The Fundamentals for IP Storage
In many ways, IP storage systems work the same as any storage network — they use the serial small computer system interface (SCSI) protocol to communicate with their disk drives, they require redundancy to ensure their data are secure and they need an experienced storage area network (SAN) administrator to keep everything running smoothly.
Yet, in some fundamental facets, IP storage systems are very different from their Fibre Channel counterparts. Marc Farley, author of “Building Storage Networks,” said the main feature that distinguishes these two systems is the way their addressing components work.
“The addressing model of Fibre Channel is tied to the physical devices that you have,” he said. “An IP address doesn’t have to be associated with a piece of hardware. As it turns out, because of that, it’s got some great flexibility advantages — you can manage the network at a higher level, and you can assign IP addresses and manage them in a way that you just can’t do in an addressing scheme that’s tied to hardware.”
An IP storage system gives an organization the room to grow because it makes it easy to move and enhance its storage components, Farley said.
For example, if a small Internet company begins to attract investors and wants to upgrade its equipment, it’s much easier for an administrator using an IP storage system to quickly move the business’ information to a new piece of hardware because all he or she must do is transfer the IP address.
With other types of storage systems, upgrades can be more complicated, Farley said.
“With IP, in a matter of seconds, the new machine is working, and everybody in the world can get to it through this IP address that you moved from one machine to another,” he said. “If you had to do that with an Ethernet MAC (media access control) address or a Fibre Channel address, you would either have to let everybody know that you got a new address, or you’d have to do some kind of funky other things to fake everybody out.”
Another benefit of IP storage is that it can use many different types of networks to communicate storage information, Farley said.
“IP is network independent, and it can run on Ethernet, it can run on Fibre Channel and it can run on ATM (asynchronous transfer mode) networks, but now on top of the IP layer, you also have storage information,” he said. “So, now two systems that have the ability to communicate by IP, say, over Ethernet, can now also carry storage information.”
Outside of the addressing components that enhance the flexibility and manageability of IP storage systems, the fundamentals of IP storage are similar to those of any other storage system. Redundancy, scalability, virtualization and management are crucial to keeping data safe and accessible at all times. And, of course, data security is the prime concern for all SAN administrators, no matter what type of storage network they use.
“The big difference between a storage network and any other type of network is that problems in a storage network can ruin data and end up getting you fired, whereas problems in most other networks just look like blips, and you can connect again later — not that big a deal,” Farley said. “In a storage network, a blip can be a catastrophe, so you have to be very careful.”