Apr 1999 NI222: Networks and the Internet (Question 5)

(a)	What is the purpose of routing? Of which layer is it a part?	[2]
	Routing is to find the best possible, cost effective path between two nodes communicating with each other. Routing can be viewed as static, dynamic centralised and decentralised. It is a part of the network layer.
(b)	Routing algorithms use a number of link attributes in computing path costs, to determine the minimum cost path. List three such attributes.	[3]
	Physical length of the link Communication cost Propagation delay
(c)	Explain the concept of hierarchical routing, and how it applies in the Internet.	[7]
	In Internet, there is mainly two routing levels : Router Level and Domain Level. It is difficult for a router to compute. The routing paths for the whole network. In Internet where large number of networks are connected, it is not even feasible. So we can divide Internet into different domain. Each domain will be conisdered as a sparate network and router nodes all suppose to compute the path in their own network. Routers are used to do routing of network are not the same. Paths are computed by routers. If Domain is bif,we can divide domain into sub-domain. Hence, we have Exterior gatways and interior gatways. Exteriros gateways are for different AS (automonous systems) while Interior gateways are for the sams AS. They are being replaced by Border gateways and IRP,etcs.
(d)	Asynchronous Transfer Mode (ATM) has some similarities with IP, and many differences. Explain the key features of ATM, being sure to mention its cell format, and approach to routing.	[5]
	Asynchronous Transfer Mode is the fastest technique available now having transmission speed of 660 Mbs as compare to 100 Mbs of the conventional methods. It is a standard technique which solve the incompatibilities problem faced by others. The best thing is that it support different kind of media i.e, coaxial, fibers, wires, etc. It is independent of physical media going to be used. It is fixed size cell 53 bytes where 5 bytes are used for header and 49 bytes used for data payload. It is good for real time i.e video conferencing because of fixed size cell, no data can get priority. One feature is that it is connection oriented transmission which adds odd on the packet nature of data. These are three layers : ATM adaptation layer :- mapping information into and out of cell. ATM layer : - do addressing Physical layer : - physically transmit data bits on a physical media. There are two identifiers used in ATM: Virtual path identifier Virtual channel identifier
(e)	Suppose you could run the transport layer TCP directly on top of ATM. Would this be better than using TCP over IP, for an interactive telnet session? Explain your answer.	[3]
	This would probably be a good match. The telnet/TCP connection will need a long-term connection from one host to another, and ATM would provide that as its network layer service. Being interactive, the telnet session will probably send very few characters at a time, which is ideal for ATM's small cells, compared to IP's large packets.