Bloom Filter Based Privacy Preserving Deduplication System

Deduplication is a data reduction technique which eliminates uploading and storing redundant data. Therefore, it is widely adopted in cloud storage services to reduce communication and storage overhead. However, deduplication can be used as a side channel to learn the existence of a particular data in cloud storage thereby raising significant privacy issues. The existing solutions delay deduplication process to hide information regarding the presence of data. These solutions increase communication overhead as the client needs to send data even if it is present on storage. In this paper, we present a novel privacy preserving deduplication approach using bloom filter. In our approach, bloom filter (containing blocks of data) is used as a deduplication identity. When a client sends an upload request, the server responds by sending a genuine bloom filter (if data exists) along with some dummy filters. Now, that client who genuinely owns the data, can learn the existence information by computing bloom filter of the data. Further, client does not need to send the data if it exists on storage. Security analysis proves that our approach provides privacy to data at deduplication system. We implement the approach and demonstrate that communication overhead is significantly less in our approach than the existing approaches.