In this podcast, the Radio Free HPC team looks at the coming wave of declustered RAID solutions designed to address the problem of ever-increasing RAID rebuild times. With 5 Terabyte and larger drives in the wings, RAID cluster rebuild times are becoming impractical.
- Dan is impressed with what IBM is doing in this area with GPFS.
- Henry thinks these larger drives harken the death of RAID-5 and RAID-6.
- Rich has been keeping a close watch on Xyratex and their ClusterStor GridRaid implementation.
Read the original paper on Declustered RAID by Garth Gibson and Mark Holland * Download the MP3 * Subscribe on iTunes * RSS Feed