Get some code here and start a Hekaton
Memory optimalization code named Hekaton:
So we have this idea that our data belongs on disks and the data we need will be loaded into the memory of your machine for manipulation. In Hekaton transactions never get rolled back. when you change data, you create a new version in memory. If the transaction fails, the changes are never written to the disk. When a transaction does commit, it goes into the transaction log and a background process copies it back to the set of sequential files, the Hekaton data. These files are read at database start-up, when they get copied into memory. the downsite is, it has to fit in your RAM, if not, the new rows can’t be added, you can’t change it to the old write to disk mode. So, your old database probably will not work with hackaton, but Microsoft is working on a solution for this to optimize al your data.
to be continued…