Faculty of Technical Sciences

In this paper, we introduce HEKATE a tool for gauging the performance of data deduplication. In particular, HEKATE allows to test classic, chunk-based deduplication as well as the emerging concept of generalized deduplication, which leverages different data transformations prior to deduplication. We show that HEKATE can be used to evaluate performance with different system configurations and deduplication schemes. In fact, we use six real-world data sets to showcase HEKATE ’s
ability to test performance over a wide range of conditions. We also use this evaluation to showcase how HEKATE can provide insights into data deduplication performance in real systems and optimal configurations given specific data characteristics of an expected workload. HEKATE can also be used to determine the suitability of various deduplication approaches for each data set.