ORLANDO (10/28/2003) - Maj. Gen. Dale Meyerrose, CIO and director of architecture and integration for the U.S. Northern Command at Peterson Air Force Base in Colorado, was Tuesday morning's opening speaker here at the Storage Networking World conference. After his presentation, Meyerrose spoke with Computerworld about an IT strategy that supports sharing information with civil authorities as part of the Northern Command's homeland security mission. Excerpts from the interview follow:
You said your main IT focus is on information sharing. What technologies do you need to standardize on in order to accomplish that?
It's not specific technology, but families of technology. They have to be commercially available, and they have to be deployable on a wide, wide range and scale. If you think about it, there are two-and-a-half million first responders in the United States. Conceivably, there may be instances in which potentially every first responder in the entire United States might need to send information to me. With that kind of environment, if it's too complex or based on DOD (Department of Defense) encryption standards (for example), all of those elements end up being barriers. So open-system-type technologies as a family of technologies are the ones that are most helpful to us.
You said in your presentation that you see Unix as the "Betamax of software." Does that mean it has no place in your IT environment?
It currently has a place, and in fact we do use Unix on several of the DOD systems. My thought behind Unix is the march of technology is what has made Unix less and less relevant. The Y2k rollover, in fact, killed lots of Unix. There are a finite number of Unix engineers and software writers in the world, and it is not big enough to support the information technology demands not only of our society, but of the world's economies in general. It costs a lot to train somebody; it takes many years to get them educated. It takes even more for them to get experience so that you can use them, when in fact a lot of software engineering and design is a lot less complicated and in my view will outpace Unix.
There are still lots of very important DOD Unix-based systems in existence. But I predict that five to 10 years from now, there may be almost no Unix-based systems, just because the intellectual investment it takes to maintain a cadre of Unix-trained, -educated and -experienced engineers is too difficult a problem.
How about Linux?
I think Linux has less investment baggage than Unix, and therefore I see that being a lot longer-enduring technology. It has an important role in DOD environments; it does not have that much of an important role in my environment yet. Most Linux systems are fairly complex -- lots of security parameters and things like that. We have to boil down to a common denominator that allows us to exchange information in a trusted information exchange environment with a very, very broad population.
Along the same lines, but on the storage front: Is Fiber Channel going to go the same way as Unix?
In candor, I don't know, and neither does anybody else. But if you're looking at something that's widely deployable, then it has to be simple in architecture. I believe that the more complex certain technologies are, either they're not mature enough to be deployed or they're going to have a finite life.
We see that cycle played out all the time. Something starts out as very complicated; a breakthrough happens and the same technology becomes simpler. And if it doesn't become simpler, like I maintain Unix has not become simpler, then it marks an endpoint into how long it's going to be useful.
Do you foresee the wide adoption of IP-based storage networks?
Again, I'm not a forecaster. But in the way I view the world, anything you deal with ends up being an IP address on a network. So the problem becomes network management.
Do you use your IP networks for both voice and data?
We are primarily using our IP networks for data -- very limited voice. We're being paced by the progress of DOD in using IP voice. We do use some, just like we use some wireless. So while we may be strong advocates for it, our pacing is (tied) to the Department of Defense.
You stressed in your presentation that procedures for information access need to be established within an organization in advance of technology. Can you elaborate on that?
In my experience, every time you come up with a problem you have to solve, invariably we initially solve it with procedures, and then somebody may come along later with technology and improve that solution. So when you come up with a problem, because you can't instantly deploy technology in most cases, you can instantly deploy procedures which allow you to do work-arounds until somebody comes up with a better solution. It may be a better procedural solution, or it may be a better technological solution.
What do vendors need to provide in order to enable object-based access to information?
There's a guide (of steps) that I suggest to them that they (follow). The first is to come up with a concept of operation. Oftentimes, we give folks IT tools, and then they have to discover how to use them. The second thing is a life cycle sustaining plan. That's what makes it real -- that's what changes it from PowerPoint to reality. That (determines) the decision about whether you're going to invest in that technology.
Most folks come and immediately want to sell me an IT tool because they think that I will automatically see the utility of the IT tool. The utility I may or may not see may not be what the intention of the tool is. In order for anybody to make an investment, you have to have an idea about how you're going to pay for it, how you're going to train to it, and how you're going to sustain it through the life cycle. Those are going to be the (criteria) that you place on your IT - not the capability of the IT.
Does Northern Command have a storage-area network?
I don't have a storage-area network. I use a portion of the Air Force Space Command's storage network. So it really is about Northern Command accessing information, not running infrastructure.
Have you standardized on any one vendor's technology to do that?
We have worked very hard to be system-agnostic and policy-dependent. To a great degree we've managed to do that, and we will continue to push for that.