Towards Secure, Dynamic, and Practical Auditing for Decentralized Storage Network

Decentralized storage (DS) is emerging as an appealing alternative to commercial cloud storage. It envisions open storage marketplaces where anyone can join and monetize unused disk space to store and serve data for others. As commonly acknowledged by all DS proposals today, such openness mandates effective auditing mechanisms to assure owners of their remotely stored data integrity. Ideally, by employing certain auditing mechanisms to routinely check the data integrity, the DS can remunerate or fine individual storage providers accordingly, enforcing correct functioning and health for the ecosystem. For auditing transparency, a standard architecture here is on-chain auditing, where a general-purpose blockchain serves as an unbiased auditor to publicly verify storage proofs submitted by storage providers and fairly resolve potential disputes between parties.   Despite the promises, the stringent requirements of this blockchain-reliant architecture make realizing on-chain auditing paradigm a daunting task: 1) The immutability and public accessibility of blockchain pose unique security threats towards the on-chain auditing trails; 2) Scalable on-chain auditing designs require the storage proofs to be as small and easy to verify as possible; 3) The ubiquitous data dynamics in real-world storage systems calls for efficient auditing of dynamic data; 4) Implementing a fully functional on-chain auditing system demands non-trivial systemization efforts. In this project, we plan to push forward the frontier of this under-explored area with secure, dynamic, and practical on-chain auditing for decentralized storage, with three research tasks: 1) Develop a secure and efficient on-chain storage auditing framework: a) enhancing security guarantees against unforeseen threats that exploit the on-chain auditing trails, such as unsolicited data recovery and storage freeriding; b) achieving new auditing schemes that are more affordable with much improved on-chain efficiency. 2) Devise a scalable on-chain auditing protocol to support dynamic data: a) designing a near-optimal indexing translation data structure, to efficiently turn an auditing scheme into a dynamic one with minimal changes; b) improving indexing data structure with performance trade-off, potentially making the design more scalable. 3) Establish a full-fledged working prototype for practical deployment: a) enabling a secure and trustworthy middleware service to alleviate the workload for data owners and storage providers respectively; b) building a data abstraction layer to enable the auditing support for diversified data applications. Our project will greatly advance the field of decentralized storage ecosystem, contributing new insights to the trending on-chain auditing paradigm. Our research outputs will also serve as solid grounds for future decentralized data applications.