Zhibo Wu

A Novel Aco-Based Multicast Path Algorithm In Hypercube Networks

Abstract This paper proposes an ACO (ant colony optimization)-based multicast path algorithm (AMPA) to reduce the multicast communication traffic in hypercube networks. According to the high regularity of hypercube networks, an optimized distributed algorithm (Opt-AMPA) for AMPA is proposed by clustering strategy, whose time cost can fill the requirements of the actual networks. The simulation results indicate that effectiveness of the algorithm can be improved drastically using Opt-AMPA.

A multi-cycle checkpointing protocol that ensures strict 1-rollback

In this paper, a checkpointing protocol based on loose synchronization is proposed. The protocol enables processes to take checkpoints at different frequencies so that each process can control its rollback distance. In traditional asynchronous and quasi-synchronous checkpointing protocols, the checkpoints that are not up-to-date may be used for recovery. As a result, the rollback distance is often difficult to control. In the proposed protocol, the checkpoint cycle of each process is dynamically adjusted using a pessimistic scheme so that strict 1-rollback is achieved; namely, one of the last two checkpoints of each process can be utilized for recovery.

A Comprehensive Approach of SA Based Software Deployment Reliability Estimation in Neural Networks

Software architecture (SA) has been widely advocated as an effective abstraction for modeling, implementing, and evolving complex software systems such as those in distributed, decentralized, heterogeneous and pervasive environments. We intend to investigate two problems related to the domain of those environments: software deployment and reliability. Though many approaches for Architecture-based reliability research, little work has been done in incorporating software deployment into SA based reliability estimation. This paper presents a new comprehensive approach of reliability estimation considering software deployment in neural networks. Additionally, we explain software component replica and different software component deployment architectures in neural networks.

Area Difference Based Recovery Information Placement for Mobile Computing Systems

In a mobile computing system, mobile hosts may move around cells, resulting in a considerable cost for locating and retrieving the recovery information, which is necessary for fault tolerance. To speed up the recovery, traditionally, recovery information is migrated according to the location of the mobile host. In this paper, a scheme for efficiently handling the recovery information is proposed. When a mobile host moves out of a certain range, only partial recovery information of the mobile host needs to be migrated to mobile support stations. It can avoid the unnecessary migration of recovery information. Moreover, the performance of the proposed scheme is evaluated and compared with the traditional movement based scheme.

An Effective Bi-Criteria and Contention Awareness Scheduling in Heterogeneous Distributed Systems

Abstract Fault-tolerant scheduling is an important issue for optimal heterogeneous distributed systems because of a wide range of resource failures. In this paper we propose a fault-tolerant scheduling heuristics for precedence task that is based on primary-backup replication scheme on a realistic platform model where communication contention is taken into account. We focus on a bi-criteria approach, where we aim at minimizing makespan (or the schedule length), and the other way take account into the failure probability of the application. We are able to let the user choose atrade-off between reliability maximization and makespan minimization. Major achievements include a low complexity and reduction of the number of additional communications by the replication and clustering mechanism. Simulation results show that in comparison to existing scheduling algorithms, the proposed scheduling algorithm improves the reliability and performance.

A New Method to Improve the Dependability of Key Nodes in Pervasive Computing Environment

Pervasive computing has been a hot point in researches for decades which integrates digital and physical devices in everyday life. The availability of key nodes which provide service is momentous. HA Cluster computing systems have become significantly practical for being server node. The availability of heartbeats, a main fault detection method, plays a momentous role in existing HA systems. However, the HA system may work on incorrect status when a fault exists in heartbeat links in the HA architecture. This paper presents a HA system with an independent monitor-based mechanism which monitors the leader nodes and manage the system to avoid the fault operation resulting from failures of heartbeat links. Its architecture is discussed, and the dependability of the system is analyzed by a continuous time Markov chain model; the dependability of a HA system and the HA system with an independent monitor module are compared. The results show that the proposed method is reasonable and will significantly improve the availability of the server node to fulfill the requirement in pervasive computing environment.

SA based software deployment reliability estimation: Problem space, challenges and strategies

Software architecture (SA) has been widely advocated as an effective abstraction for modeling, implementing, and evolving complex software systems such as those in distributed, decentralized, heterogeneous, mobile, and pervasive environments. We intend to investigate two problems related to the domain of those environments: software deployment and reliability. Software deployment is a post-production activity that is performed for or by the customer of a piece of software. Though many approaches for Architecture-based reliability research, little work has been done in incorporating software deployment into architecture-based reliability estimation. Thus, this paper describes which information constitute SA based software deployment reliability model and how to obtain available information for the model. The system reliability estimation at an architectural level includes software deployment, architectural style and component replica.

Dependency mining-based causal message logging

In fault-tolerant computing, the approach of causal message logging provides on-demand stable logging and enables the independent recovery of nodes. It imposes the requirement that the dependency between non-deterministic events needs to be known for nodes. The dependency information is disseminated through messages. A central problem in traditional causal message logging is that if the message logging progress of nodes cannot be effectively tracked, some redundant dependency information will be piggybacked on the messages. In this paper, a dependency mining based approach is proposed. It tries to detect the message logging progress of nodes only from the dependency between non-deterministic events, without needing to piggyback an additional dependency vector or dependency matrix on each message.

Communication-Based Prevention of Non-P-Pattern

An issue pertinent to the design of checkpointing protocols is how to improve the autonomy of checkpointing and keep computation loss under control. To address the problem, a time-based multi-cycle checkpointing protocol is proposed in this paper. In this protocol, processes are allowed to take checkpoints with desired checkpoint cycles. To enable recent checkpoints to be used to form a consistent global checkpoint, a communication-based checkpoint cycle adjustment approach is also proposed. In this approach, the checkpoint cycle adjustment of each process follows a P-pattern. Simulation results show that the rollback deviation of the proposed protocol can be well controlled under a low checkpointing overhead.