Component-Based Dynamic QoS Adaptations in Distributed Real-Time and Embedded Systems

Building for the Future: China's Progress in Space Technology During the Tenth 5-Year Plan and the U.S. Response

“…We instantiated QoS mechanisms as software components called qoskets [23,24], including compress, scale, and crop qoskets that shape image information objects; CPU and Diffserv qoskets that set operating system and network priorities, respectively; and control of the rate at which information objects are published. Each QoS mechanism takes as input a level to configure it, which is provided by the LRM dynamically based on the resource allocation, role, and QoS policies.…”

“…We conducted a demonstration of our prototype software at the Operational Information Management Principal Investigators meeting held in Washington, DC, April [24][25]2007. The demonstration was similar to the previous one, but showcased the allocation of QoS levels using multiple resources, using the approximation algorithm described in Section 4.3.…”

Applying model-driven development to distributed real-time and embedded avionics systems

“…The CFG also displays the documentation for each option. 2 The OCML options as applied to CIAO are available from $COS-MIC ROOT/PSM/CCM/OCML directory in the CoSMIC distribution. OCML tool alleviates complexities associated with configuring component middleware by ensuring both (1) syntactic correctness, i.e., options reflecting actual middleware setting and (2) semantic validity, i.e., eliminating inconsistent options.…”

“…This paper describes how CoSMIC's integrated MDD toolsuite supports correct-by-construction techniques for the design, development, configuration, deployment, and validation of DRE systems by (1) providing expressive power equivalent to CORBA 3.x Interface Definition Language (IDL) in the specification of component interfaces, including import/export of IDL to/from PICML models, (2) implementing constraints on the modeling of assemblies, e.g., ensuring certain ports of components are always connected, and allowing connections only between compatible ports, (3) allowing the configuration of the underlying middleware on which the artifacts will run, (4) modeling the deployment target at multiple levels of abstraction, including the hardware level (where nodes are modeled), the communication level (where the interconnects between the nodes are modeled), and the resource level (where resources used by the nodes are modeled), and (5) empirically evaluating the performance implications of configuring the underlying middleware in a particular configuration and deploying components on this middleware. Paper organization: The remainder of this paper is organized as follows: Section II uses an avionics mission computing application as a representative example to describe common challenges associated with applying QoS-enabled component middleware to DRE systems; Section III shows how the CoSMIC DSMLs and generative tools help resolve these challenges and analyzes the results of experiments that evaluate the application of CoSMIC to our avionics application; Section IV compares our work with other MDD tools that support DRE systems; and Section V presents concluding remarks and summarizes lessons learned.…”

DAnCE: A QoS-Enabled Component Deployment and Configuration Engine

“…In large-scale distributed real-time and embedded (DRE) systems, such as shipboard computing environments [2], inventory tracking systems [3], and intelligence, surveillance and reconnaissance systems [4], component middleware features can help make the software more flexible by separating (1) application functionality from (2) system lifecycle activities, such as component configuration and deployment. Conventional component middleware platforms, such as J2EE and .NET, is not well-suited for these types of DRE systems since they do The remainder of this paper is organized as follows: Section 2 provides an overview of inventory tracking system (ITS) case study that elicits many challenges of deploying large-scale DRE systems; Section 3 describes how we designed and applied DAnCE to resolve key challenges in our ITS case study; Section 4 compares our work with related efforts; and Section 5 presents concluding remarks.…”