|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
– |
some
consequences from model:
|
|
|
|
n |
communication
costs are no longer uniform:
|
|
|
|
– |
relatively
cheap to communicate between two MSS
|
|
|
|
– |
relatively
expensive to communicate MSS-MH or MH1-MH2
|
|
|
|
n |
mobile
hosts may disconnect (or doze off), reappear later
|
|
|
|
n |
underlying
network structure changes all the time
|
|
|
– |
model
applied to study a well-known distributed mutual
|
|
exclusion
algorithm (Lamport’s Algorithm):
|
|
|
|
n |
if
original algorithm implemented directly, poor performance:
|
|
|
|
– |
high
search costs for all MH-MH messages
|
|
|
|
– |
all
hosts required to participate, so disconnections not allowed
|
|
|
|
– |
data
structures need to be maintained at the mobile hosts, so high
|
|
|
|
power
usage
|
|