ON HIERARCHICAL CONFIGURATION OF DISTRIBUTED SYSTEMS ON MESH AND HYPERCUBE

2004 ◽  
Vol 15 (03) ◽  
pp. 517-534 ◽  
Author(s):  
DAJIN WANG ◽  
JIANNONG CAO
Keyword(s):  
Distributed Systems ◽  
Distributed Computing ◽  
Traffic Flow ◽  
System Performance ◽  
Interconnection Networks ◽  
Heuristic Algorithms ◽  
Network Routing ◽  
Distributed Monitoring ◽  
Hierarchical Partition ◽  
Distributed Processes

We study hierarchical configuration of distributed systems for achieving optimized system performance. A distributed system consists of a collection of local processes which are distributed over a network of processors, and work in a cooperative manner to fulfill various tasks. A hierarchical approach is to group and organize the distributed processes into a logical hierarchy of multiple levels, so as to coordinate the local computation/control activities to improve the overall system performance. It has been proposed as an effective way to solve various problems in distributed computing, such as distributed monitoring, resource scheduling, and network routing. The optimization problem considered in this paper is concerned with finding an optimal hierarchical partition of the processors, so that the total traffic flow over the network is minimized. The problem in its general form has been known to be NP-hard. Therefore, we just focus on distributed computing jobs which require collecting and processing information from all processors. By limiting levels of the hierarchy to two, we will establish the analytically optimal hierarchical configurations for two popular interconnection networks: mesh and hypercube. Based on analytical results, partitioning algorithms are proposed to achieve minimal communication cost (network traffic flow). We will also present and discuss heuristic algorithms for multiple-level hierarchical partitions.

AN EMPIRICAL STUDY OF HIERARCHICAL DIVISION FOR MESH-STRUCTURED NETWORKS

2007 ◽  
Vol 08 (03) ◽  
pp. 285-304
Author(s):  
DAJIN WANG
Keyword(s):  
Data Aggregation ◽  
System Performance ◽  
Network Routing ◽  
Distributed Monitoring ◽  
Network Loading ◽  
Distributed Processes ◽  
Multi Level ◽  
Simulation Results ◽  
Multiple Levels ◽  
Hypercube Networks

A parallel/distributed system consists of a collection of processes, which are distributed over a network of processors, and work in a cooperative manner to fulfill various tasks. A hierarchical approach is to group and organize the distributed processes into a logical hierarchy of multiple levels to achieve better system performance. It has been proposed as an effective way to solve various problems in distributed computing, such as distributed monitoring, resource scheduling, and network routing. In [21], we studied hierarchical configuration for mesh and hypercube networks to the end of achieving better system performance. In particular, we proposed theoretically optimal hierarchy for mesh and hypercube, so that the total traffic flow over the network is minimized. In this paper, we present the experimental results to establish the practical relevance of mesh hierarchy proposed in [21]. We simulated situations for multi-level division, real network loading scenarios, random data aggregation rates, and different division sizes other than derived in [21]. The simulation results not only show that the analytically obtained hierarchy works well for many realistic settings, but also offer some useful insights into the proposed hierarchy scheme.

Decentralization in Distributed Systems

2011 ◽  
pp. 386-399 ◽  
Author(s):  
Mustafizur Rahman ◽  
Rajiv Ranjan ◽  
Rajkumar Buyya
Keyword(s):  
Distributed Systems ◽  
Distributed Computing ◽  
System Performance ◽  
Distributed Computing Systems ◽  
Computing Systems ◽  
Cloud Systems

In recent years, decentralization in distributed computing systems, such as Grids and Clouds has been widely explored in order to improve system performance in terms of scalability and reliability. However, the decentralized nature of the system also raises some serious challenges. This chapter discusses the major challenges of designing and implementing decentralization in Grid and Cloud systems. It also presents a survey of some existing decentralized distributed systems and technologies regarding how these systems have addressed the challenges.

scholarly journals ДОСЯГНЕННЯ ЕФЕКТИВНОГО РОЗПОДІЛЕНОГО ПЛАНУВАННЯ ЗА ДОПОМОГОЮ ЧЕРГ ПОВІДОМЛЕНЬ У ХМАРІ ДЛЯ БАГАТОЗАДАЧНИХ ОБЧИСЛЕНЬ ТА ВИСОКОПРОДУКТИВНИХ ОБЧИСЛЕНЬ

2020 ◽  
Author(s):  
Старовойтенко Олексій Володимирович
Keyword(s):  
Distributed Systems ◽  
System Performance ◽  
State Of The Art ◽  
Computation Time ◽  
Cloud Services ◽  
Distributed Scheduling ◽  
Message Processing ◽  
Computational Problem ◽  
Control Units ◽  
And Control

Due to the growth of data and the number of computational tasks, it is necessary to ensure the required level of system performance. Performance can be achieved by scaling the system horizontally / vertically, but even increasing the amount of computing resources does not solve all the problems. For example, a complex computational problem should be decomposed into smaller subtasks, the computation time of which is much shorter. However, the number of such tasks may be constantly increasing, due to which the processing on the services is delayed or even certain messages will not be processed. In many cases, message processing should be coordinated, for example, message A should be processed only after messages B and C. Given the problems of processing a large number of subtasks, we aim in this work - to design a mechanism for effective distributed scheduling through message queues. As services we will choose cloud services Amazon Webservices such as Amazon EC2, SQS and DynamoDB. Our FlexQueue solution can compete with state-of-the-art systems such as Sparrow and MATRIX. Distributed systems are quite complex and require complex algorithms and control units, so the solution of this problem requires detailed research.

Security Issues in Distributed Computing System Models

2017 ◽  
pp. 211-259 ◽  
Author(s):  
Ghada Farouk Elkabbany ◽  
Mohamed Rasslan
Keyword(s):  
Distributed Systems ◽  
Distributed Computing ◽  
Computing System ◽  
Computing Environment ◽  
Distributed Computing Systems ◽  
Computing Systems ◽  
Research Directions ◽  
System Models ◽  
Security Issues ◽  
Systems Security

Distributed computing systems allow homogenous/heterogeneous computers and workstations to act as a computing environment. In this environment, users can uniformly access local and remote resources in order to run processes. Users are not aware of which computers their processes are running on. This might pose some complicated security problems. This chapter provides a security review of distributed systems. It begins with a survey about different and diverse definitions of distributed computing systems in the literature. Different systems are discussed with emphasize on the most recent. Finally, different aspects of distributed systems security and prominent research directions are explored.

scholarly journals The Wiener Index of Circulant Graphs

2014 ◽  
Vol 2014 ◽  
pp. 1-4 ◽  
Author(s):  
Houqing Zhou
Keyword(s):  
Distributed Computing ◽  
Interconnection Networks ◽  
Distance Matrix ◽  
Wiener Index ◽  
Important Class ◽  
Circulant Graphs ◽  
Parallel And Distributed Computing ◽  
Harary Index ◽  
Largest Eigenvalues ◽  
Reciprocal Distance Matrix

Circulant graphs are an important class of interconnection networks in parallel and distributed computing. In this paper, we discuss the relation of the Wiener index and the Harary index of circulant graphs and the largest eigenvalues of distance matrix and reciprocal distance matrix of circulants. We obtain the following consequence:W/λ=H/μ;2W/n=λ;2H/n=μ, whereW,Hdenote the Wiener index and the Harary index andλ,μdenote the largest eigenvalues of distance matrix and reciprocal distance matrix of circulant graphs, respectively. Moreover we also discuss the Wiener index of nonregular graphs with cut edges.

scholarly journals Information theory inspired optimization algorithm for efficient service orchestration in distributed systems

PLoS ONE ◽  
2021 ◽  
Vol 16 (1) ◽  
pp. e0242285
Author(s):  
Matheus Sant’Ana Lima
Keyword(s):  
Information Theory ◽  
Distributed Systems ◽  
Optimization Algorithm ◽  
Travelling Salesman Problem ◽  
Optimal Solution ◽  
Network Routing ◽  
Solution Quality ◽  
Single Node ◽  
Service Orchestration ◽  
Algorithmic Information

Distributed Systems architectures are becoming the standard computational model for processing and transportation of information, especially for Cloud Computing environments. The increase in demand for application processing and data management from enterprise and end-user workloads continues to move from a single-node client-server architecture to a distributed multitier design where data processing and transmission are segregated. Software development must considerer the orchestration required to provision its core components in order to deploy the services efficiently in many independent, loosely coupled—physically and virtually interconnected—data centers spread geographically, across the globe. This network routing challenge can be modeled as a variation of the Travelling Salesman Problem (TSP). This paper proposes a new optimization algorithm for optimum route selection using Algorithmic Information Theory. The Kelly criterion for a Shannon-Bernoulli process is used to generate a reliable quantitative algorithm to find a near optimal solution tour. The algorithm is then verified by comparing the results with benchmark heuristic solutions in 3 test cases. A statistical analysis is designed to measure the significance of the results between the algorithms and the entropy function can be derived from the distribution. The tested results shown an improvement in the solution quality by producing routes with smaller length and time requirements. The quality of the results proves the flexibility of the proposed algorithm for problems with different complexities without relying in nature-inspired models such as Genetic Algorithms, Ant Colony, Cross Entropy, Neural Networks, 2opt and Simulated Annealing. The proposed algorithm can be used by applications to deploy services across large cluster of nodes by making better decision in the route design. The findings in this paper unifies critical areas in Computer Science, Mathematics and Statistics that many researchers have not explored and provided a new interpretation that advances the understanding of the role of entropy in decision problems encoded in Turing Machines.

Sign in / Sign up

Export Citation Format

Share Document