Facebook updates Flashcache data storage project

Demonstrating its love for (and dependence upon) sharing, Facebook is once again sheding more light about the inner workings of its data infrastructure.

Today's news comes in the form of an update about its ongoing data storage project, dubbed Flashcache, with the latest improvements focused on three particular areas: read-write distribution, cache eviction, and write efficiency.

While most of the nitty-gritty details require some engineering experience to comprehend, the bottom line is that these performance improvements have improved the average hit rate from 60 percent to 80 percent, and the disk operation rate has been slashed nearly in half.

That essentially means faster operations on servers (and then the front-end portals on desktop and mobile screens) in the long run.

In providing a history about the decisions made over hard disk versus flash storage solutions, the Facebook engineering team explained in a blog post on Wednesday that there is still a long way to go given current hardware industry challenges:

Also, while the cost per GB for flash is coming down, it’s still not where it needs to be, and this introduces more complexities into capacity planning. SSDs have limited write cycles, so we have to make sure that we’re not writing too much. Every cache miss results in a flash data write, so having flash devices that are too small may be problematic. In this case it’s nice to have hard disks that aren’t too fast, as any cache hierarchy has to depend on steep performance differences between multiple levels.

Engineers slipped in that they have also "already spent some time restructuring metadata structures to allow for more efficient data access," translating to improved searching capabilities -- and possibly even more results hidden deep in the archives when deemed relevant.