Context-Awareness Ephemeral Group Communication Services
- Part of Adaptive Middleware Services for Situation-Aware Communication in Ubiquitous Computing Environments

Energy-Efficient Multicasting Framework

Application Requirement

We are investigating the applications to facilitate group collaborations among users using handheld devices, such as conference meeting and classroom application. In these applications, in order to form a logical communities, members of a logical community need to interact. Such interactions will involve group communications of various types in order to identify and agree on a group situation. As opposed to groups in traditional distributed systems, these groups are dynamically formed and are ephemeral. Further, traditional approaches to group communication, such as multicasts using multicast trees may not be feasible in such environments due to short lived and opportunistic connections. As such, an efficient group collaboration service is needed to enable the establishment of device communities, which will use context as the collaboration agreement.

Research Challenges

  • Context-Awareness
  • Group Membership Management
  • Group View Maintenance
  • Energy-Efficiency
  • Bandwidth Efficiency
  • Light-weight
  • Distributed and Ad hoc Solutions.

Research Approaches

  • Cross Issues Design
  • Cross Layer Design
  • Adaptation
  • Application QoS Consideration


Faculty Advisor

Sandeep K. S. Gupta
    Associate Professor
    Department of Computer Science and Engineering
    Arizona State University

Graduate Students

   Bin Wang: PhD candidate and Group Leader (
   Guofeng Deng: PhD student (

Undergraduate Students

   Andy Lim ( supported by Research Experience for Undergraduate (REU) Program at NSF

Our Achievements

We proposed a group management model that represents the role of this kind of group and provides a mapping between user's social communities to device communities. Using the group management model, we proposed a Context-Awareness Ephemeral Group (CAEG) service and implemented it as a test bed (CAEG service is one of the two middleware services for an NSF-funded project In the test bed, devices utilize multiple on-board sensors to gather context information from surrounding environments. Contexts are made up from the sensor information and user's input (such as user's social context). The test bed is implemented on Dell and Compaq PDAs (each PDA obtains sensor data via USB port from a FPGA board). Based on the CAEG service, we also developed a Chat application. The CAEG service and Chat application have been used in Software Engineering Project I (CSE 461) and Software Engineering Project II (CSE 462) courses at Arizona State University in Fall 2003 and Spring 2004. The students are very excited to use CAEG service and Chat application. They provide very useful feedbacks to increase the CAEG service's capability and flexibility.


Bin Wang, John Bodily, and Sandeep K. S. Gupta, Persistent Social Groups in Ubiquitous Computing Environments Using Context-Aware Ephemeral Group Service, Proc. of Second IEEE International Conference on Pervasive Computing and Communications (PerCom), Orlando, FL, Mar. 2004, pp. 287-296. [ PDF | PPT]

Stephen S. Yau, Fariaz Karim, Yu Wang, Bin Wang, and Sandeep K. S. Gupta, Reconfigurable Context-Sensitive Middleware for Pervasive Computing, IEEE Pervasive Computing, joint special issue with IEEE Personal Communications on Context-Aware Pervasive Computing , 1(3), July-September 2002, IEEE Computer Society Press, Los Alamitos, USA, pp. 33-40. [ PDF]

Stephen S. Yau, Sandeep K. S. Gupta, Fariaz Karim, Sheika I. Ahamed, Yu Wang and Bin Wang, Smart Classroom: Enhancing Collaborative Learning Using Pervasive Computing Technology, Proc. of 6th WFEO World Congress on Engineering Education & Second ASEE International Colloquium on Engineering Education (ASEE), June 2003, Nashville, Tennessee. [ PDF | PPT]

This work has also been used in the following courses:



Ambient Intelligence

Massachusetts Institute of Technology  

Mobile Computing Systems & Applications Fall 2004

Carnegie Mellon University  

Distributed Systems Spring 2004

University of Colorado at Boulder

Parallel and Distributed Computing Spring 2003 

Rutgers University

Wireless Networks and Mobile Systems Spring 2003 and 2004

Viginia Polytechnic Institute and State University

Foundations for Future Mobile Computing 

University of Helsinki

Mobile and Ubiquitous Multimedia Fall 2002

University of Oulu

Pervasive Computing Fall 2004

National Tsing Hua University

Pervasive Computing Fall 2004

University of Ulster

Ubiquitous Computing Fall 2002

Arizona State University

Wireless Networks and Mobile Systems Lecture March 2005



Test bed of Hardware Setup

PDA connects FPGA (integrates with light, noise and motion sensors) through USB cable

Application Scenarios

1. Group Text Chat in Classroom (Completed)

Instructor assigns students into groups for a Term project. The students may get together regularly, e.g. every week, and collaborate through the PDAs. PDAs detect the group meeting event and form a device community to support the collaboration. When the meeting is finished, the device community will be dispersed and group meeting report is available at every student's PDA.

2. Radio Station-based Chat in Traffic Jam (Finished Prototype)

John encounters traffic jam at downtown New York. He turns on the radio to FM 101.5. His car discovers this event and helps John to form a chat group which includes nearby people listening to the same radio station. When John tunes the radio to another radio station, the car will help him to join another chat group on the new radio station.

Following is our implemented prototype on PDAs.

User adds a new "music" interest.

Found existing groups nearby and join the group. Then start chatting with each other.

3. Music Sharing (Ongoing Work)

A group of friends come together and have a song they want everyone to listen to. The owner of the song will distribute and play the song for all the friends. The owner may fast forward and the application will fast forward on the other devices too, so that if there was a good segment that everyone should hear, the owner could fast forward it and everyone could listen. After playing the song, all the copies on the devices will be destroyed except the owner's one. To avoid annoying other nearby people, the application will mute the speaker when song is playing.