255.255.255, but the address 224.0.0.0 is retained and cannot be assigned to any group.

In the multicast communication model, two new types of addresses are required: an IP multicast address and an Ethernet multicast address. An IP multicast address represents a group of receivers, they need to receive the data sent to the entire group. Because the IP packet is encapsulated in an Ethernet frame, an Ethernet multicast address is also required. To make the multicast model work properly, the host should be able to receive unicast and multicast data at the same time. The host needs multiple IP addresses and Ethernet addresses, unicast IP addresses and Ethernet addresses are used for unicast communication, while Ethernet multicast addresses are used for multicast communication. If the host does not want to receive multicast addresses, it is set to zero. Therefore, the main difference between unicast and multicast addresses is that each host has a unique Unicast address, but not the multicast address.

Ing Class d ip addresses to Ethernet MAC addresses is completed by the data link layer. The process from group broadcasting to the layer-4 address of the Lingling Ring Network is a work procedure adopted by the CISCO router, while the ing between the Ethernet and FDDI networks from multicast to layer-4 is quite direct.

During the ing process, a total of 9 bits are not involved in the replacement of the multicast IP address, including 8 bits in the high byte and a flag next to the byte, the first 4 bits 1110 indicate that they belong to the class d ip address, and the remaining 23 bits are replaced. Replace the low 23 bits in the IP multicast address with the Ethernet multicast address 01: 00: 5E: 00: 00: the value of 00 is 23 bits. Therefore, five digits are not involved in the ing. No matter what the values of these bits are, the multicast Ethernet addresses are the same. Since a total of 32 different combinations can exist for five single digits, The ing is not unique.

2. IGMP

Before a multicast router establishes a route and transmits its multicast group member information, it must determine whether one or more hosts are added to a multicast group on the local network. Therefore, the multicast router and the host that implements multicast must use the IGMP (Internet Group Management Protocol) Protocol to communicate Group member information. Using IGMP, the multicast router can determine whether some multicast group members exist on any network connected to itself, such as group members, the multicast router can join a specific multicast group and forward multicast data to the host that joins the group. Therefore, IGMP is used by the host to notify the directly connected router to join a multicast group to make the multicast network dynamic and flexible.

The initial IGMP specification was defined in RFC 1112 in detail. We generally refer to this specification as "IGMP version 1", written by S. Deering of Stanford University in August 1989. Later, W. Fenner of Xerox PARC significantly updated the earliest IGMP version 1. The updated result is RFC 2236 file, that is, IGMP version 2. Two versions of IGMP can perform a few operations on each other. IGMP version 2 is nearing official approval. Idmrhas started igmp00003 research work. Currently, draftis draw.ietf-idmr-igmp-v3-05.txt. IGMPv1 defines the basic process of querying and reporting group members. IGMPv2 adds a mechanism for members to leave quickly, the main function added in IGMPv3 is that a member can specify to receive or not to receive packets from some multicast sources.

3. layer-2 multicast protocols

Layer-2 multicast protocols of IP networks include IGMP Snooping and CGMP.

IGMP Snooping listens to igmp members of the vro to report messages through the vswitch to form a correspondence between the group members and the vswitch interfaces and put them in the multicast CAM table. Based on this relationship, the vswitch transfers multicast packets to APIs with group members only.

CGMP (Cisco Group Management Protocol) is a private Protocol developed by Cisco Based on the client/server model. It runs on routers and switches, and allows member relationship information to communicate from routers to switches. With the support of CGMP, the multicast router can notify the vswitch when the host is added to or out of the multicast group based on the received IGMP packet, the vswitch uses the forwarding table built by the information to determine the interfaces to which multicast packets are forwarded. GMRP is the standard protocol between the host and the Ethernet switch. It enables multicast users to register multicast members on the second-layer switch.

4. IP Multicast Routing Protocol

In a routing network, IP multicast routing protocol is a crucial issue for transmitting multicast information, it overcomes the bandwidth bottleneck caused by transmitting multicast information using the unicast communication model and reduces the communication cost for sending identical data information to multiple receivers, this is also the main reason for the development of IP multicast applications. The data flow in the multicast network must establish a spanning tree based on the multicast routing protocol, so that a separate forwarding path is formed between the sending source and multicast group members to ensure that each packet can be forwarded to the destination.

IP multicast routing protocols are classified into intra-domain and Inter-Domain protocols. Domain protocols include PIM-SM, PIM-DM, DVMRP, and CBT. Inter-Domain protocols include MBGP, MSDP, and BGMP.

Related Articles]

• IP multicast relies on management to prevent data packet delay or loss

• Use IP multicast technology to transmit video information

• Solution to loop problems caused by multicast Static Routing