Quality-based adaptive resource management architecture (QARMA): a CORBA resource management service

Design and Performance of Configurable Endsystem Scheduling Mechanisms

“…That creates a slightly richer approach to describing some aspects of system scheduling semantics, particularly with respect to middleware enforcement mechanisms, but the fundamental assumption of priority scheduling semantics remained unchanged. A slightly higher level abstraction using essentially the same dominant decomposition is the CORBA-based resource broker service [5]. The higher level abstractions given by the resource broker include the ability to consider a range of system state information and to use either priority or share based underlying scheduling semantics, but the approach still assumes a static and uniform underlying scheduling semantics from the endsystem.…”

Integrated CORBA Scheduling and Resource Management for Distributed Real-Time Embedded Systems

“…Our integrated approach consists of a distributed scheduling (DS) service [4], which coordinates the activities of multiple endsystems, a distributed resource management (DRM) service [6], which allocates resources on each endsystem, and several alternative endsystem scheduling options, e.g., Real-Time CORBA 1.2 fixed priority schedulers [17], Kokyu [14], and group scheduling [8]. We have integrated our DS and DRM services atop the TAO [30] CORBA [24] object request broker, and have performed experiments to validate and evaluate this integration, which we present in Section 4.…”

“…We have implemented a distributed resource management service in CORBA as part of the Qualitybased Adaptive Resource Management Architecture (QARMA) [6] project. The QARMA DRM service consists of the following types of elements:…”

“…Integration of the DS and DRM services: The integration of the DS service with the QARMA Resource Manager [6], and with TAO's RTC 1.2 scheduling service, is shown in Figure 2. In the interactions depicted in Figure 3, the application makes a begin_scheduling_segment() call (labeled bss in the figure) to a distributable thread (labeled DT).…”

Robust Task Allocation for Dynamic Distributed Real-Time Systems Subject to Multiple Environmental Parameters

“…High latency may result from a poorly allocated system, when changes in external variables frequently trigger reallocations, and the overhead may cause task deadline misses and serious consequences. The algorithm developed in this research will be used for task allocation decisions in the Resource Management Service of QARMA, the Quality-based Adaptive Resource Management Architecture [6]. Figure 2 shows the architecture, which consists of three major components: the System Repository Service, the Resource Management Service, and the Enactor Service.…”