Hardware measurement device for IBM system/360 time sharing evaluation
ACM '67 Proceedings of the 1967 22nd national conference
An analysis of some time-sharing techniques
Communications of the ACM
The instrumentation of multics
Communications of the ACM
Microprogrammed implementation of computer measurement techniques
MICRO 5 Conference record of the 5th annual workshop on Microprogramming
Interactive task behavior in a time-sharing environment
ACM '72 Proceedings of the ACM annual conference - Volume 2
An experimental study of computer system performance
ACM '72 Proceedings of the ACM annual conference - Volume 2
A tandem queueing model of a time-sharing computing system
ACM '72 Proceedings of the ACM annual conference - Volume 2
AMS: A software monitor for performance evaluation and system control
SIGME '73 Proceedings of the 1973 ACM SIGME symposium
Gathering and analyzing data from a computer system: A case study
ACM '75 Proceedings of the 1975 annual conference
A simulation model of GECOS III
ACM '71 Proceedings of the 1971 26th annual conference
Scheduling TSS/360 for responsiveness
AFIPS '70 (Fall) Proceedings of the November 17-19, 1970, fall joint computer conference
SPY: a program to monitor OS/360
AFIPS '70 (Fall) Proceedings of the November 17-19, 1970, fall joint computer conference
Experience gained in the development and use of TSS
AFIPS '72 (Spring) Proceedings of the May 16-18, 1972, spring joint computer conference
Tuning a virtual storage system
IBM Systems Journal
User behavior on an interactive computer system
IBM Systems Journal
Uses of virtual storage systems in a scientific environment
IBM Systems Journal
Investigation into scheduling for an interactive computing system
IBM Journal of Research and Development
Hi-index | 0.03 |
Analysis of the operation of a large-scale, complex operating system requires data that can be obtained only from “inside” the system as it is running. Such data can supplement that obtained via hardware measurements, and can support simulation models by providing realistic calibration data. The primary attribute of an internal software measurement technique is that it has access to, and can selectively record, system data stored in memory. This paper describes the methods employed in developing and using such a technique within the IBM System/360 Time Sharing System (TSS/360). A prime consideration in developing a software measurement technique to obtain internal data is that a suitable compromise between resolution and system degradation must be achieved, since the traditional hardware tools for gathering external data provided high resolution without degradation. This was accomplished in the design of the System Internal Performance Evaluation (SIPE) program, which functions as a physical extension to the TSS/360 resident supervisor and records significant events as they occur. When control reaches one of the SIPE “hooks,” which are strategically located in the supervisor, SIPE is entered to record an event. When SIPE has accumulated a quantity of events, it transfers the data to tape. Off-line reduction techniques subsequently permit detailed reconstruction of system data and logical flow. SIPE has been operational for over a year. Experience in using it has indicated that the actual degradation is so small that continuous recording during development, primarily for debugging, can readily be justified.