Quantizing Radio Link Data Rates to Create Ever-changing Network
Conditions in Tactical Networks
Abstract
Several sources of randomness can change the radio link data rate at the
edge of tactical networks. Simulations and field experiments define
these sources of randomness indirectly by choosing the mobility pattern,
communication technology, number of nodes, terrain, obstacles and so on.
Therefore, the distribution of change in the network conditions is
unknown until the experiment is executed. We start with the hypothesis
that a model can quantize the network conditions, using a set of states
updated within a time window, to define and control the distribution of
change in the link data rate before the experiment is executed. The goal
is to quantify how much variation in the link data rate a tactical
system can handle and how long it takes to resume IP data-flows after
link disconnections. Our model includes functions to combine patterns of
change together, transforming one pattern into another, jumping between
patterns, and creating loops among different patterns of change. We use
exemplary patterns to show how the change in the data rate impacts other
link metrics, such as latency and jitter. Our hypothesis is verified
with experiments using VHF radios over different patterns of change
created by our model. We compute the inter-packet latency of three types
of IP data-flows (broadcast, unicast and overlay) to highlight the time
to resume data-flows after long link disconnections. The experimental
results also support the discussion on the advantages and limitations of
our model, which was designed to test tactical systems using military
radios.