A Robust Distributed Termination Detection Algorithm

IEEE Transactions on Software Engineering, 2000

He is currently working toward the Ph.D. degree in computer science at the State University of New York at Stony Brook. His research interests include distributed programming languages, operating systems, network protocols, and distributed algorithms.

… and Data Engineering, IEEE Transactions on, 2001

The paper presents a probabilistic performance analysis of a deadlock detection algorithm in distributed systems. Although there has been extensive study on deadlock detection algorithms in distributed systems, little attention has been paid to the study of the ...

IOSR Journal of Engineering, 2016

In a distributed system, a deadlock may occur when set of processes wait for resources from each other. A process involved in a deadlock waits indefinitely unless a special action is taken. Deadlock leaves the system into a blocking state with no process getting complete and also it reduces the throughput. In this paper a technique is presented that will improve the performance of Srinivasan distributed deadlock algorithm for multiple executions. In Srinivasan algorithm whenever multiple initiators invoke the algorithm one after the other same deadlock cycle and similar message transfers may be reported in more than one execution. So, in order to handle multiple executions an algorithm has been proposed which will not only reduce the number of message transfers but will also reduce the deadlock detection time. The proposed algorithm gives priority to different algorithm execution depending on their process id"s along with this it allows lower priority executions to continue so that a deadlock that are not directly reachable from the higher priority execution could be detected.

2004

This paper presents a discussion of the Kahn process network (PN) model and the challenges in distributing it onto multiple networked workstations. Compared to a non-distributed PN model, a distributed PN model is more scalable in terms of performance and has significantly lower cost. Although additional distribution-specific challenges exist, existing distributed systems algorithms can solve them.

Lecture Notes on Software Engineering, 2015

In this paper we have performed an analysis of existing deadlock detection algorithm for distributed systems and did some improvement on them. The algorithm proposed in this paper is an extension of previous works with an introduction of an identity set S in the probe initiation. The rate of dependency table clearance is determined by our algorithm, and how over-killing of processes can be avoided by using one additional data-structure is also shown by our algorithm. The algorithm shows how this data-structure should be updated with probe-messages with the identity set, and how it helps in determining which process should be selected as victim for resolving the deadlock, and which entries should be cleared in the dependency table after successful detection of a deadlock. The study indicates that in this algorithm, rate of probe initiation is a dominant factor in determining system performance and rules need to be framed for determining best value of it. 

A distributed algorithm for the detection of deadlocks in store-and-forward communication networks is presented. At first, we focus on a static environment and develop an efficient knot detection algorithm for general graphs. The knot detection algorithm uses at most O(n2 + m) messages and O(log (n)) bits of memory to detect all deadlocked nodes in the static network. Using the knot detection algorithm as a building block, a deadlock detection algorithm in a dynamic environment is developed. This algorithm has the following properties: It detects all the nodes which cause the deadlock. The algorithm is triggered only when there is a potential for deadlock and only those nodes which are potentially deadlocked perform the algorithm. The algorithm does not affect other processes at the nodes.

1994

with false deadlocks. 13, 16, 17, 18, 20, 211. There are four categories of deadlock algorithms [12]: path-pushing, edge-chasing, diffusing computations and global state detection. In this paper, we restrict our attention to edge-chasing (or probe-based) algorithms [5, 13, 16, 18, 20, 211. 'This work was supported in part by the Consortium for International Earth Science Information Networking.

Distributed systems deadlock is similar to single-processor system deadlock, but is worse. It is harder to avoid, prevent or detect and is harder to cure, when it is tracked down because all the relevant information is scattered over many machines. In some systems, such as distributed database systems, it can be extremely serious, so it is important to understand how it differs from ordinary deadlock and what can be done about it. Two important deadlock prevention algorithms in distributed systems are wait-die and wound-wait. Their problem is that they just attend to the time stamp of processes, but not priority of them. In a real operating system, attending to priority of processes is very important. The proposed improved algorithms are attending to both priority and time stamp of processes.

Distributed systems exhibit a high degree of resource and data sharing creating a state in which deadlocks might make their appearance. Since deadlock detection and resolution is one of the important concerns in distributed systems which lead to minimizing available resources, therefore instigating the system throughput decrease. Our proposed algorithm detects and resolves the multi-cycle deadlocks, whether the initiator is involved in the deadlock cycle directly or indirectly. Also the chance of phantom deadlock detection is minimized. This algorithm not only can manage the simultaneous execution of it but also detects the multi-cycle deadlocks in the distributed systems. Our algorithm introduces a modified probe and victim message structure. Moreover, no extra storage required to store probe message in each node which is known as memory overhead in the distributed systems.

Cata, 2003

Classification is not new in this middleware area. Many authors have suggests classifying by various communication techniques, protocols (session versus sessionless), the programming interface, [5], specific categories included database-oriented, virtual system, middle-tier, gateways, and Web-enabled [25], communication with programming execution and number of participants [32], or ten specific categories [21]