In recent years, with the rapid popularization and explosive development of the Internet, many new applications have emerged on the Internet, many of which are high-bandwidth multimedia applications, such as online video conferencing, online audio / video broadcasting, stock market quotations, multimedia remote Education, large-scale collaborative computing, etc. This brings about the problem of rapid consumption of bandwidth and network congestion. In order to alleviate network bottlenecks, people have proposed various solutions, one of which is the use of IP MulTIcast (multicast, multicast or multicast) technology, which is a technology that can maximize the use of existing bandwidth. In comparison, IP multicast technology has its unique advantages. In a multicast network, even if the number of users grows exponentially, the backbone bandwidth does not need to increase.
1. Brief History of IP Multicast Development
In the mid-1980s, Stanford University implemented the first multi-purpose call. PhD student SE Deering published two papers: Host group: A mulTIcast extension to the Internet Protocol (RFC0966) and Host extensions for IP MulTIcasTIng (RFC0988). The possibility of IP multicast is proposed.
In 1988, D. Waltzman, C. Portridge, SE Deering published an article entitled "Distance Vector Multicast Routing Protocol" (RFC1075), which was the first practice of multicast routing protocol;
In December 1991, SE Deering published his doctoral thesis "Multicast Routing in Datagram Interconnection Networks" (RFC1112). It laid the foundation of the multicast network architecture and routing protocol. This article has also become the prototype of the Internet Group Management Protocol (IGMP);
In March 1994, the OSPF protocol extension protocol MOSPF (RFC1584) was formed;
In 1995, Cisco began selling routers and switches that support multicast
In November 1997, the group management protocol IGMPv2 was approved by the IETF as a standard (RFC2336);
In June 1998, the IETF standard for evaluating the reliable multicast transmission protocol RMTP was introduced (RFC2357);
In July 1998, when formulating the IPv6 address system standard, the IPv6 multicast address allocation scheme (RFC2373) was determined, which made the necessary preparations for the application of multicast technology on the next generation Internet;
From the end of 2000 to the beginning of 2001, people began to develop various multicast MIB libraries, which indicates that the multicast technology is developing in a manageable and controllable direction.
Second, the multicast network architecture
The multicast network architecture includes: the basic working principle of multicast, the conditions for implementing multicast, the address allocation scheme of multicast and the mapping with MAC addresses, and the Internet group management protocol.
1. The working principle of multicast
Multicast is a network technology that allows one or more senders (multicast sources) to send a single data packet to multiple receivers (once, simultaneously). The multicast source sends the data packet to a specific multicast group, and only the addresses belonging to the multicast group can receive the data packet. general speaking,
The host joins the wildfire group by using the INTERNET group management protocol, and can leave the group dynamically, that is, the membership relationship often changes. The router tracks this relationship and attempts to form a loop-free path to the multicast members. Some multicast channels have been used to obtain the routers on the path of the multicast group being used, and the best path information to reach these multicast groups. Once the message reaches the target LAN, the message may be flooded or forwarded to the host. The three transmission methods are compared as follows:
Unicast transmission: A separate data channel is required between the sender and each receiver. If a host transmits data to a small number of recipients at the same time, there is generally no problem. However, if there are a large number of hosts wishing to obtain the same copy of the data packet, it is difficult to achieve. This will cause the sender to have a heavy burden, long delay, and network congestion. To ensure a certain quality of service, hardware and bandwidth need to be increased.
Multicast (Multicast) transmission: It improves the efficiency of data transmission. Reduced the possibility of congestion in the backbone network. The hosts in the multicast group can be on the same physical network or they can be from different physical networks.
Broadcast (Broadcast) transmission: refers to the broadcast data packets in the IP subnet, all hosts inside the subnet will receive these packets. Broadcasting means that the network delivers a packet of data to all subnet hosts, regardless of whether these hosts are willing to receive the packet. The scope of broadcast is very small, and it is only valid in the local subnet, because the router will isolate the broadcast communication. Broadcast transmission increases the overhead of non-receivers.
2. Prerequisites for IP multicast
To achieve IP multicast transmission, the multicast source and receiver and the underlying network between the two must support multicast. That is, the host's TCP / IP implementation supports sending and receiving IP multicast; the host's network interface supports multicast; there is a set of group management protocols for joining, leaving, and querying, that is, IGMP (v1, v2); there is a set of IP address allocation strategies , And can map the third layer IP multicast address to the second layer MAC address; application software that supports IP multicast; all routers and switches between the multicast source and receiver must support multicast; Cisco ’s The router not only supports DVMRP, PIM routing protocol, IGMP group management protocol, but also supports Cisco's proprietary Cisco group management protocol CGMP. For intermediate routers that do not support IP multicast transmission, IP tunneling technology is used as a transition solution.
Three, the realization of multicast
There are four problems in IP multicast technology: the first is to whom it is sent, the second is how the receiver receives multicast information, the third is how the user host informs the router that it is no longer interested in a group, and the fourth This is how the router forwards multicast information.
1. Multicast address allocation and MAC address
In multicast communication, we need two kinds of addresses: an IP multicast address and an Ethernet multicast address. Among them, the IP multicast address identifies a multicast group. Since all IP packets are encapsulated in Ethernet frames, a multicast Ethernet address is also required. In order for multicast to work properly, the host should be able to receive unicast and multicast data at the same time, which means that the host needs multiple IP and Ethernet addresses. The IP address scheme specifically delineates an address range for multicast. In IPv4, it is a class D address, ranging from 224.0.0.0 to 239.255.255.255, and class D addresses are divided into local link multicast addresses, reserved multicast addresses, Management authority multicast address.
Partial link address: 224.0.0.0 ～ 224.0.0.255, used for local area network, the router does not forward IP packets belonging to this range;
Reserved multicast address: 224.0.1.0 ～ 238.255.255.255, used for global scope or network protocol;
Management authority address: 239.0.0.0 ～ 239.255.255.255, used internally by the organization, to limit the scope of multicast;
Ethernet multicast MAC address mapping method: All IP multicast frames use MAC layer addresses starting with the 24-bit prefix of 0X0100.5EXX.XXXX, but only half of the MAC addresses can be used by IP multicast, and the remaining MAC addresses The 23 bits of the space are used as the mapping of the layer 3 IP multicast address into the layer 2 MAC address. Since the 28-bit address of the third-layer IP multicast cannot be mapped to the available MAC address space of only 23 bits, 32: 1 addresses are not clear, so the host CPU must make a judgment on each multicast packet received . This increases the CPU overhead of the host. In addition, there is also the problem of suppressing the multicast proliferation of Layer 2 LAN switching.
2. Group Management Protocol IGMP
The host uses IGMP to notify the subnet multicast router that it wants to join the multicast group; the router uses IGMP to query whether there is a host belonging to a multicast group in the local subnet.
1) Join a multicast group
When a host joins a multicast group, it notifies the multicast router of the IP subnet where it is located through the "Membership Report" message, and prepares its own IP module to start receiving multicast data from the multicast group. The data from the group. If this host is the first host in the IP subnet where it joins the multicast group, the multicast router joins the multicast distribution tree through the exchange of routing information.
After joining, the network interface card of the receiver's host begins to listen to the multicast MAC address related to the multicast group address. The router sends the sender's packet to the receiver's network segment hop by hop. The group address in the information packet is converted into a MAC address related to it. The receiver listens to this address, and after receiving the information packet, the multicast packet at the IP layer is taken out and transmitted to the upper layer.
2) Exit the multicast group
In IGMP v1, when a host leaves a certain multicast group, it will exit on its own. The multicast router periodically uses the "membership query" message to query the group address (224.0.0.1) of all hosts in the IP subnet. If a multicast group does not have any members in the IP subnet, the multicast router is in After confirming this event, the data of this multicast group will no longer be forwarded on the subnet. At the same time, through the exchange of routing information, the corresponding multicast router is deleted from the specific multicast group distribution tree. This method of leaving quietly without notifying anyone makes the multicast router know that there is no member in the IP subnet for a period of time. In IGMP v2.0, when each host leaves a multicast group At this time, the subnet multicast router needs to be notified, and the multicast router immediately queries all the multicast groups in the IP subnet, thereby reducing the delay in processing the system to stop the multicast.
3. Multicast forwarding
1) Reverse Path Forward (RPF)
When the multicast data packet arrives at the router, the router performs an RPF check to decide whether to forward or discard the data packet. If it succeeds, it is forwarded, otherwise it is discarded. The RPF check process is as follows:
Check the source address of the packet to determine whether the interface through which the packet passes is on the path from the source to this;
If the data packet arrives from the interface that can return to the source host, the RPF check is successful, and the data packet is forwarded to all the interfaces on the output interface table, otherwise the RPF check fails and the data packet is discarded.
2) Multicast forwarding cache
Performing RPF check on each incoming multicast packet will cause a large loss of router performance. Therefore, when establishing a multicast forwarding cache, the RPF interface is usually determined by the multicast route. Then change the RPF interface to the input interface of the multicast forwarding cache entry. Once the routing table used by the RPF check program changes, the RPF interface must be recalculated; and the multicast forwarding cache entry must be updated.
3) TTL threshold
Whenever a router forwards a multicast packet, the TTL (Time To Live) value in the IP packet decreases by 1. If the TTL of a packet decreases to 0, the router will discard the packet. The TTL threshold can be used for each interface of the multicast router to prevent the forwarding of multicast packets below the TTL threshold on this interface. This can control the scope of multicast.
4) Management authority boundary
In addition to the TTL threshold, multicast provides another address mechanism called management authority as a boundary to limit the forwarding of multicast information outside the domain. The multicast address for administrative rights is from 239.0.0.0 to 239.255.255.255. This address is considered to be locally allocated (similar to 192.168.xx.xx in unicast) and cannot be used on the Internet. This mechanism makes it possible to reuse multicast addresses within the intranet and improve the utilization of multicast address space.
4. Multicast tree
In the unicast model, data packets are transmitted from the source host to the target host along a single path through the network, but in the multicast model, the multicast source transmits data packets to a certain group of addresses, but this address represents a group . In order to transfer data to all receivers, a multicast distribution tree is generally used to describe the path that IP multicast travels through the network.
There are two basic types of multicast distribution trees: active trees and shared trees.
1) Active tree
Active trees are also called source-based trees or shortest path trees (Shortest Path Tree: SPT). It is a distribution tree with the shortest path from the root to all receivers constructed with the multicast source as the root. If there are multiple multicast sources in the group, a multicast tree must be constructed for each multicast source. Since the data packets sent from different multicast sources are scattered on separate multicast trees, the use of SPT is beneficial to the balance of data traffic in the network. At the same time, because the path from the multicast source to each receiver is the shortest, the end-to-end delay performance is better, which is beneficial to real-time media applications with large traffic and high delay performance requirements. The disadvantage of SPT is that it is necessary to construct its own distribution tree for each multicast source. When the data traffic is not large, the overhead of constructing SPT is relatively large.
2) Shared tree
The shared tree, also known as RP tree (RPT), refers to a multicast tree established by selecting a common root (convergence point RP or core) for each multicast group with RP as the root. The multicast source of the same multicast group unicasts the data to be multicast to the RP, and then the RP forwards it to other members. At present, the two most shared and most representative shared trees are the Steiner tree and the core tree (CBT).
The shared tree has better performance in both the amount of state information that the router needs to store and the total cost of the routing tree. When the size of the group is large and the data transmission rate of each member is low, it is more appropriate to use a shared tree. But when the traffic is heavy, the use of shared trees will cause traffic concentration and bottlenecks near the root (RP).
Multicast, multicast is a thing. He is similar to broadcast, but only interested hosts will receive it. Those who are not interested are directly filtered out by the network card. Broadcast is different. Broadcasts always reach the network protocol layer. In addition, multicast can block network segments and reach any network, but broadcast can't.
There is a certain range of broadcast addresses.
white Color paraffin wax Fluted Candle also called velas, bougies, house hold candles, the candles surface with lines. Very beatiful .usually packed by color polybag. as client design, we also can design for client.
and 6pcs in polybag ,and 25bag/.ctn the weight 30-75g per piece to africa countries.
different market need different candles size ,we can supply best price as popular size
the shipment in the 30days after get the depsoit .
Fluted Candle
Fluted Candle,Large Fluted Candle,White Decor Fluted Candle,Bright Fluted Candles
Shijiazhuang Zhongya Candle Co,. Ltd. , https://www.zycandlefactory.com