NJIT logo
Mobius: A Multi-Tier Socially-Aware Network Infrastructure
  Overview
  Research Directions
  People
  Funding
  Publications
   
 
Overview
This research aims to explore the benefits of embedding social knowledge in the network protocols and services that support mobile social computing applications. To this end, we will investigate, design, and prototype Mobius, a self-organizing, self-adaptive, community-oriented, multi-tier network infrastructure for mobile social computing. Similar to the Mobius band that has one surface despite the appearance of two, this infrastructure consists of two tiers that converge into one network: (1) a Mobile (wireless) human-centric tier, which runs mobile applications and collects geo-social context information, and (2) a Peer-to-Peer (wired) system tier, which runs services in support of mobile applications and adapts to the geo-social context to enable energy-efficient, scalable, secure, and reliable mobile applications. Similar to traditional Peer-to-Peer (P2P) networks, users will contribute their resources, both wireless mobile devices and wired PCs, to create this infrastructure. Unlike existing mobile service provisioning solutions, Mobius is decentralized, adaptive, and flexible. As a necessary departure from the dependence of mobile applications on cellular network operators, it allows mobile applications on the phones (i.e., in the Mobile tier) to interact with user-deployed mobile services in the P2P tier. This independence is achieved using community-supported services running on a collection of user-owned resources in a P2P manner. Our architecture does not preclude existing or new services offered by network providers or third parties.
Project web page at USF

Research Directions

::: Socially-Aware Peer-to-Peer Overlay
We plan to identify, design, and prototype a set of socially-aware P2P adaptation strategies. These strategies provide resilience to node churn, and improved performance and resource management. Examples of such strategies include: (i) mapping community-specific content, services, and state to peers in such a way as to optimize the resource consumption and the user-perceived response time, (ii) deciding when and where to instantiate a service to better serve the community of users accessing it, and (iii) granting or denying offloading requests and deciding where to run the offloaded application based on users' social ties.

::: Socially-Aware Data Management
We conjecture that social knowledge is beneficial for guiding the design of a data management system that supports mobile social applications. We assert that including social knowledge in the supporting infrastructure for mobile social applications has the following benefits:

  • Sensitivity to the geo-social context: The ubiquity of mobile devices in our lives leads to undiscerning competition for our attention. We believe that by embedding social knowledge in system design, mobile devices can self-configure to reduce inopportune requirements for user attention. Similarly, data delivery can also be more sensitive to the current social context of the user, such that, for example, fun videos from YouTube and VoIP calls are not delivered during professional meetings.
  • Incentives for resource sharing and participation in the p2p infrastructure, as socially connected users (such as family members or friends) are expected to share better quality resources with each other.
  • Increased trust and privacy can be achieved by avoiding centralized storage of social knowledge (big brother fear) and instead placing social information or personal data onto trusted peers.
  • Geo-socially-aware load transfer from mobile devices to peer-to-peer resources: delivering data closer to where the mobile device is likely to reduce data access latency and battery consumption. The P2P layer provides improved data persistency compared to the highly intermittent mobile device participation.
  • Enabling novel approaches for filtering unwanted content by using user’s social community feedback to build a personalized social firewall.
  • Improved resource management in the network (mobile and P2P) by replicating data in a manner that is aware of mobile user co-location, and cooperation capabilities via ad-hoc communication.

::: Geo-Social Data Collection
Inferring social information from the use of technology is possible due to two main factors. First, the highly popular online social networking applications, such as MySpace or Facebook, can provide rich information about the social fabric of the Internet users. Second, the mobile devices equipped with communication and localization capabilities can help refine reported social relationships. We are exploring ways to infer the social context of a user by combining these two sources of information.

::: Application Offloading
Battery will continue to be a costly resource in mobile applications. At the same time, more applications are more and more demanding in terms of computation and communication, thus making more and more demands on battery power. We investigate ways to offload applications from mobile devices onto the P2P infrastructure in such a way that saves battery and improves performance. Application placement is again a decision that cab benefit from social knowledge for exploiting inherent incentives and trust for resource sharing between users.

::: Privacy and Security Policy Specification and Enforcement
An important concern in this project is privacy protection. While knowledge about the social behavior of an individual or a community can fundamentally enhance the impact of mobile social applications in our lives, it can also be used in ways that violate privacy or even safety. An important component of this project is thus the design and implementation of a framework for defining and enforcing individual, community-specific, or global privacy policies. These policies can specify constraints in: (i) accessing user content, (ii) accessing social data, (iii) service instantiation and execution, and (iv) application or component offloading.


 
People


Faculty:

Students:

  • Juan (Susan) Pan (NJIT)
  • Manoop Talasila (NJIT)
  • Mohammad Khan (NJIT)
  • Hillol Debnath (NJIT)
  • Nicolas Kourtellis (USF)
  • Jeremy Blackburn (USF)

Past Student Members:

  • Daniel Boston (NJIT)
  • Steve Mardenfeld (NJIT)
  • Bidisa Rai (NJIT)
  • Gezhi Zhong (NJIT)
  • Joshua Finnis (USF)
  • Paul Anderson (USF)
  • Mayur Palankar (USF)
  • Lydia Prieto (USF)

Collaborators:

 
Funding

» Mobius: A Multi-Tier Socially-Aware Network Infrastructure. NSF CNS-0831753 and CNS-0831785, 2008-2011. Collaborative research between University of South Florida and New Jersey Institute of Technology.

 
Publications with NJIT Participation

» Enabling Social Applications via Decentralized Social Data Management
Nicolas Kourtellis, Jeremy Blackburn, Cristian Borcea, and Adriana Iamnitchi
ACM Transactions on Internet Technology, Special Issue on Foundation of Social Computing,Vol. 15 , No. 1, February, 2015.

» Collaborative Bluetooth-based Location Authentication on Smart Phones
Manoop Talasila, Reza Curtmola, and Cristian Borcea
Elsevier Pervasive and Mobile Computing Journal, Vol. 17, February, 2015

» Leveraging Bluetooth Co-location Traces in Group Discovery Algorithms
Daniel Boston, Steve Mardenfeld, Juan (Susan) Pan, Quentin Jones, Adriana Iamnitchi, and Cristian Borcea
Elsevier Pervasive and Mobile Computing Journal, Vol. 11, April 2014, Special Section on Mobile Social Networks

» TIE: Temporal Interaction Explorer for Co-presence Communities
Daniel Boston and Cristian Borcea
Proceedings of the International Conference on Social Computing and its Applications (SCA 2011), December 2011.

» Analysis of Fusing Online and Co-presence Social Networks
Susan Juan Pan, Daniel Boston, and Cristian Borcea
Proceedings of the 2nd IEEE Workshop on Pervasive Collaboration and Social Networking (PerCol 2011), March, 2011.

» Prometheus: User-Controlled P2P Social Data Management for Socially-Aware Applications
Nicolas Kourtellis, Joshua Finnis, Paul Anderson, Jeremy Blackburn, Cristian Borcea and Adriana Iamnitchi
Proceedings of the 11th ACM/IFIP/USENIX International Middleware Conference (Middleware 2010). December 2010.

» LINK: Location verification through Immediate Neighbors Knowledge
Manoop Talasila, Reza Curtmola, and Cristian Borcea
Proceedings of the 7th International ICST Conference on Mobile and Ubiquitous Systems (MobiQuitous 2010). December 2010.

» GDC: Group Discovery using Co-location Traces
Steve Mardenfeld, Daniel Boston, Susan Juan Pan, Quentin Jones, Adriana Iamntichi, and Cristian Borcea
Proceedings of the 2nd IEEE Symposium on Social Computing Applications (SCA-10) held in conjunction with IEEE SocialCom 2010. August 2010.

» MobiSoC: A Middleware for Mobile Social Computing Applications
Ankur Gupta, Achir Kalra, Daniel Boston, and Cristian Borcea
ACM/Springer Mobile Networks and Applications (MONET) Journal, Vol 14, No. 1, 2009.

» MobiDew: Socially-Aware Data Management for Mobile Users
Nicolas Kourtellis, Joshua Finnis, Adriana Iamnitchi, and Cristian Borcea.
Poster presented at the 10th International Workshop on Mobile Computing Systems and Applications (HotMobile'09)

» P2P Systems Meet Mobile Computing: A Community-Oriented Software Infrastructure for Mobile Social Applications
Cristian Borcea and Adriana Iamnitchi.
Proceedings of the Workshop on Decentralized Self Management for Grids, P2P, and User Communities (SELFMAN 2008) held in Conjuntion with the Second IEEE International Conference on Self-Adaptive and Self-Organizing Systems (SASO 2008), October 2008.

 
 

2012 - UbiNetS Lab (New Jersey Institute of Technology) and Distributed Systems Group (University of South Florida)