Hardware Reference
In-Depth Information
Web server via the telephone system using ADSL, which was discussed in Chap. 2.
(Alternatively, cable TV can be used, in which case the left-hand part of Fig. 8-14
is slightly different and the cable company is the ISP.) The user's computer breaks
the data to be sent to the server into packets and sends these packets to the user's
ISP
(
Internet Service Provider
), a company that offers Internet access to its cus-
tomers. The ISP has a high-speed (fiber-optic) connection to one of the regional or
backbone networks that comprise the Internet. The user's packets are forwarded
hop-by-hop across the Internet until they arrive at the Web server.
Most companies offering Web service have a specialized computer called a
firewall
that filters all incoming traffic in an attempt to remove unwanted packets
(e.g., from hackers trying to break in). The firewall is connected to the local LAN,
typically an Ethernet switch, which routes packets to the desired server. Of course,
reality is a lot more complicated than we have shown, but the basic idea of
Fig. 8-14 is still valid.
Network software consists of multiple
protocols
, each one being a set of for-
mats, exchange sequences, and rules about what the packets mean. For example,
when a user wants to fetch a Web page from a server, the user's browser sends a
packet containing a
GET PAGE
request using the
HTTP
(
HyperText Transfer
Protocol
) to the server, which understands how to process such requests. Many
protocols are in use and often combined. In most situations, protocols are struc-
tured as a series of layers. Upper layers hand packets to lower layers for proc-
essing, with the bottom layer doing the actual transmission. At the receiving side,
the packets work their way up the layers in the reverse order.
Since protocol processing is what network processors do for a living, it is nec-
essary to explain a little bit about protocols before looking at the network proc-
essors themselves. Let us go back for a moment to the
GET PAGE
request. How
is that sent to the Web server? The browser first establishes a connection to the
Web server using a protocol called
TCP
(
Transmission Control Protocol
). The
software that implements this protocol checks that all packets have been correctly
received and in the proper order. If a packet gets lost, the TCP software assures
that it is retransmitted as often as need be until it is received.
In practice, what happens is that the Web browser formats the
GET PAGE
re-
quest as a correct HTTP message and then hands it to the TCP software to transmit
over the connection. The TCP software adds a header in front of the message con-
taining a sequence number and other information. This header is naturally called
the
TCP header
.
When it is done, the TCP software takes the TCP header and payload (con-
taining the
GET PAGE
request) and passes it to another piece of software that im-
plements the
IP protocol
(
Internet Protocol
). This software attaches an
IP
header
to the front containing the source address (the machine the packet is com-
ing from), the destination address (the machine the packet is supposed to go to),
how many more hops the packet may live (to prevent lost packets from living for-
ever), a checksum (to detect transmission and memory errors), and other fields.
