H. Tokuda, M. Kotera and C. W. Mercer, `A real-time monitor for a distributed real-time operating system', ACM SIGPLAN/SIGOPS Workshop on Parallel and Distributed Debugging, published in ACM SIGPLAN Notices, 24, (l), 68--77 (1989).  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... on monitoring and debugging facilities for distributed and parallel systems, run time monitoring of real time systems has received little attention (with a few exceptions) Special hardware monitoring for collecting run time data in real time applications has been considered in a number of papers [CJD91, MoL97, RRJ92 and TKM88]. These approaches introduce specialized co processors for the collection and analysis of run time information. The use of special purpose hardware allows for non intrusive monitoring of a system by recording the run time information in a large repository, often for post analysis. In related work ....

Tokuda H., Kotera M., Mercer C.W. (1988); "A Real-Time Monitor for a Distributed Real-Time Operating Systems"; Proceedings of the ACM Workshop on Parallel and Distributed Debugging; 1988

....efficient and deterministic observations. 4.1 Monitoring Research on testing of shared memory multiprocessor systems and distributed systems have been penetrated in some detail over the years. The focus has mainly been on monitoring, i.e. gathering of run time information for debugging [86][111][112] and performance measurements [11] The research issues have been: The intrusiveness (perturbation) of observations in software [31] 64] 78] and how to eliminate the perturbations using special hardware [113] How to deterministically reproduce the observations using mechanisms in ....

....in the software (application and kernel) we can observe significantly more than possible with hardware approaches. Software monitoring of real time systems have been covered by Chodrow et al. 14] distributed systems by Joyce et al. 46] 80] and distributed real time systems by Tokuda et al. [111][90] In general it is necessary to leave the software probes in the target system in order to eliminate the probe effect. If the target system is a real time system, which can be scheduled e.g. using fixed priority or static scheduling it is straightforward to analyze the effects that the probes ....

Tokuda H., Kotera M., and Mercer C.W. A Real-Time Monitor for a Distributed Real-Time Operating System. In proc. of ACM Workshop on Parallel and Distributed Debugging, Madison, WI, pp. 68-77, May, 1988.

....by a monitor like HMON as soon as the system has to interact with other computers that are not part of the monitored domain. Tokuda, Kotera, and Mercer proposed in 1988 a real time monitor featuring the visualization of the internal behavior of a distributed real time operating system ARTS [Tok88]. It consists of a real time monitor debugger to visualize the target systems scheduling decisions in quasi real time by Gantt diagrams. Information is gathered by a software sensor, called an Event Tap, embedded into the realtime operating system kernel. To predict and reduce the monitoring ....

Tokuda, H., Kotera, M., Mercer, C. W., A real-time monitor for a distributed real-time operating system, Proc. of ACM Workshop on Parallel and Distributed Debugging, Madison, WI, USA, pp. 68-77, May 1988.

.... effects which cause non deterministic behaviour in programs with race conditions and poor synchronisation [6, 13] The research efforts on real time monitoring has over the past decade been mostly devoted to dealing with probe effects and timing interference in various applications of monitoring [17, 18, 2, 8, 7]. Hence, a wide spectrum of monitoring approaches have been proposed, ranging from pure software techniques [17, 8] to the use of special hardware support [12, 18, 7] Software monitoring systems offer the cheapest and most flexible solution where a common technique is to insert instrumentation ....

.... efforts on real time monitoring has over the past decade been mostly devoted to dealing with probe effects and timing interference in various applications of monitoring [17, 18, 2, 8, 7] Hence, a wide spectrum of monitoring approaches have been proposed, ranging from pure software techniques [17, 8] to the use of special hardware support [12, 18, 7] Software monitoring systems offer the cheapest and most flexible solution where a common technique is to insert instrumentation code at interesting points in the target software. When the instrumentation code is executed the monitoring process ....

[Article contains additional citation context not shown here]

H. Tokuda, M. Koreta, and C. Mercer. A real-time monitor for a distributed real-time operating system. ACM Sigplan Notices, 24(1):68--77, January 1989.

....timing of the monitored programs. The majority of monitoring and debugging tools (e.g. the RED [4] or DCT [1] methods) require dedicated hardware. By software approaches, the source of the monitored program is modified, inserting extra instructions which collect trace data [6] or extra processes [12]. Hybrid techniques like [2] or [3] combine software modification and hardware based information collection. Some error detection techniques widely used in highly dependable systems often provide mechanisms to derive the trace of the program with minor cost and additional effort. Our goal is to ....

H. Tokuda, M. Kotera, and C. W. Mercer. A real-time monitor for a distributed real-time operating system. In Proc. ACM/ONR Workshop on Parallel and Distributed Debugging, pages 68--71, 1988.

....multiprocessors, current research is almost exclusively concerned with performance issues and monitoring of control related (internal) and or environmental (external) constraints. A related work, however, is the real time monitor developed for the ARTS distributed operating system (Tokuda et al. [14]) The proposed monitor requires support from the kernel such as notification of process state changes, e.g. awakening from a blocked state, being scheduled, etc. The collected events are then sent to a remote host for displaying the execution history. The main difference between this monitor and ....

H. Tokuda, M. Koreta, and C. Mercer. A real-time monitor for a distributed real-time operating system. ACM Sigplan Notices, 24(1):68--77, January 1989.

....the next version, the microprocessors in the modules will also serve as low level controllers for the joint actuators, making the control system, to some extent, distributed. A number of testbeds have been reported in the literature, e.g. SPRING and ARTS (Stankovic and Ramamritham, 1991; Tokuda, 1988). These testbeds, that we are aware of, are computer science oriented and focus primarily on scheduling and debugging. In the field of applied and theoretical control it is quite common to use dedicated and resource adequate computer architectures and to focus on control concepts. One ....

Tokuda H., M. Kotera, C.W. Mercer (1988). A real-time monitor for a distributed real-time operating system.

....s = a b; printf( c= d n , c) return(s) 12 4 MONITORING Research on testing of shared memory multiprocessor systems and distributed systems have been penetrated in some detail over the years. The focus has mainly been on monitoring, i.e. gathering of run time information for debugging [47][66][67] and performance measurements [6] The research issues have been: The intrusiveness (perturbation) of observations in software [18] 36] 42] and how to eliminate the perturbations using special hardware [68] How to deterministically reproduce the observations using mechanisms in ....

....code in the software (application and kernel) we can observe significantly more than possible with hardware approaches. Software monitoring of real time systems have been covered by Chodrow et al. [9] distributed systems by Joyce et al. [24] 44] and distributed real time systems by Tokuda et al. [66][51] In general it is necessary to leave the software probes in the target system in order to eliminate the probe effect. If the target system is a real time system, which can be scheduled e.g. using fixed priority or static scheduling it is straightforward to analyze the effects that the probes ....

Tokuda H., Kotera M., and Mercer C.W. A Real-Time Monitor for a Distributed Real-Time Operating System. In proc. of ACM Workshop on Parallel and Distributed Debugging, Madison, WI, pp. 68-77, May, 1988.

....system wide scheduling analysis which includes communication and synchronization among real time tasks. It can predict and analyze the schedulability of the application task set and perform runtime monitoring and debugging [3] Currently, the toolset consists of Scheduler 1 2 3 [4] and ARM [5]. 2.1 Scheduler 1 2 3 Scheduler 1 2 3 is an X11 window based interactive schedulability analyzer for creating, manipulating, and analyzing real time task sets. It employs analysis methods ranging from closed form analysis to simulation to determine whether a feasible schedule exists for a given ....

H. Tokuda, M. Kotera, and C. W. Mercer. A Real-Time Monitor for a Distributed Real-Time Operating System. In Proceedings of ACM SIGOPS and SIGPLAN workshop on parallel and distributed debugging, May 1988. 62

....work on timing analysis has mainly concentrated on execution timing analysis. This includes compiler support for computing WET estimates [9, 10, 12] and the run time refinement of WET estimates based upon a combination of compile time information [15, 6] and run time information [3, 5] In [19] authors present the straightforward approach of introducing the monitoring activity at the user selected points and [20] presents a hardware solution to non intrusive monitoring by providing architectural support. Our approach can enhance the effectiveness of monitoring systems such as the ISSOS ....

H. Tokuda, M. Kotera, and C.W. Mercer, "A Real-Time Monitor for a Distributed Real-Time Operating System," Technical Report, Carnegie Mellon University, CMU-CS-88-179, 1988.

....interface for real time purposes and synchronizes on breakpoints by suspending the execution on all processors. DCT [9] is a tool that allows practically non intrusive monitoring but requires special hardware for bus access and does not extend to non intrusive debugging. Both RED [30] and ART [67] provide monitoring and debugging facilities at the price of software instrumentation. RED dedicates a co processor to collect trace data and send it to the host system. The instrumentation is removed for production code. In ART, a special reporting task sends trace data to a host system for ....

H. Tokuda, M. Kotera, and C. W. Mercer. A real-time monitor for a distributed realtime operating system. In ACM/ONR Workshop on Parallel and Distributed Debugging, pages 68--77, 1988.

....(rt; ls) rt; dl) in VTD ) where ls denotes the latest sending time of the message ( and dl denotes the deadline) Finally, the message is put back into the message queue. 2. 2 Monitoring Methods Monitoring techniques, long used for system testing, debugging and performance monitoring [Cas91, HW90, LSMC90, TFCB90, TKM88], have been applied to critical and or real time systems to detect system errors or property violations [Gor91, CJD91, JG90, MM88, WS88, Lu82] Similar techniques have also been applied in real time systems to aid scheduling of real time tasks [GG90, HS90] the basic idea is to enable scheduling ....

H. Tokuda, M. Kotera, and C.W. Mercer. A real-time monitor for a distributed real-time operating system. Proceedings of the ACM SIGPLAN and SIGOPS Workshop on Parallel and Distributed Debugging, pages 68--77, May 1988.

....specification and access to monitoring information at run time [22] The assumption is that the relational model is an appropriate formalism for structuring the information generated by a distributed system. A real time monitor developed for the ARTS distributed operating system is presented in [24]. The proposed monitor requires certain support from the kernel, such as notification of the state changes of a process, including waking up, being scheduled. In particular, the ARTS kernel records certain events that are seen by the operating system as the state of a process changes, e.g. ....

Tokuda, H., Koreta M. and Mercer, C. W. Jan. 1989. A Real-Time Monitor for a Distributed Real-Time Operating System, ACM Sigplan Notices 24,1, pp. 68-77.

....monitoring task with an application task, the compiler can analyze the busy idle profiles to determine whether the requests can be handled. Consider the situation shown in Figure 13a in which monitoring code is to be accommodated in a real time schedule. The straightforward approach, suggested by Tokuda, Kotera, and Mercer (1988), is simply to insert the monitoring task directly at the user specified point, as shown in Figure 13c. However, this technique has the following drawback. By placing the code directly at the specified point, the completion of application task may be delayed and the task s deadline may be missed ....

....On the other hand Gerber and Hong s approach requires analyzing the timing behavior of the tasks resulting from the slicing transformation. Our approach for non intrusive monitoring of real time applications can enhance the effectiveness of existing systems (Ogle, Schwan, and Snodgrass, 1993, Tokuda, Kotera, and Mercer, 1988, Haban and Shin, 1989) In contrast to the non intrusive monitoring techniques developed by Tsai, Fang, and Chen (1990) which require specialized hardware support, our approach is based entirely upon software techniques. 6. Concluding Remarks We presented techniques for the timing analysis of ....

Tokuda, H., M. Kotera, and C.W. Mercer, "A Real-Time Monitor for a Distributed Real-Time Operating System," ACM Workshop on Parallel and Distributed Debugging, May, 1988, pp.

....[14] addresses the customization of the debugging interface for real time purposes and synchronizes on breakpoints by suspending the execution on all processors. DCT [5] is a tool that allows practically non intrusive monitoring but requires special hardware for bus access. Both RED [8] and ART [17] provide monitoring and debugging facilities at the price of software instrumentation. RED dedicates a co processor to collect trace data and send it to the host system. The instrumentation is removed for production code. In ART, a special reporting task sends trace data to a host system for ....

H. Tokuda, M. Kotera, and C. W. Mercer. A realtime monitor for a distributed real-time operating system. In ACM/ONR Workshop on Parallel and Distributed Debugging, pages 68--77, 1988.

....[1, 11, 13, 21, 26] These approaches use dedicated hardware to detect event occurrences by snooping and matching bus signals of target systems and storing event data for post processing. These methods are especially suitable for monitoring hard realtime systems because they are non intrusive. In [3, 8, 25], software monitoring approaches are proposed which insert event detection and event data collecting code into application programs, operating system kernels or monitoring systems. In [25] a real time monitor (ARM) for a distributed real time operating system (ARTS) is described. Events, which ....

....methods are especially suitable for monitoring hard realtime systems because they are non intrusive. In [3, 8, 25] software monitoring approaches are proposed which insert event detection and event data collecting code into application programs, operating system kernels or monitoring systems. In [25], a real time monitor (ARM) for a distributed real time operating system (ARTS) is described. Events, which are generated whenever a process changes state, are captured by the ARTS kernel and sent to a visualizer on another machine for displaying and analysis. In [3] a monitoring system is ....

H. Tokuda, M. Kotera, and C.W. Mercer, "A RealTime Monitor for a Distributed Real-Time Operating System", Proc. ACM Workshop Parallel and Distributed Debugging, 1988, pp.68-77.

....They also cause little interference to the monitored system. However, hardware monitors do not support the interactive modification of task execution that is necessary to support debugging. A number of monitoring and debugging systems have been developed specifically for real time systems [22, 24, 28, 29]. However, interactive debugging of real time programs without deterministic replay is not sufficient because debugging commands destroy the timing dependent nature of real time systems. Similarly, breakpoints must only be inserted during replay, not normal execution, to maintain the consistent ....

H. Tokuda, M. Kotera, and C. W. Mercer, "A real-time monitor for a distributed real-time operating system," Proc. ACM SIGPLAN/SIGOPS Workshop on Parallel and Distributed Debugging, published in ACM SIGPLAN Notices, vol. 24, no. 1, pp. 68--77, January 1989.

....tool in debugging logical errors as well as assessing the accuracy of the WCETs. The runtime monitoring can collect information on such aspects as message arrival and contents, external input reception, the execution time of various statements and statement groups and changes in variable status [3, 4, 5]. However, a significant drawback Supported in part by the National Science Foundation through Grant CCR 9212020 and a Presidential Young Investigator Award CCR 9157371, NASA Connecticut Space Grant Consortium and Trinity College of runtime monitoring is that it introduces intrusion, that is, ....

....execution time of the modules being monitored as well as potentially alter synchronization activities among the modules. In static scheduling in which each task is allotted a fixed period in which to execute, one approach is to perform monitoring of a task as possible within its allocated period [5]. However, a problem arises if insufficient time is available for monitoring, leading to an inability to carry out the required monitoring in the time period allotted. To increase the effectiveness of a runtime monitoring tool, the tool must provide the user with a view of the module s behavior ....

H. Tokuda, M. Koreta, and C. Mercer, "A Realtime Monitor for a Distributed Real Time Operating System," ACM Sigplan Notices, vol. 24, no.1, pages 68-77, January 1989.

.... of distributed parallel (real time) programs [9] The work is based on an earlier system for exploring the use of an extended E R model for specification and access to monitoring information at run time [11] A real time monitor developed for the ARTS distributed operating system is presented in [13]. The rest of this paper is organized as follows. Section 2 describes an aircraft tracking system and discusses issues in detecting a violation of a timing assertion. Section 3 presents our event based computation model, and discusses specification of timing assertions. Section 4 presents a ....

H. Tokuda, M. Koreta and C.W. Mercer. A RealTime Monitor for a Distributed Real-Time Operating System, ACM Sigplan Notices 24,1, pp. 6877, Jan. 1989.

....their structure or modifying their parameters. The tool does not exercise this control directly. System monitoring tools typically separate the functions of capturing performance data, analyzing the data, and having an effect on the system that was measured. The Advanced Real Time Monitor (ARM) [123], which was originally designed for the ARTS Kernel [124] and more recently updated for RT Mach [125] takes this approach. ARM uses a kernel mechanism to capture scheduling events and then sends those events over the network to an ARM application running on a different machine. ARM then displays ....

H. Tokuda, M. Kotera, and C. W. Mercer. A Real-Time Monitor for a Distributed Real-Time Operating System. In Proceedings of ACM SIGOPS and SIGPLAN workshop on parallel and distributed debugging, May 1988.

....a system wide scheduling analysis which includes communication and synchronization among real time tasks. It can predict and analyze the schedulability of the application task set and perform runtime monitoring and debugging [3] Currently, the toolset consists of Scheduler 1 2 3 [4] and ARM [5]. 2.1 Scheduler 1 2 3 Scheduler 1 2 3 is an X11 window based interactive schedulability analyzer for creating, manipulating, and analyzing real time task sets. It employs analysis methods ranging from closed form analysis to simulation to determine whether a feasible schedule exists for a given ....

H. Tokuda, M. Kotera, and C. W. Mercer. A Real-Time Monitor for a Distributed Real-Time Operating System. In Proceedings of ACM SIGOPS and SIGPLAN workshop on parallel and distributed debugging, May 1988.

....an integrated time driven scheduler (ITDS) real time synchronization, and memory resident objects. The real time thread model is based on the ARTS real time thread model [21, 25] and our real time scheduling theories. Real Time Mach was also integrated with our real time toolset, Scheduler 1 2 3[23] and Advanced Real Time Monitor, ARM[22] In this paper, we describe new system facilities in RealTime Mach and the current status. In Section 2, we first introduce the real time thread model, real time synchronization primitives, integrated time driven scheduler, and support for memory resident ....

H. Tokuda, M. Kotera, and C. W. Mercer, "A realtime monitor for a distributed real-time operating system", In Proceedings of ACM SIGOPS and SIGPLAN workshop on parallel and distributed debugging, May, 1988.

....it does not provide any safe mechanisms for creating a periodic thread and avoiding priority inversion problems. Yet another uniqueness of RT Mach is that a user can use a set of real time toolset, Scheduler 1 2 3 and ARM, which allows us to analyze, simulate, and monitor the real time task set[28, 29, 30]. The proposed real time thread model is different from many other thread models. In particular, 1) our model distinguishes between real time and non real time threads, 2) we provide explicit timing constraints for each real time thread, 3) a scheduling policy for a real time application can be ....

H. Tokuda, M. Kotera, and C. W. Mercer, "A real-time monitor for a distributed real-time operating system", In Proceedings of ACM SIGOPS and SIGPLAN workshop on parallel and distributed debugging, May, 1988.

No context found.

H. Tokuda, M. Kotera and C. W. Mercer, `A real-time monitor for a distributed real-time operating system', ACM SIGPLAN/SIGOPS Workshop on Parallel and Distributed Debugging, published in ACM SIGPLAN Notices, 24, (l), 68--77 (1989).

No context found.

H. Tokuda, M. Kotera, and C.W. Mercer, "A RealTime Monitor for a Distributed Real-Time Operating System", Proc. ACM Workshop Parallel and Distributed Debugging, 1988, pp.68-77.

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