Data deduplication is a specialized data compression tech-nique for eliminating coarse-grained redundant data, typ-ically to improve storage utilization. In this paper, we ex-plore another direction of using deduplication. That is, we use deduplication in the swapper. Before swapping out any page from memory to the swap area in disk, dedupli-cation checks that whether a page of the same contents
... [Show full abstract] has been written to the swap area. If so, we can avoid one I/O. We implement that idea in the linux kernel. Our experi-mental results indicate that using deduplication is able to reduce the overhead of swappers by orders of magnitude when there are many duplicate pages comparing to that of not using deduplication. However, we also notice that deduplication also incurs an overhead when few duplicate pages are present.