28.06.2014 Views

Learning by Doing: CISCO Certified Network ... - SCN Research

Learning by Doing: CISCO Certified Network ... - SCN Research

Learning by Doing: CISCO Certified Network ... - SCN Research

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Ethernet Packet Structures<br />

Objective:<br />

To learn about the structure of Ethernet packets.<br />

Background:<br />

So far we have been talking about networking and packets passing over the network. In<br />

this lab we will look at the precise structure of packets. Later when we use protocol<br />

inspectors you will be able to understand the information better.<br />

Ethernet<br />

Ethernet generally refers to a standard developed <strong>by</strong> a consortium of the Digital<br />

Equipment Corporation (DEC), Intel, and Xerox. It is one of the most widely used<br />

encapsulation standards in use for networking today. There have been many versions and<br />

revisions to it over the past twenty years. So trying to “nail-down” the exact structure of<br />

an Ethernet packet is as easy as nailing jello to the wall. Simply put, you need to be more<br />

specific about which Ethernet packet structure you want to examine. There have many<br />

different types of Ethernet, or “flavors” if you will, and we will look at the two most<br />

common ones: the “generic Ethernet” and “Ethernet SNAP.” Basically our two Ethernet<br />

packet structures are the same except the SNAP packet uses part of the data field for LLC<br />

sub-layer and SNAP information. In either case the minimum/maximum size of our<br />

Ethernet packet is 64-1518 <strong>by</strong>tes. If the information in the data field will be smaller than<br />

the minimum size allowed then it will be “padded” with contiguous zeros to fill the data<br />

field up to the minimum size.<br />

802.2/802.3 Ethernet (RFC 894)<br />

Preamble SOF DA SA Type Data FCS<br />

Figure 1—Generic Ethernet packet structure.<br />

This “Standard for the Transmission of IP Datagrams Over Ethernet <strong>Network</strong>s” was<br />

written <strong>by</strong> Charles Hornig in 1984 (ftp://ftp.isi.edu/in-notes/rfc894.txt ).<br />

Stripped <strong>by</strong> the NIC:<br />

The preamble can vary in length. The preamble basically is used to help set up the<br />

transmission and reception of the information through synchronization. The actual<br />

amounts of bits have varied over the years but the principle is still the same: a series of<br />

alternating zeroes and ones encompass the preamble. Some of these can be lost during<br />

transmission but that is ok. The incoming stream of bits “establishes” that the reception<br />

of a packet has started. Most agree on 62 bits. (In hex: 1555555555555 In binary:<br />

010101010101010101010101010101010101010101010101010101010101010101010101<br />

0101010101). You will not see this with a protocol sniffer because it is stripped and<br />

dumped.<br />

75

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!