One of the things I love about the storage industry is the continual emergence of innovative technology. New features in our market can go from inception to implementation in as little as 18 months. Seeing this continuous hyper-evolution of storage features, it has occurred to me that the storage industry doesn’t really innovate at all, it mimics existing innovations. No, not mimicking the technology itself, but the evolution of the technology. Let me explain.

Remember back in 1997 when 10-megabit Ethernet was the new kid on the block? The ISP I worked at finally dumped our antiquated hubs and bought a shiny new USRobotics 10-megabit switch for our core network. Soon after that came 100-megabit hubs, then 100-megabit switches, then core switches with backplanes, so on and so forth.

Then something special happened. Switches weren’t just switches anymore. Now they had integrated routers, packet filters, intrusion detection, cache engines and load balancers. Switches were becoming intelligent. These days, the network switch does a lot more than simply network things together.

Compare that to the evolution of Fibre Channel switches over the last 8 years. See any similarities?

Along those same lines, aren’t Zones basically VLANs? Aren’t WWNs just MAC addresses? FC doesn’t seem to have a lot of the problems that TCP/IP and Ethernet have. For one, FC is a combined protocol, but it also eliminates the need to assign IP addresses, set up DHCP servers, tweak MTU sizes, or worry about dropped packets.

For management protocols, look at the how similar SNMP is to SMI/S. For certifications, there are CISSP and SNIA.

Another great example is De-duplication. I remember first seeing this technology about 15 years ago when it was called compression.

On a side note, I always secretly smile when I hear the term “De-dupe.” De-dupe is a great marketing term. I doubt it would be so popular if it were called compression2.

The only problem with mimicking another industry is, there doesn’t tend to be a whole lot of natural innovation. For example, there are some features that have long since been desired but still aren’t quite there. Thin Provisioning comes to mind – what took so long?

Another is ILM.

ILM is a great concept on paper, but not really practical. Currently it is just too difficult to move data from expensive disk to cheap disk and vice versa. And now that Solid State is becoming affordable, customers are going to want to move data around their environment easily and online, without having to engage project managers and data migration appliances. This is changing of course, there are some great LUN virtualization products maturing, and I suspect in 12 months I will be singing a different tune on this topic.

Even though there is an intelligent evolution happening in the storage market, there is still a fair amount of stumbling occurring. For example, this constant churn about iSCSI vs. FCOE vs. FC vs. Infiniband is getting a little old. I see this topic in industry ‘expert’ blogs 100 times more than I hear about it from actual customers with money. I don’t think people realize that until there is something deployable in a customer environment, all this discussion is just academic. Let the companies innovate, the customers purchase, and the market decide. Worry not Chicken Little, FC is not going away anytime soon.

I am happy that my industry is trying to evolve in a more efficient way than other industries. Fast emerging technologies keep us early-implementers employed. It will be interesting to see what is innovated, er, mimicked next.