I must be an optimist, because when I hear storage vendors promising – yet failing to deliver – interoperability year after year, I simply shrug and say, “Better luck next year.” I can afford to be so detached because I don’t have to make all the pieces of the storage puzzle work. But I feel for those who do.
Fortunately, the state of storage interoperability is indeed changing for the better. IP networking vendors such as Cisco Systems Inc., along with traditional Fibre Channel and storage vendors such as Brocade Communications Systems Inc., EMC Corp., Emulex Corp., and McData Corp., are finally figuring out what LAN vendors learned a decade ago: Customers want interoperability that really works, instead of just acronyms on spec sheets.
Nevertheless, storage-networking players have a long row to hoe before storage networks become as plug-and-play as we expect our LANs to be. LAN interoperability didn’t happen overnight, as anyone who broke into networking between 1980 and 1995 will tell you – at length if they’re allowed. Thankfully, many of the lessons learned then can be applied to the building of storage networks.
The most important lesson for me as a network designer during those years revolved around the KISS (keep it simple, stupid) principle: Find an equipment vendor or two you can live with and then stick with them, because few companies have time to be interoperability guinea pigs although most end up doing some experimental work whether they realize it or not. The road to sanity in a time without standards is best found with as few people in the car as possible.
The big consolation for everyone struggling with storage networking is that it will get easier soon. I expect that within a few years, we’ll be wondering what all the fuss was about. Still, it’s premature for vendors to make claims of interoperability today, given the kinks remaining in IP implementations of Fibre Channel and SCSI.