It's very likely that they are using Intel SSDs (likely 320/330, maybe 510+) for a variety of reasons:
Intel has a good drive reliability history, and is very enterprise friendly in bulk purchasing. Intel has had excellent firmware for their drives, which at the datacenter scale is valuable--people who have dealt with raid controller firmware(including Amazon) know all about this.
Traditionally, Amazon has not used SAS drives in EC2, opting for lower cost SATA drives. It's also unlikely that Amazon is using small numbers of high capacity (>=500GB) drives because they still aren't perfectly price effective; price per gb is ok, but replacing a failed drive is more costly.
Also keep in mind that to get to today, Amazon has been rolling these drives in huge numbers out through two enormous data centers, so it's unlikely that Amazon has picked a brand new drive (say the latest OCZ Vertex 4).
There are other factors that Amazon has bumped into while testing drives, but they remain unreported and internal.
Intel has a good drive reliability history, and is very enterprise friendly in bulk purchasing. Intel has had excellent firmware for their drives, which at the datacenter scale is valuable--people who have dealt with raid controller firmware(including Amazon) know all about this.
Traditionally, Amazon has not used SAS drives in EC2, opting for lower cost SATA drives. It's also unlikely that Amazon is using small numbers of high capacity (>=500GB) drives because they still aren't perfectly price effective; price per gb is ok, but replacing a failed drive is more costly.
Also keep in mind that to get to today, Amazon has been rolling these drives in huge numbers out through two enormous data centers, so it's unlikely that Amazon has picked a brand new drive (say the latest OCZ Vertex 4).
There are other factors that Amazon has bumped into while testing drives, but they remain unreported and internal.