No earlier issue with same topic
Issue
Previous article
Article
SIGCHI Bulletin
Vol.29 No.4, October 1997
Next article
Article
No later issue with same topic
Issue

HCI Design for Network and System Management

A CHI 97 Special Interest Group

Thomas M. Graefe and Dennis Wixon

Introduction

All too often the people responsible for the care and feeding of the information technology infrastructure are poorly supported by the very technology they must manage, even as the popularity and use of networks (such as for the World Wide Web) grows. Corporate MIS staffs spend billions of dollars just on managing their computing infrastructures, and still they must continually cope with ineffectual products that do not support them in their work. A $2,000 PC may cost $5,000 to $10,000 a year to support.(1)

This Special Interest Group (SIG) provided an opportunity for over 30 HCI practitioners and researchers in the domain of network and system management to share information about the problems faced by operators, system managers, administrators, and end users, and to explore new techniques in user interface design that might provide better support in the future.

The group spent the majority of its time sharing information about design problems in a structured brain-storming exercise. Candidate areas for solutions were considered in response to the defined problem.

Design Challenges

Each participant was asked to report on the most important task or key problem area their users faced. These were then clustered to define focus areas.

Information Visualization

Representation and manipulation of large collections of objects is a common requirement. There is a need for both broad views and rapid drill-down, for real-time monitoring and trend-analysis, and for customization of the display for specific tasks and user preferences.

Capture and Access to Trouble Shooting Information

Diagnosis and problem solving in a chaotic heterogeneous networking environment with widely varying information needs to be supported within applications. Information capture is critical and automated support of correlation across objects and time key.

Scalability of Solutions

Organizations need to be supported with applications that can span entire companies while at the same time allowing subset domains to be established and controlled. As companies grow or change the tools should be able to do likewise.

Education and Training

For small and medium businesses where the network or system management is often a part time job for an employee there needs to be a way to supply expertise and training as part of the application. These users are not experts, which is the typical requirement for use of management products.

Customization of Interface

Users and organizations need to be able to tailor their applications to suit scope of responsibility and task. This includes customization of reports and displays of collections of managed objects.

Inadequacies of the Window Paradigm

Window proliferation make complex nested displays unwieldy. Lack of a command language inhibits experts and frustrates home grown automation.

Managing Relationships among Domain Objects

Networks and systems are defined by relationships among their elements, but few applications enable management of key relationships in a robust fashion. (e.g., `provide services for', `linked to', etc.) This leaves it to the user to recreate such relationships

Lack of Standards

This is an ongoing problem because it aggravates the already bad problem of heterogeneity. The lack of standard management protocols or uniform implementations is typically directly manifested to the operator or administrator.

Inadequate Product Engineering

Many network and system products are not designed to be easily managed. Product designs do not take into account the context of use, and often the groups designing management applications are separate from those designing the products being managed resulting in seemingly full featured products that are difficult to install, costly to upgrade, and hard to configure or monitor.

Where to Find Solutions

Discussion of challenges was so enthusiastic it left little time for solutions. In short, the group discussed general topics such as the addition of performance support to applications, the use special purpose displays, and incorporating expert system technology into applications as areas needing broader use.

Future Meetings

The attendees were extremely enthusiastic about continuing the SIG at CHI 98, and were also interested in a forum where some of the challenges and solutions could be pursued in depth. A workshop on this area was discussed, and the authors invite anyone interested in such a workshop to contact them at the addresses provided below.

About the Authors

Tom Graefe has been in the Usability Design Center at Digital Equipment since February, 1995. He has worked on HCI design and issues in networking and system management since 1981

Dennis Wixon has been working in User Interface design for over 20 years. He is currently program manager for usability at Digital Equipment Corporation.

Authors' Addresses

Thomas M. Graefe
Digital Equipment Corporation
550 King St. Littleton, MA 01540 USA
email graefe@tnpubs.enet.dec.com
tel 1+508-486-7335

Dennis Wixon
Digital Equipment Corporation
110 Spitbrook Rd.
Nashua, NH 03062 USA
tel +1 603 881 2276
email wixon@orion.enet.dec.com


Footnotes

(1)
for example, see Gibbs, W.W. Taking Computers to Task. Scientific American, April, 1997.

No earlier issue with same topic
Issue
Previous article
Article
SIGCHI Bulletin
Vol.29 No.4, October 1997
Next article
Article
No later issue with same topic
Issue