Switching in IP Networks: IP Switching, Tag Switching, & Related Technologies
By Yakov Rekhter
Chapter 7: Comparison of Label Switching Approaches
Chapter 7: Comparison of Label Switching Approaches
Overview
In this chapter we will examine the major differences between the four approaches to label switching covered in the previous chapters. We begin by reexamining the taxonomy that was introduced in Chapter 2 and seeing where each of the approaches sits in that taxonomy. Then we endeavor to evaluate the consequences of the most important design decisions made by the inventors of each approach. By far the most significant design choice is between a data-driven (or flow-driven) model and a control-driven model. After looking at the effects of that design choice, we look at the significant differences between the approaches in the data-driven and control-driven camps.
After comparing the various label switching techniques, we briefly compare the overall label switching approach with more conventional methods of forwarding IP packets. Because there are several new products that provide high performance IP forwarding without the use of label switching, it is reasonable to ask what benefits, if any, label switching offers over these approaches.
7.1 Taxonomy
In Table 7.1 we have classified the four label switching approaches described in this book in terms of the fundamental design decisions that were discussed in Chapter 2 , such as the type of binding distribution. In Table 7.2, we list the features and functionality that the designers have chosen to support.
Table 7.1 Architectural comparison of label switching approaches.
CSR
IP Switching
Tag Switching
ARIS
Data- or control-driven
Data-driven
Data-driven
Control-driven
Control-driven
Loop prevention
Unspecified
Supported
Copyright Morgan Kauffmann Publishers, Inc. 1998 under license agreement with Books24x7