Question Set 2 for Crowcroft's DTN/Social/Mobile Networking Summer
School course@Lipari, July 2010

1. Forwarding algorithms for opportunistic networks can be stateless or rely
on observed patterns of encounters between nodes. Describe the classes of
forwarding schemes and the costs in terms of state they incur.

2.  A feature of human encounters is that the distribution of intervals between
encounters, and the distribution of duration of encounters, are both heavy
tailed (at least that is a reasonably first approximation). Why might this be?
How does it relate to the degree distribution of nodes in many naturally
occurring networks?

3. "Preserving privacy in forwarding is hard enough. Preventing denial of
service attacks in infrastructure-less networks is almost impossible". Is this
true?

4. Sketch out a systems design for a P2P Ad Hoc Delay Tolerant Mobile
Social Network.

5. Could you design a distributed computation that achieved "eventual
consistency" over a set of nodes communication via opportunistic (DTN) system.

