1. Technical background
In many LAN technologies, each station can “hear” each packet transmission, so the “destination” field must be included in each packet. The “source” field is also required in order for the destination site to know which site sent the packet. Also, to prevent each packet from generating a software interrupt, the LAN adapter filters out those not intended for the site.
2. Propose a plan – also check the 9tut pdf
The 802 committee needs to standardize addresses for its LANs. The first decision is to set the length of the address field. The committee apparently felt that it would be better if the standard had a fixed length, and it would be better to standardize on several lengths. The 802 committee recommended that the LAN (except 802.6) use 48-bit or 16-bit addresses, and after comprehensive consideration, 48-bit was finally adopted.
The 48-bit address view is that they enable the device to obtain a globally unique identifier from the factory. This makes the network truly plug-and-play, because users can directly access the network without assigning addresses manually after purchasing the device.
Globally unique addresses are possible because there is a global agency responsible for distributing blocks of addresses. First Xerox, now IEEE. Every equipment manufacturer who wants to manufacture a plug-and-play LAN device must first contact the agency and obtain an address.
The 802 committee wasn’t sure that everyone was willing to spend the energy (and money) to acquire blocks of addresses from global institutions, so it designated another bit of the 48 bits to identify whether addresses were allocated globally or locally. The global/local flag will be set to 0 when the vendor purchased the address block from the global organization. If the global/local flag is set to 1, people can use the address however they want. When using local addresses, address assignment should be performed by network administrators and ensure that there is no address conflict (that is two sites use the same address).
3.New problems and solutions
Address conflicts are a significant problem when two networks merge.
Group addresses are also sometimes called “multicast addresses”. Often used to discover appropriate neighbors (nodes on the same link), which can be solved in one of two ways:
1) Request: Suppose there are one or several sites of a certain type on the network, such as domain name servers, routers, file servers, etc., and site A wants to access these servers. Site A could be assigned the addresses of all these nodes, but it would be better if site A didn’t need to know a particular server in advance, it only needed to know a single group address Z server (where Z refers to any kind of service, like the one mentioned above Those ones).
When site A wants to access a Z server, it sends packets destined for the Z server. All Z servers will receive and respond to packets addressed to this address.
2) Advertisement: Another way to use group addresses is to define an address for the site listening service. Clients of the Z server do not need to send to the Z server for help, but the Z server periodically sends packets to the Z client at this address. The Z client will receive all packets addressed to the Z client, and then based on the source address of the packet or some other field explicitly included in the data part, the Z client will know the address of a Z server.
spoto Practice Exam