UMass DOME 

Networks composed of mobile nodes inherently suffer from intermit-
tent connections and high delays. Performance can be improved by
adding supporting infrastructure, including base stations, meshes,
and relays, but the cost-performance trade-offs of different designs
is poorly understood. To examine these trade-offs, we have deployed
a large-scale vehicular network and three infrastructure enhance-
ment alternatives. The results of these deployments demonstrate
some of the advantages of each kind of infrastructure; however,
these conclusions can be applied only to other networks of similar
characteristics, including size, wireless technologies, and mobility
patterns. Thus we complement our deployment with a demonstrably
accurate analytical model of large-scale networks in the presence
of infrastructure.

Based on our deployment and analysis, we make several funda-
mental observations about infrastructure-enhanced mobile networks.
First, if the average packet delivery delay in a vehicular deployment
can be reduced by a factor of two by adding x base stations, the
same reduction requires 2x mesh nodes or 5x relays. Given the high
cost of deploying base stations, relays or mesh nodes can be a more
cost-effective enhancement. Second, we observe that adding small
amount of infrastructure is vastly superior to even a large number
of mobile nodes capable of routing to one another, obviating the
need for mobile-to-mobile disruption tolerant routing schemes.
use of open base stations is free, finding volunteers that permit