Dan Houston: Stamp of approval at USPS
- By Frank Konkel
- Oct 21, 2013
There is no such thing as a normal day when you're the guy responsible for managing database and storage services at a major federal agency.
So says Dan Houston, who oversees 3,300 databases, 400 applications and all 22 petabytes (and counting) of data collected by the U.S. Postal Service, mainly from its public-facing website and mail-tracking system.
Whether he is implementing the latest version of Oracle on hundreds of systems or supporting new business projects for customers, Houston faces new challenges almost daily. And with USPS' precarious financial situation, data and IT have never been more important to the agency's mission.
In terms of size and scope, few organizations in the world rival USPS' IT infrastructure, and Houston handles the data at the heart of it.
"With an environment that size, you don't have average days," he said. "It's a new challenge every day."
A self-professed data junkie, Houston designed and developed the USPS DMS Operation System, which features detailed configuration management of databases and storage, automatic deployment and monitoring of database administration and service management functions, on-demand performance reports and diagnostics, and encrypted administrator password management for more than 1,800 databases.
With the system, it takes only three employees to manage those databases during off-business hours and 12 employees to handle the 22 petabytes of multiplatform storage technology. That saves USPS 40,000 man-hours annually while delivering database and storage services at 99.999 percent availability.
Houston "is a guy we want to keep around," said John Edgar, vice president of IT at USPS. "He is tremendous at analyzing problems and figuring out technical solutions to what is going on, and he communicates well."
Houston plans to pursue an advanced degree as a data scientist and would like to see USPS further its data analytics initiatives.
"We're good at using data to tell us what happened," he said. "But we're not quite to the point of having the data tell us what [will happen] tomorrow."
Frank Konkel is a staff writer covering big data, mobile, open government and a range of science/technology issues. Connect with him on Twitter at @Frank_Konkel.