Automated Systems Administration: Taming the Chaos in IT

Most IT organizations still install and maintain computers the same way the automotive industry built cars in the early 1900’s: An individual craftsman manually manipulates a machine into being, and manually maintains it afterward. In most IT shops, relatively little effort is allocated for automating changes made to systems in order to gain economies of scale. In losing repeatability, we also lose safety — reliability of standby systems is mixed, and even disaster recovery planning is often considered to be a paper exercise rather than a technology driver.

But outside the systems administration field, it’s not well understood that computer systems administration is generally manual as well as ad-hoc. The CIOs of some of our best and brightest Silicon Valley companies are blissfully unaware of the chaos in their own ranks. In a global economy dependent on information technology, the friction generated by ineffective IT practices impacts productivity, jobless rates, quality of life, and the growth of emerging industries.

The automotive industry discovered first mass production, then mass customization using standardized tooling. This talk describes similar techniques that are being applied in IT today, ranging from tools for systems administrators to workable business models of interest to the executive officer. Developed over the last several years by a growing, loose-knit group of active UNIX and Linux systems administrators, architects, and IT executives, these methodologies have been proven in mission-critical environments as well as in recovery from major disasters. They lower the cost of providing IT infrastructure, increase data center scalability and efficiency, and make for rapid, reliable, and repeatable deployments and changes.

Presenter(s)

Steve Traugott's 1998 USENIX Large Installation Systems Administration (LISA) paper, "Bootstrapping an Infrastructure", helped launch the IT Infrastructure Architect career field. He is a former Vice President of trading floor engineering for Chase Manhattan Bank and a U.S. Air Force Special Operations veteran. His IT industry experience spans over 20 years and covers platforms ranging from embedded systems to supercomputers. He helped port the Mach kernel to mainframes, and UNIX System V to PC's. After September 11th, 2001, he returned to New York for three months to work with World Trade Center survivors.

Today Steve is a consulting Infrastructure Architect, and publishes tools and techniques for automated systems administration and disaster recovery. His clients have included Chemical Bank, Cisco, NASA, IBM, AT&T, DEC, Netscape, Sun, Caterpillar, Morgan Stanley, and the Central Bank of Trinidad.