Peerasak Wangsom

Peerasak Wangsom

Auto-generated profile, based on publicly available information

Peerasak Wangsom

Data Science and Engineering Laboratory, School of Information Technology, King Mongkut’s University of Technology Thonburi, Bangkok 10140, Thailand

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
5 December 2019 in IEEE Access
Verified
Journal Article
Open access
PWPeerasak Wangsom
Peerasak Wangsom
KLKittichai - Lavangnananda
Kittichai - Lavangnananda
Pascal BouvryPascal Bouvry
Pascal Bouvry
Due to serving several purposes simultaneously, running scientific workflows on dynamic environments such as cloud computing, has become multi-objective scheduling. Among these purposes, Cost and Makespan are probably the most two primitive objectives. Another critical factor in a large-scale scientific workflow is tremendous amount of data during execution. Therefore, this work also includes Data Movement as an additional objective as it has a major impact on network utilization and energy consumption in network equipment in cloud data center. In considering these three objectives, this work proposes a framework for scheduling solutions which combines a new nodes clustering technique in Directed Acyclic Graph (DAG) model known as Multilevel Dependent Node Clustering (MDNC) and the multi-objective optimization, Extreme Nondominated Sorting Genetic Algorithm-III (E-NSGA-III). E-NSGA-III is the recent extension of Nondominated Sorting Genetic Algorithm (NSGA-III). Five well-known scientific workflows, CyberShake, Epigenomics, LIGO, Montage, and SIPHT are selected as testbeds, while the commonly known Hypervolume is chosen as the performance metric. In this work, MDNC is also experimented with both NSGA-III. Comparison among three approaches, E-NAGA-III alone, E-NAGA-III with Peer-to-Peer clustering and E-NAGA-III with MDNC are carried out. The superiority of the proposed framework among them and its limitation are discussed.

1 December 2018
Verified
Conference Paper
Closed access
PWPeerasak Wangsom
Peerasak Wangsom
PBPascal Bouvry
Pascal Bouvry
KLKittichai Lavangnananda
Kittichai Lavangnananda
The execution of scientific workflows on dynamic environments such as cloud computing has become multi-objective scheduling in order to satisfy user demands from several perspectives. Among these objectives, Cost and Makespan are probably the most common. This research also includes Data Movement as an additional objective as it has significant effect to network utilization and energy consumption in network equipment in cloud data center. This paper proposes a multi-objective scheduling, Extreme Nondominated Sorting Genetic Algorithm (E-NSGA-III). It is an extension of the Nondominated Sorting Genetic Algorithm (NSGA-III). E-NSGA-III utilizes extreme solutions in the population generation module in order improve quality of solutions. Five well-known scientific workflows are selected as testbeds. Hypervolume and the Pareto front are chosen as the performance metrics. E-NSGA-III is evaluated by comparing its performance against the two previous versions (NSGA-II and NSGA-III). The comparison reveals that E-NSGA-III yields the best performance among them in multi-objective scheduling of the five scientific workflows.

1 January 2019
Verified
Conference Paper
Closed access
PWPeerasak Wangsom
Peerasak Wangsom
KLKittichai Lavangnananda
Kittichai Lavangnananda
PBPascal Bouvry
Pascal Bouvry
Running scientific workflows, often involves large-scale data-intensive applications, on cloud. The reduction of data movement is necessary because of the major impact to network utilization and energy consumption in network equipment in cloud data center. In order to reduce the Data Movement, this paper proposes Peer-to-peer clustering technique which can be applied to Directed Acyclic Graph (DAG) workflows. This research also includes Cost and Makespan which are probably the most common objectives for workflow scheduling. Due to dealing with the multi-objective optimization problem, Nondominated Sorting Genetic Algorithm-III (NSGA-III) is selected for finding solutions. Three well-known scientific workflows including Epigenomics, LIGO, and Montage are chosen as testbeds. Pareto front is adopted as the mean for visualization of the results while Hypervolume is selected as the performance metric. DAGs with Peer-to-peer clustering is evaluated by comparing its performance against the original DAG versions. The comparison reveals that DAGs with Peer-to-peer clustering enable NSGA-III to generate solutions having better Hypervolume and Pareto front in all three scientific workflows.

1 July 2017
Verified
Conference Paper
Closed access
PWPeerasak Wangsom
Peerasak Wangsom
KLKittichai Lavangnananda
Kittichai Lavangnananda
PBPascal Bouvry
Pascal Bouvry
The data locality is significant factor which has a direct impact on the performance of MapReduce framework. Several previous works have proposed alternative scheduling algorithms for improving the performance by increasing data locality. Nevertheless, their studies had focused the data locality on physical MapReduce cluster. As more and more deployment of MapReduce cluster have been on virtual environment, a more suitable evaluation of MapReduce cluster may be necessary. This study adopts a simulation based approach. Five scheduling algorithms were used for the simulation. WorkflowSim is extended by inclusion of three implemented modules to assess the new performance measure called `data locality ratio'. Comparison of their results reveals interesting findings. The proposed implementation can be used to assess `data locality ratio' and allows users prior to efficiently select and tune scheduler and system configurations suitable for an environment prior to its actual physical MapReduce deployment.