Memory compression allows for greater overcommit than other memory management techniques by compressing the memory page and storing it as a cache in a small portion of memory. When the page is needed, it's retrieved from the cache, decompressed and placed back into working memory. The amount of space savings can vary, depending on the amount of redundant data. A potential downside of memory compression is the processing overhead on the server, but in virtualized environments, memory is typically harder to come by than processor cycles and compression performance is better than page swapping to disk. Because the total space savings and overhead aren't set and ultimately depend on the data, it's important to experiment in a test lab before deploying in production.