| Quality of Service (QoS) is one of the most elusive, confounding, and confusing topics in
data networking today. Why has such an apparently simple concept reached such
dizzying heights of confusion? After all, it seems that the entire communications industry
appears to be using the term with some apparent ease, and with such common usage, it
is reasonable to expect a common level of understanding of the term.
Our research into this topic for this book has lead to the conclusion that such levels of
confusion arise primarily because QoS means so many things to so many people. The
trade press, hardware and software vendors, consumers, researchers, and industry
pundits all seem to have their own ideas and definitions of what QoS actually is, the
magic it will perform, and how to effectively deliver it. The unfortunate byproduct,
however, is that these usually are conflicting concepts that often involve complex,
impractical, incompatible, and non-complementary mechanisms for delivering the desired
(and sometimes expected) results. Is QoS a mechanism to selectively allocate scarce
resources to a certain class of network traffic at a higher level of precedence? Or is it a
mechanism to ensure the highest possible delivery rates? Does QoS somehow ensure
that some types of traffic experience less delay than other types of traffic? Is it a dessert
topping, or is it a floor wax? In an effort to deliver QoS, you first must understand and
define the problem. If you look deep enough into the QoS conundrum, this is the heart of
the problem—there is seldom a single definition of QoS. The term QoS itself is an
ambiguous acronym. And, of course, multiple definitions of a problem yield multiple
possible solutions. |
