A Synchronous Scheduling Service for Distributed Real-Time Java

Concurrency and Computation

et al. 2011

Self Cite

SUMMARYService-based approach has been successfully applied to distributed environments, modelling them as pieces of functionality that exchange information by means of messages in order to achieve a common goal. The advantages of this approach can be also be applied to distributed real-time systems, increasing their flexibility and allowing the creation of new brand applications from existing services in the system. If this is an online process, then time-bounded composition algorithms are needed to not jeopardize the performance of the whole system. Different composition algorithms are studied and proposed, two of them optimal and another two based on heuristics. This paper presents an analytical solution that selects, depending on the structure of the application and on the load of the whole system, the most suitable composition algorithm to be executed in order to obtain a composed application in bounded time.

Section: Introductionmentioning

confidence: 99%

A hybrid approach for selecting service‐based real‐time composition algorithms in heterogeneous environments

Concurrency and Computation

et al. 2011

Self Cite

“…The second important result refers to the case with several handlers subscribed [i.e., D.E(4) and D.E (16)]. There is a linear relationship among the number of subscribers and the blocking experienced at the server, as suggested by the model proposed in Section III, which is supported by the empirical evidence.…”

Section: A Performance Patternsmentioning

confidence: 65%

A Dual Programming Model for Distributed Real-Time Java

2011

IEEE Trans. Ind. Inf.

More-and-more, the use of Java as a programming language for real-time applications is becoming of interest in industrial middleware. This is mainly because it offers advantages for the programmer as reduced deployment times, increased portability, and a number of APIs that may be integrated in large distributed applications. This paper contributes a dual communication model for distributed real-time Java applications. Current efforts in distributed real-time Java [e.g., The Distributed Real Time Specification for Java (DRTSJ)] are mainly focused on remote invocations (R.I.) and set aside other valuable approaches such as distributed events. The proposed model offers two choices for developing distributed real-time Java applications (one based on R.I., and another on distributed events), which may be used to develop applications. Both models include an additional support for asynchronism in communications, a feature that may speed up their communication performance. This paper includes a description for the two models, the changes that are required in the current API to accommodate them, and an empirical evaluation of their performance on a reference implementation.

“…NTP). For instance, in the Synchronous Scheduling Service (SSS) [16] of DREQUIEMI, S-choreographies execute a synchronization protocol between the master and its corresponding slaves. Depending on the quality of the infrastructure the jitter may be around ±100 µs in most cases (a performance similar to the second option).…”

Section: ) Time Synchrnonizationmentioning

confidence: 99%

Towards Propagation of Non-functional Information in Distributed Real-Time Java

2010 13th IEEE International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing

2010

Many real-time systems use preemptive priorities in their internals to guarantee certain real-time performance. This includes technologies that range from RTSJ (The RealTime Specification for Java) to middleware like Real-Time CORBA (Common Object Request Broker Architecture) which offers additional models and policies that blend client and server information. This decision makes easier the integration of real-time acceptance tests and dispatching policies in these kinds of infrastructures. In this paper, we analyze different tradeoffs that emerge from the definition of different propagation models for distributed real-time Java. The paper covers technological integration aspects, impact on interfaces, and other practical issues more related to the performance that this model offers to a real-time application.