Design  | SFC  |  NOS  |  Code |  Diary  | WWW |  Essay |  Cisco | [Home]



[Bill's Home Page]
[ Old page]


Click on me, if you just want to see the magazine

Date: Wednesday, 15 August 2001
Ed: Billatnapier
Source: Chapter 14, Mastering Computing
Price: £2 (UK)/ $3 (US) - Price reduced
This is another special edition of the Internet Agent, and presents some of the most asked questions relating to the Internet. More information can be found in Mastering Computing.
[Show only magazine] [Previous edition]
Question. Okay. I understand that both the MAC address and the IP address need to be specified for a node to receive data, but how does a node know the MAC address of the remote destination?
A host looks up its local ARP cache (which is in its own RAM, and not stored to the permanent storage) to see if it knows the MAC address for a known IP address.
If it does not find the MAC address, it transmits an ARP request to the whole of the network (ARP requests do not travel over routers). The host who matches the transmitted IP address then responds with an ARP reply with its own MAC address in the source address field in the data frame. This is received by the originator of the request, which updates its local ARP cache, and then transmits with the required MAC address.

Question. If a computer has no permanent stor-age, how does it know its own IP address?
Diskless hosts use the RARP protocol, which broadcasts a message to a RARP server. The RARP server looks-up the MAC address in the source address field in the data frame and sends back its IP address in a reply to the host.


Question. Oh, yes. I think I see it now, but what if the destination is on another network, possibly in another country, how does it determine the address of the destination?
The host knows the IP address of the gateway for the network (normally a router). It then uses the MAC address of the gateway, but with the destination IP address of the host that the data is destined for. The gateway senses that the data frame is addressed to it-self, and forwards it to the next gateway, and so on.
If the node does not know the MAC address of the gateway it will send out an ARP request to the network with the IP address of the gateway.

Question. How is it possible to simply connect a computer to an Ethernet network, and all the computers on the network are able to communicate with it, and how do they know when a computer has been disconnected?
Computers use the ARP protocol, which allows nodes to determine the MAC address of computers on the network, from given IP addresses. Once they discover the destination MAC address, they update their ARP cache. After a given time, the entries in the table are updated (known as aging the entry).


Question: What's the difference between a data segment and a data packet?
The transport layer uses data segments, whereas the network layer uses data packets. Data segments allow two or more applications to share the same transport connection. These segments are then split into data packets which have a given maximum size (typically for IP packets this is 64KB) and each are tagged with a source and destination network address. Different applications can send data segments on a first-come, first-served basis.

Question: How does a node broadcast to the network?
There are two types of broadcasts. The first is a flooded broadcast, which has 1's in all parts of the IP address ( The other is a directed broadcast, which has all 1's in the host part of a IP address. For example, to broadcast to the network, the broadcast address is 146.176.151. 255, as all 1's in the host part of the address specifies a broadcast. Routers forward directed broadcasts, but not flooded addresses (as these are local). All hosts and routers must thus know what the subnet mask is.