comob: a scenario focusing on pervasiveness, distribution, and scale gianpaolo cugola & matteo...
Post on 22-Dec-2015
216 views
TRANSCRIPT
CoMob: A Scenario Focusing on Pervasiveness, Distribution, and Scale
Gianpaolo Cugola & Matteo RossiDeepSE GroupDipartimento di Elettronica e InformazionePolitecnico di Milano, Italy
Milan, 23 Sept. 2009 Meeting on scenarios 2
Why a new scenario
• We feel the need of a scenario centered around pervasiveness, ubiquity, strong distribution, coordination without centralized control– Typical domains: p2p applications, WSNs, IoT, ...
• Scenarios centered on services and service-oriented computing stress a different form of distribution and a different kind of dynamics
• Clearly, we need both kind of scenarios. Possibly, but not necessarily, integrated together in a single “macro-scenario”
Milan, 23 Sept. 2009 Meeting on scenarios 3
CoMob: Cooperative Mobility
• People roam around with their gps-equipped phone
• Collect information about current traffic (along their route) based on the position of others (FCD)
• Collect information about others’ route to predict future traffic
• Coordinate with others to avoid future jams
Milan, 23 Sept. 2009 Meeting on scenarios 4
CoMob: An analysis
• Good– Strong distribution– Large scale– Ubiquitous– Dynamic– Discourages centralized solutions (efficiency, administrative
issues)• Game theory may suggest optimal solution with zero coordination
– Small (limited) devices• Bad
– Fixed requirements– No “open-world”– Small devices but not “really small”
Milan, 23 Sept. 2009 Meeting on scenarios 5
Per-CoMob• Sensors (pressure,
cameras, ...) build a WSN that contribute information about traffic
• Traffic lights contribute controlling routes and limiting traffic jams
• Drivers may prefer a route if they know they will “sync” with green
– But they must coordinate – the number of cars on a street determines the average speed
Milan, 23 Sept. 2009 Meeting on scenarios 6
Open-CoMob
• Before getting in the street, people define what cooperative application they want to run (e.g., a tourist might want to maximize the number of locations visited)– The system builds the application by setting up a
collaboration with other devices• E.g., with devices of other people belonging to a community
“tourists”, to determine at what time it is best to visit a monument