SNW - Weather Channel"s VP on horizons

25.10.2005 von Lucas Mearian

Dan Agronow, vice president of technology at The Weather Channel Interactive Inc. in Atlanta, has been focused on rolling out more open-source software and commodity hardware as a cost-cutting effort over the past few years. At the same time, he"s been working on a more robust disaster recovery plan as well as employee retention. In an interview with Computerworld at Storage Networking World Tuesday, Agronow discussed how he"s addressing key industry challenges.

Have blade servers created any issues? I"m more concerned about heat. We run a lot of 1U pizza boxes. It"s great to have faster and more powerful boxes. But if I have to spread them out because of the heat, then I have an issue with how many servers I can stick in my racks. It"s great to see the competition between Intel and AMD. I think ultimately that"s going to help the end users. I most dislike all the lawsuits going around. Cisco, IBM, AMD, Intel. It just seems to delay projects going forward.

What"s your biggest concern these days? Losing skilled staff. The past couple of years, the industry hasn"t been that good. Now things are starting to pick up. We"ve had almost zero turnover the past couple of years. My fear is I"ll lose some of these talented individuals I"ve built up over the past few years. In the past, it was very easy to fill positions. Most times, you got a lot of resumes and high-quality candidates. What we"re seeing lately is the number of resumes are down and the quality of the candidates are down. And we"re getting outbid. We"d make an offer, and they"d already have offers and we"d lose. The amount to money we"ve been paying for people to come in has come up.

I spend a lot of time working on employee retention. My hope is that when they get calls from headhunters, they won"t even give them a call back. That"s when I know I"ve had success.

How do you ensure that? We"re putting a more realistic telecommuting policy in place. Especially with such high gas prices, it makes more sense. We"re constantly trying to find and fund new and fun projects for people to work on.

I just took my IT team out rock climbing for lunch last Friday. It was great. It was team building. Everyone gets excited about racing up the wall. We have a ping pong table at work, and at night they constantly play and have matches. It develops relationships across teams. Before you"d have a development team and a site operations group that sits by itself at lunch. Now they"re spending more time with each other. I try to do more team building. My thought is that people don"t want to leave their friends.

What"s the greatest technology challenge these days? Security never goes away as an issue. Lately we"re spending more time on network- and security-related issues. We"ve spent so much time securing our networks and building them up and monitoring them, but I"m always concerned that security issues can occur within the organization and we may not have spent as much time as we should have on it.

Has disaster recovery become a key issue? We"ve got some very good disaster recovery plans. It"s never fast enough. I know we"re going to get there. Even when you"ve gotten to the point when you"ve got disaster recovery that works, if you"re not consistent in your testing, the environment is always changing. Then when you really need it, you"re going to say, "Shoot, we changed that but didn"t change our high-end systems." We need to test it frequently enough so that we know it will work when we need it. I don"t feel 100 percent there yet. We"re headed down that path, but it"s all process and procedures. When things change, you have to update the documentation in every task.

If you"re not testing it in a way that validates it, I"m not sure you have a 100 percent reliable solution.

What things in disaster planning are most key: data redundancy, long-distance replication? It"s convincing the executive management to properly invest to do [disaster recover]. It"s really an insurance policy. It"s about how much you want to pay for insurance if that happens. We"re starting to see more natural disasters, so it"s becoming an easier project to get proper funding [for]. As for the technology piece of it, for us it"s the replication of data. Making sure we"ve got all the data we need in the right location.

If you need more servers, worst case, you go to your vendor and ask for another hundred. Either you sell them to me or I go somewhere else. Physical infrastructure is relatively easy. The data is not. I personally would have a real concern -- and I"m not loading data from tapes -- but if I was, I remember when I worked for IBM I used to participate in [disaster-recovery] drills where we"d load data off tapes. Every time we tested, there was always some tape that got left behind or had an I/O error.

So we"re going disk-to-disk in different locations. I wouldn"t be happy if I were dependent on recovering data from tapes.

So are you 100 percent on disk-to-disk backup? We have some data going disk-to-disk, not all of it. But that will be completed before the end of the year.

Are you getting rid of tape all together, or are you keeping it for archival only? Yes. We"re getting rid of it all together.

Are you going with EMC"s Centera, a virtual tape library. What? I"m piggybacking on the Weather Channel"s system right now. They"re picking, and I"m just using it. They"re still evaluating it.

We"ve bought some Hitachi [virtual tape libraries] and we"re running NetBackup [from Symantec"s Veritas division]. We like simple solutions. We"re going over Gigabit links. It"s a really easy configuration. It"s a triangle between two data centers and our home location.

What other issues have your attention? We need more support for Linux and open-source platforms. I think slowly you"re starting to see that, but it"s slow.