Non-volatile storage support for data deduplication (Invited talk)

Data deduplication has the salient features of decreasing data redundancy and improving system performance. Existing deduplication systems fail to efficiently offer cost-effective storage services due to the potential performance bottleneck of indexing data redundancy. In this invited talk, we consider non-volatile storage devices to provide large capacity and support fast indexing performance. In the meantime, data redundancy also causes potential energy inefficiency in deduplication operations and network transmission. In the system implementations, we leverage software-defined schemes to support in-network deduplication with the aid of non-volatile storage devices.