Information Technology Reference
In-Depth Information
Information Systems in the Global Economy
Telecom Giant Uses Knowledge Management and Expert Systems
A corporation's success depends on the knowledge it maintains and uses through its
employees, executives, board of directors, and information systems. Building that body of
corporate knowledge is a primary goal of most businesses—a goal that requires a con-
certed effort. If left untended, corporate knowledge pools in certain people, who gain it
through time and experience, and is no longer accessible when those people leave the
organization. Information systems, such as knowledge management systems and expert
systems, can store knowledge gained by those within a corporation over time. For expert
systems, artificial intelligence (AI) techniques can automate expert reasoning and activi-
ties. Consider how an expert system allows telecom giant Ericsson to monitor telecom
networks in a manner that would overwhelm a human expert.
Established in 1876 and headquartered in Stockholm, Sweden, Ericsson is a world
leader in telecommunications services, networking, multimedia solutions, and core tech-
nologies for mobile handsets. In fact, ten of the world's largest mobile phone operators use
Ericsson technologies and 40 percent of all mobile traffic travels through Ericsson sys-
tems. Ericsson's 50 percent share of Sony Ericsson Mobile Communications has further
expanded its power in the market.
Ericsson's busy telecom networks, as with all complex networks, encounter many
obstacles over the course of the day. Lines may go down; some percentage of the
thousands of routers, junction boxes, transmitters, and other telecom hardware devices
used may fail; and systems software may experience bugs. Each time a problem, or fault,
occurs, network administrators are alerted through dashboards that test and monitor
the network.
Ericsson network administrators work in a stressful environment, where streams of
alarms indicating network faults need to be evaluated and acted upon, sometimes imme-
diately. Therein lays the challenge. It is difficult, if not impossible, for network adminis-
trators to determine which alarms are important and need immediate attention and which
can wait. Given time, the administrators could read, interpret, and analyze each alarm to
determine its level of importance, but with hundreds of thousands of alarms flowing in
each day, there isn't even enough time to read one before ten more have arrived. This type
of problem is well suited for an expert system.
Ericsson worked with an expert system company to assist with network fault man-
agement. The expert system company used real-time rule technology to automate the
process of monitoring network alarms and determining which alarms needed immediate
attention and which could wait. The human experts at Ericsson assisted the company in
designing the software by sharing the secrets to the complicated process of alarm inter-
pretation. With more than 50 types of equipment sending more than 500,000 alarms each
day, the expert knowledge that was collected needed to be complete and executed quickly.
Today the expert system for managing network faults is deployed in 500 Ericsson
systems in over 100 countries. The stream of alarms faced by human experts has been
slowed to a tolerable amount of only the most important problems that require immediate
intervention. The performance and quality of Ericsson networks has improved, its admin-
istrators can focus on priorities, and important corporate knowledge has been digitized to
benefit the company over generations of employees.
