Rothstein Associates Inc.

ENSURING FACILITY INFRASTRUCTURE RELIABILITY: MAKING COMPUTER
ENVIRONMENTAL SYSTEMS FAULT TOLERANT
by Kenneth G. Brill

"An information outage is becoming the corporate equivalent of a heart attack. The risk of
such outages is rapidly increasing because there is an unrecognized and growing divergence
between end-user requirements for information availability and the limitations of the physical
facility and its environmental infrastructure to support uninterrupted computer operation. Ten
years ago, site reliability problems accounted for just 3% of total downtime. Today it is up to
10%, and by 1995, I project that site malfunctions will account for 20% to 30% of total
processing downtime."

"Reversing these trends requires top management to redirect resources, but not necessarily
to spend more money. This paper explores current problems and develops a conceptual
framework and strategy for the management, organization, and engineering steps required to
make the facility and its environmental infrastructure more reliable and fault tolerant."

CONTENTS
Summary
Problems Addressed
The Site Is Becoming the Weakest Reliability Link
Downtime Continues until All Users Are Back up and Working Productively
Nonstop Data Center Facilities
Facilities and Infrastructures That "Work' Versus Fault Tolerant Ones That Never Fail
Different Objectives Are Used to Evaluate Performance
A Fundamental Organizational Principle Is Being Violated
Facility Service Level Agreement
Reliability Criteria Are an Essential Communication Tool
If Nonstop Operation Is a Requirement, How Much Redundancy Is Enough?
Reliability Criteria Are Applicable to All Site Subsystems
A New Analytical Approach Is Needed
Typical Vulnerabilities
Create a Level Playing Field for Competitive Bidding
Experience Is Uniquely Critical in Evaluating Reliability
Case Study: the American Express Approach to Ensuring Reliability
A Recommended Course of Action

ABOUT THE AUTHOR

Kenneth G. Brill is an uptime consultant specializing in the engineering and management
issues of building fault tolerant systems that never fail. He is also an advocate of making
disaster avoidance the first step in disaster recovery planning and has written extensively on
issues of infrastructure reliability. Brill is the President of Computersite Engineering and is the
founder of the Uninterruptible Uptime Users Group. Brill has a degree in electrical engineering
from the University of Redlands and an MBA from the Harvard Business School.

1990, 30 pages Order #DR-110.

4 Arapaho Rd.

Brookfield, CT 06804-3104

1-888-ROTHSTEin

Telephone: 203.740.7444; 888.768.4783

Fax: 203.740.7401

All bookstore enquiries should be sent to Rothstein Associates at the above address.

Looking for Practical Knowledge?