Transcript Guohong Cao

Research Projects in the Mobile Computing
and Networking (MCN) Lab
Guohong Cao
Department of Computer Science and Engineering
The Pennsylvania State University
http://www.cse.psu.edu/~gcao
Mobile Computing and Networking (MCN) Lab
• MCN lab conducts research in many areas of wireless
networks and mobile computing, emphasis on designing and
evaluating mobile systems, protocols, and applications.
– Current Projects: smartphones, wireless sensor networks, vehicular
networks, wireless network security, data dissemination/access in
wireless P2P networks, resource management in wireless networks.
– Support: NSF (CAREER, ITR, NeTS, NOSS, CT, CNS), Army
Research Office, NIH, DoD/Muri, DoD/DTRA, PDG/TTC and
member companies Cisco, Narus, Telcordia, IBM and 3ETI.
• Current students: 10 PhD, 1 MS, and 2 visiting scholars
– Alumni: 11 PhD, including faculty members at Iowa State University,
University of Tennessee, Frostburg State University, and students in
Qualcomm, Cisco, Microsoft.
– 12 MS students went to various companies
Outline
• Efficient Energy-Aware Web Browsing in
Wireless Networks
• In-Network Storage
2
Web Browsing in 3G Networks
• Smartphones in 3G networks:
– Increasingly used to access the Internet
– Consume more power
• Current status:
– Limited computation capability, data
transmission distributed whole loading time
– 3G radio interface always on, timer control
– Radio resource is not released, reduce network
capacity
3
Characteristics of 3G Radio interface
T1 = 4 sec
T2 = 15 sec
Intuitive Approach
Latency
Power
Whether to switch DCH->IDLE: Need predict next interval
Our Solution
• Reorganize the computation
sequence of the web browser when
loading a web page.
– First run the computations that will
generate new data transmissions.
– 3G radio interface into low power
state, release the radio resource
– Then run the remaining computations
• After a webpage is downloaded,
predict the user reading time on the
webpage (Gradient Boosted
Regression Trees (GBRT)
– This time > a threshold: switch
into low power state
Evaluations
• The prototype:
– Android Phones
– T-Mobile 3G/UMTS network
• Implement the prototype and collect real traces
• Experimental results:
– Reduce power consumption:
– Reduce loading time:
– Increase network capacity:
30%
17%
19%
7
Computation Limitation
8
VM-based Proxy
• The prototype:
– Xen virtual machines
– Android Phones
– UMTS network
• Experiment Result:
– Reduces the delay by 60%
– Reduces the power by 35%.
10
Outline
• Efficient Energy-Aware Web Browsing in
Wireless Networks
• In-Network Storage
– Social-Aware Data Dissemination in Mobile
Opportunistic Networks
– Cooperative caching in MANET
11
Data Dissemination in Mobile
Opportunistic Networks
• System development: recording users’ interests
– Data access via Samsung Nexus S smartphones
– Categorized web news from CNN
• Application scenarios
– Disaster recovery, military environment
– Public commute systems: bus, subway
– Public event sites: stadium, shopping mall
webpage
XML format
Android
2.3.3
phone display
Social Interest
• User interests: dynamically updated by users’ activities
• System execution
– 20 users at Penn State, 5-month period
– 11 categories, 86,914 transceived, 25, 872 read by users
A
Contact
C
B
Caching in Mobile Ad Hoc Networks
• In Battlefield, mobile devices of the soldiers form
a MANET.
• After a soldier obtains enemy information (e.g.,
battlefield map, enemy distribution) from the
commander (data center), it is very likely that
nearly soldiers also need the same information.
– Bandwidth and power can be saved if these data
accesses are served by the soldier with the cached data
instead of the data center which may be far away.
14
• CachePath: Cache the data path
• Suppose N1 has requested a data item from N11. N3 knows
that N1 has the data. Later if N2 requests for the data, it
forwards the request to N1 instead of N11.
• CacheData: Cache the data
• In the above example, N3 caches the data, and forwards the
data to N2 directly.
• Many technical issues not shown here
15
Social Contact
• System development
 802.15.4/ZigBee compliant
 10kB RAM, 250kbps data rate
 TinyOS 2.0
– Testbed: TelosB sensors
– Deployment: 1000+ sensors distributed to high school students
• Heterogeneity of centrality, community, high cluster coefficient
• Flu immunization
B
A
C
Other Projects
• Efficient Energy-Aware Web Browsing in Wireless
Networks
• In-Network Storage
• Security in Cellular Networks
– Z. Zhu, G. Cao, S. Zhu, S. Ranjan, A. Nucci, "A Social Network Based
Patching Scheme for Worm Containment in Cellular Networks," IEEE
INFOCOM, 2009.
– B. Zhao, C. Chi, W. Gao, S. Zhu, G. Cao, "A Chain Reaction DoS Attack
on 3G Networks: Analysis and Defenses," IEEE INFOCOM, 2009.
– Z. Zhu and G. Cao, "APPLAUS: A Privacy-Preserving Location Proof
Updating System for Location-based Services," IEEE Infocom, 2011.
• Data Dissemination in Vehicular Ad Hoc Networks
18