CoRaid abandoned engineering before SATA was a thing
The first CoRaid product was a physical ATA-to-Ethernet bridge that connected a cheap commodity disk to a 10Mbps ethernet port. The idea was that a good ethernet switch could be used to build a middling SAN that would cost at least 90% less than any solution involving Fiber Channel.
Back when disks had ribbon cables and companies had dedicated storage teams, our SAN Department was upset because the CoRaid solution was both excellent and disruptively inexpensive. At the time, the notion that ethernet technology could eclipse either SCSI performance or FC fabrics was novel.
After getting their AoE driver into the Linux kernel, near the time when libata was superseding libide, CoRaid pivoted into the software defined storage market. They probably did this to get an easier margin, but this space was soon overrun by OpenZFS appliances.
Afterwards, CoRaid lacked a differentiator because the AoE hardware was never updated for SATA, and because ZFS+iSCSI performance caught up through things like multi-core CPUs and offload engines on mid-range ethernet adapters.
CoRaid is a business study in how one good idea can carry a company for more than a decade through more than $100M in investment despite mediocre executive management.
Biting the hand that feeds IT
