I talked to two cloud startups today, but to my mind only one of them should get to use the term "cloud" in how it positions itself.First is ParaScale, which sells software for the creation of private clouds. I think I'm getting into a religious debate here, but I don't think the notion of a private cloud makes any sense. In my mind, a key component of the definition of a cloud is that you, the enterprise, don't have to host and run the infrastructure. All that stuff is supposed to be out of your hands.

This is not to say that ParaScale doesn't sound like an interesting company. I think it sounds great. But it's really a storage virtualization company, not a cloud storage provider.

Here's how it works. You take the ParaScale software and load it onto a bunch of servers running Red Hat Linux. The ParaScale software makes all these servers look like a single IP address and single storage address to whatever systems you connect to it. A control server keeps track of all the bothersome namespace issues, so IT can simply chuck new servers into the array whenever they need more capacity, without having to reconfigure any applications or computers that need to connect to the array to store and retrieve files.

This is essentially the same approach of file virtualization vendors such as F5 Acopia and EMC Rainfinity. But instead of fronting expensive NAS boxes, ParaScale is designed for low-cost commodity servers. ParaScale also doesn't require any specialized, high-end hardware to sit in front of the array like you need with Acopia.

Fuzzy Definitions

Great. I love the idea, and I plan to follow ParaScale closely because it sounds like a compelling offering. But if you're building all this architecture inside your own data center, and running it yourself, it's not a cloud solution.

In ParaScale's defense, the company is also pursuing the service provider market, which could use ParaScale's software to build an infrastructure to resell cloud services, but that's about as close as ParaScale gets to being a cloud storage company.

The other company I talked to today is CTERA. Founded in May, it had its official launch at CES this week. CTERA aims at the SOHO and SMB market with a neat idea. It provides a small storage device, called a Cloud Plug, that SOHOs and SMBs can use as a common file share. This device also connects to an online backup service so that all the files on the device are copied to a remote data center.

The Cloud Plug is a clean, simple, and sensible approach to storage. But CTERA is not a cloud vendor. Why? The company isn't building its own data centers and doesn't actually offer online storage. Instead, its business plan is to sell the Cloud Plug to service providers, which will in turn sell the device and the online storage to customers.

To CTERA's credit, I think the company is being more careful about how it positions itself. Yes, the word "cloud" shows up all over its Web site and press releases, but that's OK because its product is actually designed to connect to a third-party storage service that provisions and operates all the back-end infrastructure so you don't have to.

When Is A Cloud Not A Cloud?

I can see the temptation for startups to attach themselves to hot terminology, and I'm sure ParaScale will fight to defend the notion of the private cloud. And maybe it could convince me of the merits of its arguments. But for now, as this market grows, it's going to get increasingly, um, cloudy. We're all better off when terms are as clear as possible.