276
Data Integrity for Collaborative Applications over Hosted Services
Ertem Esiner and Anwitaman Datta
Nanyang Technological University, Nanyang Technological University

In this work we focus on integrity and consistency of data accessed and manipulated by multiple collaborating users, and stored in an (untrusted) hosted service. This is a problem, aspects of which have been studied in isolation in hitherto distinct communities. Consistency is one of the cardinal problems of distributed computing. Integrity of hosted data has been studied over the last decade, and numerous techniques for proof of data possession and/or retrievability have been explored. The latter line of work however have often assumed static data, and techniques to handle dynamic or versioned data have only very recently been proposed. Yet, even the existing solutions that handle mutable content do so under the assumption that only a single data owner (using a single client) manipulate and verify said data. This is a serious limitation in terms of the variety of applications that can benefit from such mechanisms for proof of data possession. The novelty, and primary contribution of this work is in filling this gap. Specifically, we extend the existing ideas of proof of possession of dynamic data, in order to support multiple users who may collaborate in real time or asynchronously. In contrast (and addition) to the challenge of an untrusted storage server that existing techniques for proof of data possession need to overcome, we had to, simultaneously account for data integrity violations that may be incurred due to all the usual challenges of maintaining consistency of collaborative data (even if the storage server was trusted).