So reading between the lines, the killer features of their software are:
1) Using LZO compression to increase cache storage efficiency (since on a current generation of CPUs LZO can decompress faster than even most SSDs can provide the data)
2) Block level deduplication within the cache
Good and interesting stuff. I'd be curious to see how this compares to ZFS which can also do both of these implicitly if you enable deduplication and compression on the volume, and it can use SSDs for caching (L2ARC).