|
IP multicasting is the capability of a network server to transmit a data stream to more than one host at a time. Consider a scenario in which a company wants to download a streaming video of a training exercise to 20 users sitting at separate workstations.
If a server has to transmit 20 individual copies to the 20 workstations (unicast), a very high bandwidth signal will be necessary. If the server could instead multicast one copy of the video stream to the 20 workstations, a much smaller bandwidth would be necessary. To multicast, the server could insert a Class D address into the IP datagram, and each of the 20 workstations could tell its IP software to accept any datagrams with this Class D address.
Although IP multicasting has some very promising advantages, it suffers from a lack of security. Because it is relatively easy for a workstation to tell its IP software to accept a particular Class D address, any workstation – even one that doesn’t have permission – could potentially receive the multicast. Nonetheless, new applications that address the shortcomings of IP multicasting have been recently announced, opening the door for future possibilities.
When a company applied for a set of Internet addresses that used classful addressing, a number of options were possible. First, the company could apply for a Class B address. With a Class B address, it could allocate 65,536 workstations on its Class B network. If this number of workstations were too large, the company could consider a Class C address. A Class C address allows for 256 computers on a single network. But if the company had 400 to 500 users, a Class C address would be too small. Therefore, the company could apply for two Class C addresses.
Suppose that a company did apply for and receive a large number of Class B addresses. How could it use these efficiently? In this case, the company can break its IP addresses into subnets using subnet masking, a technique that allows for easier network management. The basic idea behind subnet masking is to take the host ID portion of an IP address and further divide it into a subnet ID and a host ID. Using this technique, an ISP and a company can take a large number of host IDs and break them into subnets. Each subnet can then support a smaller number of hosts.
With classless addressing companies (users) do not apply for a particular class of addresses. Instead, a company will get its IP addresses from an Internet service provider (ISP). Most ISPs have already applied for a large number of IP addresses and are willing to lease those addresses to companies. Instead of applying for two Class C addresses, a company with 400 to 500 users could contact an ISP and request to lease 500 IP addresses. These addresses, at least from the company’s perspective, are not identified by any class – they are simply a contiguous block of IP addresses. As you might have guessed, classless addressing has led to a much more efficient allocation of the IP address space. A company can now lease the exact number of addresses it requires, without wasting unused addresses.
When classless addressing is combined with DHCP and NAT, a company can efficiently use a smaller number of IP addresses. This is due to the fact hat all three concepts more efficiently make use of the IP address space. Even though a 32-bit IP address supports more than 4 billion addresses, classful addressing was so wasteful that there was serious concern among industry experts that the supply of IP addresses would soon be depleted. The advent of classless addressing has considerably helped slow the IP address drain. Despite this, there is a push to incorporate a new version of the Internet Protocol, IPv6, which has a much larger address space.
|
