Mobile computing has hit its stride, and, as a result, many computing strategies are under investigation. An interesting one is Internet suspend/resume (ISR), proposed by the authors of this paper.
Basically, the suggested model focuses more on user mobility than on device mobility, considering that computers will be available everywhere, and accessible as appliances. If this will be the case in the future is a matter of debate, as many users love their privacy and mobile computing devices. However, in specific cases, as presented in the paper, the ISR model makes sense, and may prove to be cost-effective.
The model uses a virtual machine monitor (VMM) for state encapsulation: each VM state maps to a host file. After suspension, the state file can be copied to another machine with similar hardware architecture, and execution can be resumed. Data is transferred in two hops: current machine to server, then server to resume machine.
ISR uses the Coda distributed file system. Several experiments that used different transfer policies were conducted. Unfortunately, resume latency is high, and is very high in some realistic circumstances. This is one reason I think this model cannot be generally applied except in specific situations. Security, as the authors note, is another issue, and the availability of computers and resources can be another. Overall, however, this model is appealing, and is worth investigating.