Nianen Chen

Actors, roles and coordinators — a coordination model for open distributed and embedded systems

This paper presents a coordination model, the Actor, Role and Coordinator (ARC) model, to address three main concerns inherent in a pervasive Open Distributed and Embedded (ODE) system: dynamicity, scalability, and stringent QoS requirements. The model treats a pervasive ODE system as a composition of concurrent computation and coerced coordination. In particular, concurrent computation is modeled as Actors, while coerced coordination specifies the systems QoS requirements by mapping them to coordination constraints. The coordination constraints are transparently imposed on actors through message manipulations, which are carried out by the roles and coordinators. The coordinators are responsible for the coordination among roles, while the roles in our model provide abstractions for coordinated behaviors that may be shared by multiple actors and further assume local coordination responsibilities for the actors playing the roles. The roles behavior abstraction decouples the syntactic dependencies between the coordinators and the actors, thus shielding the coordinator layer from the dynamicity of underlying actors inherent in ODE systems. This paper also formally defines the role and coordinator behaviors and the composition of the actor computation model with the proposed coerced coordination model. Our formal study has shown that the ARC system is closed under composition and recursion.

Profit and penalty aware (PP-aware) scheduling for tasks with variable task execution time

As computing devices and the Internet technology advances, real-time on-line services are emerging. Different from traditional real-time applications for which the scheduling objective is to meet task deadlines, the optimization goal for on-line service systems is to maximize profit obtained through providing timely services. For this class of applications, there are two distinctive characteristics: (1) tasks, i.e., client requests, are associated with a pair of unimodal time functions, representing system accrued profit when a task is completed before its deadline, or accrued penalty if otherwise; and (2) requests execution times vary in a wide range. The paper presents a new scheduling algorithm, i.e., the Profit and Penalty aware (PP-aware) scheduling algorithm, with an objective to maximize systems total accrued profit. Our simulation results have empirically shown the advantages, in respect of system total accrued profit, of the proposed algorithm over other commonly used scheduling algorithms, such as Earliest Deadline First (EDF) and Utility Accrual (UA) algorithms.

Adaptive optimal checkpoint interval and its impact on system's overall quality in soft real-time applications

Soft real-time systems often have to consider both timing and probabilistic fault-tolerance requirements. When checkpointing techniques are used for fault tolerance purposes, the checkpointing frequency unyieldingly affects the systems overall quality measured by an integrated value of system QoS properties, such as availability, task execution time, and task deadline miss probability. In this paper, we first formally analyze the relationships between checkpoint interval and system availability, task execution time, and task deadline miss probability, respectively by considering a Poisson probabilistic fault model. We further define the systems overall quality as a weighted sum of these three QoS measures, from which an optimization problem is formulated to decide the checkpoint interval that maximizes systems overall quality. Also presented in the paper are a prototype implementation of a framework that allows adaptive checkpointing and a set of experiments executed upon the framework that further validate our analytical results.

Checkpoint Interval and System's Overall Quality for Message Logging-Based Rollback and Recovery in Distributed and Embedded Computing

In distributed environment, message logging based checkpointing and rollback recovery is a commonly used approach for providing distributed systems with fault tolerance and synchronized global states. Clearly, taking more frequent checkpointing reduces system recovery time in the presence of faults, and hence improves the system availability; however, more frequent checkpointing may also increase the probability for a task to miss its deadlines or prolong its execution time in faultfree scenarios. Hence, in distributed and real-time computing, the system’s overall quality must be measured by a set of aggregated criteria, such as availability, task execution time, and task deadline miss probability. In this paper, we take into account state synchronization costs in the checkpointing and rollback recovery scheme and quantitatively analyze the relationships between checkpoint intervals and these criteria. Based on the analytical results, we present an algorithm for finding an optimal checkpoint interval that maximizes system’s overall quality.

Using a role-based coordination model to achieve adaptive and quantifiable dependability for open distributed embedded systems

Although current middleware technologies, such as Fault-Tolerant CORBA (FT-CORBA) [1] and QuO [2], provide supports for a wide range of fault tolerance demands, they fall short for distributed embedded applications. In particular, to the authors’ knowledge, most of the current middleware technologies do not concurrently address the following critical issues while providing adaptive dependability in ODE systems: