Chapter 3.2 - Data Traffic Concepts

3.2.1 Natural Information Rate

Data is not generated at a continuous and constant flow. In fact, data flow is quite
variable, varying between a maximum or peak rate and a minimum rate. This is
known as the natural information rate (Figure 3.1).

Based on this fluctuating data rate, as packets are being formed, one can easily envision
two different scenarios. If packets have a fixed length (such as in ATM), then
packets are created at random times, the randomness of which is correlated with the
natural information rate and the packet length. Conversely, if packets have variable
length, then packets may be generated in a more periodic manner (Figure 3.2). Thus,
in addition to the natural information rate, one can define a packet rate (as the number
of packets in the unit of time), a distribution of interpacket time intervals (as the
distribution of time intervals between packets; this has a statistical profile such as
Gaussian, Poisson, etc.), distribution of packet length over time, and so on.

The ratio peak rate to average rate defines the burstiness of flow. Clearly, the
more uniform the flow is, the less the burstiness is, as the difference between maximum
and average tends to be small, as is the case for voice. Conversely, as the difference
becomes larger, the burstiness increases, as is the case for telemetry data.
However, one would expect that as the peak increases, so does the burstiness. Thus,
depending on type of traffic, a curve can be drawn (peak versus burstiness) that defines
a boundary separating the applications space in two areas, one that is better
suited for synchronous applications and one for asynchronous applications (Figure
3.3).

In addition to packet rate fluctuation, there is another basic difference between
data and synchronous traffic—the data flow is not symmetric. What this really

Figure 3.2 The information flow determines two different scenarios, one with fixed-length and one with variable-length packets.

Figure 3.3 Peak rate versus burstiness separates the application space.

means is that the traffic flow is not the same in both directions. This is particularly
true in Internet traffic or large data transfers. The traffic flow asymmetry also fluctuates.
The fluctuation depends on several factors such as time of the day, day of the
year, geographical distribution of subscribers, protocol used, and scheduled versus
unscheduled events; the latter may cause unpredictable network congestion.

< Previous Excerpt Next Excerpt >

Purchase This Book

TABLE OF CONTENTS

Chapter 3.2 - Data Traffic Concepts

Contact Preferences

This is embarrasing...

Customize Your GlobalSpec Experience

Select Your Free Newsletters

Industry Newsletters

Select Your Free Product Alerts

This is embarrasing...