Delay-tolerant networking has recently emerged as an approach that
might help achieve communication in challenged environments---such as
mobile, wireless ad hoc networks---where end-to-end connectivity
cannot always be obtained. In a delay-tolerant network, messages are
not simply routed or disseminated in the network. While travelling
from host to host, a message can be stored temporarily on certain
hosts, and be forwarded later when circumstances permit. In a dynamic
ad hoc network, where hosts can be present only sporadically, this
approach makes it possible for a message to eventually reach a target
host, even if this host was not reachable at the time the message was
sent originally. Moreover, delay-tolerant networking can allow message
delivery in a partitioned network, using mobile hosts as carriers that
can transport messages between non-connected fragments of the
network. Project SARAH relies on the assumption that delay-tolerant
networking will be widely adopted in the near future as a means to
support asynchronous communication in autonomous, partitioned ad hoc
environments. Based on this hypothesis, we plan to consider the issues
pertaining to the deployment and the utilization of distributed
services on mobile devices capable of delay-tolerant, ad hoc wireless
communication. Since no de facto standard for delay-tolerant
networking has emerged yet, we plan to devise our own model for
delay-tolerant ad hoc networking, and to implement a proof-of-concept
middleware platform complying with this model. This platform will then
serve as a framework for investigating the notion of delay-tolerant
distributed services, that is, services that can be deployed in a
partitioned ad hoc environment, and discovered and invoked thanks to a
judicious combination of proximal and delay-tolerant
communication. While devising these delay-tolerant communication
middleware and services, special attention will be paid to the
security issues introduced by the lack of end-to-end connectivity in
the environment considered. Moreover, abstract modelling and
simulation techniques will be used to evaluate and tune our
communication model, and to observe how realistic delay-tolerant
services can behave in a variety of networking conditions.
|