Building a storage centric internet infrastructure ...

P2P storage


Vision: A storage centric internet architecture (ASCII)

The primary functionality of the internet has been to provide IP based routing, which has been complemented in the last decade by overlay based routing services based on address independent arbitrary logical identifier spaces. On top of this routing-centric infrastructure, the World-Wide Web has been built as a document-centric network, to navigate and share documents between Internet users. Cloud is an emerging paradigm - where infrastructural resources (such as computation and storage), as well as other resources (such as software) can all be availed as services.

Storage as a basic service will play an important role in spurring new applications. Instead of paying cloud service providers to avail such resources, an alternative can be that end users barter thier own resources in a peer-to-peer manner to enjoy the same functionalities. Such a paradigm allows easier participation of end-users who are unable or unwilling to pay to still avail a feature rich storage service.  The ASCII vision is to facilitate such a "common man's storage cloud" - a flexible mix of provisioned as well as peer-to-peer storage infrastructure, which will enable end-users to deploy and participate in a new wave of applications which rely upon such a storage service.


P2P storage systems

Peer-to-peer (P2P) storage is a paradigm to leverage the combined storage capacity of a network of storage devices (peers) contributed typically by autonomous end-users as a common pool of storage space to store and share content, and is designed to provide persistence and/or availability of the stored content despite unreliability of the individual autonomous peers in a typically decentralized environment. Besides the basic reliability guarantees of persistence and/or availability, a feature rich P2P storage system may need to support other functional requirements depending on application needs. Some such functional requirements include:
  • Transactional access, concurrency and consistency issues
  • Security features such as:
    • Fine grained access control
    • Content validation (e.g., against pollution)
CrowdStore is our ongoing project to realize a security features rich persistent p2p storage centric internet infrastructure.

Publications


Persistence issues

Journals and Conferences

Internet-scale storage systems under churn - A study of the steady state using Markov models,
Anwitaman Datta, Karl Aberer
P2P 2006, The Sixth IEEE International Conference on Peer-to-Peer Computing, UK.

Stochastic Analysis of the Interplay between Object Maintenance and Churn,
Di Wu, Ye Tian, Kam-Wing Ng, Anwitaman Datta
The Journal of Computer Communication [Elsevier] 31(2): 220-239 (2008).

Redundancy maintenance and garbage collection strategies in peer-to-peer storage systems,
Xin Liu, Anwitaman Datta
SSS 2009, The 11th International Symposium on Stabilization, Safety, and Security of Distributed Systems, Lyon, France.


Replica Placement in P2P Storage: Complexity and Game Theoretic Analyses,
Krzysztof Rzadca, Anwitaman Datta, Sonja Buchegger
ICDCS 2010, The 30th International Conference on Distributed Computing Systems, Genoa, Italy.

Overview papers

Peer-to-Peer Storage Systems,
Anwitaman Datta
Entry in Springer's Encyclopedia of Database Systems (reference work series), September 2009

Maintaining redundancy in peer-to-peer storage systems,
Anwitaman Datta, Di Wu, Liu Xin, Adam Wierzbicki
Book chapter in the Handbook of Research on P2P and Grid Systems for Service-Oriented Computing: Models, Methodologies and Applications, IGI Global publisher.


Tutorial/Survey

Peer-to-Peer Storage Systems: Crowdsourcing the storage cloudAnwitaman Datta
Tutorial at ICDCN 2010.

Security issues

Conferences

Enabling Secure Secret Sharing in Distributed Online Social Networks
Vu Le Hung, Sonja Buchegger, Karl Aberer, Anwitaman Datta
ACSAC 2009, Annual Computer Security Applications Conference, Hawaii, USA.

Consistency and updates

Journals and Conferences

Updates in Highly Unreliable, Replicated Peer-to-Peer Systems,
Anwitaman Datta, Manfred Hauswirth, Karl Aberer
ICDCS 2003, The 23rd International Conference on Distributed Computing Systems, USA.

CMV: File Consistency Maintenance Through Virtual Servers in Peer-to-Peer Systems,
Zhijun Wang, Anwitaman Datta, Sajal K. Das and Mohan Kumar
Journal of Parallel and Distributed Computing [Elsevier] 69(4): 360-372 (2009)

Applications

Distributed online social networks

PeerSoN: P2P Social Networking - Early Experiences and Insights,
Sonja Buchegger, Doris Schiöberg, Le Hung Vu, Anwitaman Datta
SocialNets 2009, The 2nd ACM Workshop on Social Network Systems at EuroSys 2009.

Software


Simulator

P2P3S

 P2P storage systems simulator

CrowdStore

Work in progess