This chapter is concerned with the practical problem of vantage when collecting data on a network. At the conclusion of this chapter, you should have the necessary skills to break an accurate network diagram into discrete domains for vantage analysis, and to identify potential trouble spots.
As with any network, there are challenges involving proprietary hardware and software that must be addressed on a case-by-case basis. I have aimed, wherever possible, to work out general cases, but in particular when dealing with load balancing hardware, expect that things will change rapidly in the field.
The remainder of this chapter is broken down as follows. The first section is a walkthrough of TCP/IP layering to understand how the various layers relate to the problem of vantage. The next section covers network vantage: how packets move through a network and how to take advantage of that when instrumenting the network. Following this section is a discussion of the data formats used by TCP/IP, including the various addresses. The final section discusses mechanisms that will impact network vantage.
Computer networks are designed in layers. A layer is an abstraction of a set of network functionality intended to hide the mechanics and finer implementation details. Ideally, each layer is a discrete entity; the implementation at one layer can be swapped out with another implementation and not impact ...