Waheed, A., Herman D. Hughes, and Diane T. Rover, "A Resource Occupancy Model for Evaluating Instrumentation System Overheads," Proceedings of the 20th Annual International Conference of the Computer Measurement Group (CMG `95), Nashville, Tennessee, Dec. 3--8, 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....two metrics to help answer a number of what if questions. 6 2. 2 Resource Occupancy Model This subsection introduces the Resource OCCupancy (ROCC) model and its application to isolating the overheads due to non deterministic sharing of resources between the Paradyn IS and application processes [29]. The ROCC model, founded on traditional modeling techniques, consists of three components: system resources, requests, and management policies. Resources are shared among (instrumented) application processes, other user and system processes, and IS processes; for example, CPU, network, and I O ....

....to the main Paradyn process: the data forwarding policy determines the manner in which a Paradyn daemon sends data samples to the main Paradyn process. Initial back of the envelope analytical calculations of the ROCC model for the Paradyn IS indicated that the CPU is the bottleneck resource [29]. Therefore, we do not investigate the network occupancy time in this study. Our specific interest is in the analysis of the CPU time taken by the Paradyn daemon (i.e. its direct overhead) and the throughput of data forwarding (i.e. number of data samples forwarded per unit of time) We use a 2 ....

[Article contains additional citation context not shown here]

Waheed, A., Herman D. Hughes, and Diane T. Rover, "A Resource Occupancy Model for Evaluating Instrumentation System Overheads," Proceedings of the 20th Annual International Conference of the Computer Measurement Group (CMG `95), Nashville, Tennessee, Dec. 3--8, 1995.

....these metrics to help answer a number of what if questions. 2. 2 Resource Occupancy Model This section introduces the Resource OCCupancy (ROCC) model and its application to isolating the overhead due to nondeterministic sharing of resources between the Paradyn IS and application processes [29]. The ROCC model, founded on well known computer system modeling techniques, consists of three components: system resources, requests, and management policies. Resources are shared among (instrumented) application processes, other user and system processes, and IS processes; for example, CPU, ....

....Comparison of CPU overhead measurements under the CF and BF policies using two sampling period values for: a) Paradyn daemon and (b) main Paradyn process. WAHEED et al. MODELING AND EVALUATING DESIGN ALTERNATIVES FOR AN ON LINE INSTRUMENTATION SYSTEM: A CASE STUDY 465 vided by the developers [29]. The case study presented in this paper uses a more detailed workload characterization based on measurement data. Although we enhanced the scope of the what if questions in this study, e.g. to include the SMP and MPP architectures and factors such as forwarding policy and length of ....

[Article contains additional citation context not shown here]

# A. Waheed, H.D. Hughes, and D.T. Rover, "A Resource Occupancy Model for Evaluating Instrumentation System Overheads," Proc. 20th Ann. Int'l Conf. Computer Measurement Group (CMG `95), pp. 1,212--1,223, Nashville, Tenn., Dec. 1995.

....Therefore, in this paper, we take up the problem of designing an adaptive controller to maintain the overhead and intrusion of the IS within the specified desired levels for a particular real time application. We apply the Resource OCCupancy (ROCC) modeling technique proposed by Waheed et al. in [22] to model the sharing of resources in a distributed system consisting of a target application, IS, and adaptive controller. The ROCC modeling technique was chosen for this study due to its application specific workload characterization technique, which is distinctive for its short turn around time ....

.... is based on three components: shared resources of the system; occupancy requests to the shared resources; and behavior of each process of interests that determines the sequence of resource usage, resource occupancy requirements for each task, and interaction or synchronization with other processes [22]. There are three shared resources of interest at each system node: CPU, local X server (i.e. graphical display device) and network. The camera that is used by the application server process to capture the scene is not shared with any other process; therefore, Table 1. Metrics for evaluation of ....

Waheed, A., Herman D. Hughes, and Diane T. Rover, "A Resource Occupancy Model for Evaluating Instrumentation System Overheads," Proceedings of the 20th Annual International Conference of the Computer Measurement Group (CMG `95), Nashville, Tennessee, Dec. 3--8, 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC