Evaluating Alexander

We now present a performance evaluation of Alexander. We focus primarily on the Linux ext2 variant, but also include some baseline measurements of the VFAT system. We wish to answer the following questions:

$\bullet$ Does Alexander work correctly?

$\bullet$ What time overheads are introduced?

$\bullet$ How effective is access-driven diffusion?

$\bullet$ How fast is live-block recovery?

$\bullet$ What overall benefits can we expect from D-GRAID?

$\bullet$ How complex is the implementation?