TCP/IP First-Step [Electronic resources] نسخه متنی

What is a router?

Answer: A router is a specialized device designed to push IP packets (or other protocols' packets) through a network. A router relies on routing protocols to actually apply intelligence to gather information about the network and make good decisions about where to send any IP packets it receives.

What does a routing protocol do?

Answer: Routing protocols perform at least six critical functions:

Communicate information about known destination networks with neighboring routers.

Discover new destination networks automatically through communications with neighbors.

Compare different paths to the same destination.

Select the best path to each known destination.

Store all known best paths in a routing table for use in IP packet forwarding.

Send packets of data to their destination using the best known path as recorded in the routing table.

At what layer(s) of the OSI reference model does a router separate different networks?

Answer: Routers separate networks at Layers 1 and 2 of the OSI reference model. Those layers are physical and data link, respectively. Physical networks, such as Ethernet, occupy solely these two layers.

At what layer(s) of the OSI reference model does a router interconnect different networks?

Answer: Routers connect networks at Layer 3, the network layer. The Internet Protocol (IP) operates at Layer 3 and has become the world's dominant protocol for interconnecting networks.

What is a hop and why is it significant in routing?

Answer: A hop is the original distance vector. Each router in the path between a source and destination network counts as one hop. This basic mechanism is the foundation for distance-vector routing and continues to influence modern routing protocols.

What is a distance vector?

Answer: A distance vector is an arbitrary measure of distance in a network.

Why isn't hop counting a basis for efficient routing?

Answer: Hops tend not to correlate well to geographic distances in a network. Nor do hops take into account the amount of bandwidth available or other network conditions that might make certain paths undesirable. Consequently, hop counting is a rudimentary approach to selecting best paths in a network.

What is a potential problem with link-state routing?

Answer: Some types of link-state information are too volatile for direct routing decisions. Information such as a link's load level or the amount of delay currently experienced are too volatile to be useful in selecting best paths. Quite simply, the data would change many times before you could complete the decision-making process. Consequently, link-state routing has the potential to seriously disrupt a network's operation if it is not configured properly.

Which is better: distance-vector or link-state routing?

Answer: It is impossible to state definitely which approach to routing is superior. Rather, each approach to routing comes with its own strengths and weaknesses and, if you understand those qualities relative to your particular needs, you can't make a bad choice. Generally speaking, distance-vector routing is cruder and potentially ineffective, but easy to configure and use. Link-state routing is much more complex and powerful, but that complexity is its Achilles' Heel. If not properly configured, a link-state protocol can wreak havoc in a network.

When would you consider using policy- or rules-based routing?

Answer: Policy-based routing only works well when your primary goal is massive scalability. If you are not building a network the size of the global Internet, you are probably better off selecting a lighter routing protocol, such as a distance-vector or link-state protocol. Of course, if you intend to connect your WAN to the Internet, you might be required to use policy-based routing across that connection.