Advances on P2P, Parallel, Grid, Cloud and Internet Computing

Frontmatter

Citation Estimation Method Using Abstracts of Research Data Articles: A Focus on Scientific Data

Abstract

With the trend of open science, efforts have been made to openly utilize research data. Considering the use of shared research data for interdisciplinary research, developing a researcher-friendly abstract writing method in different research fields is pertinent. In this study, we focus on abstracts from Scientific Data, a journal specializing in research data. We examine the influence of each part of speech on the utilization of research data through multiple regression analysis of the number of occurrences of the part of speech, the number of words and index-keywords in the abstract, and the number of citations research data article. Based on these results, we set the explanatory variables as the number of nouns, verbs, the other parts of speech, words, and index-keywords in the abstract. Thereafter, we developed a classifier to estimate the number of citations using machine learning. An analysis of the relationship between the number of citations and index keywords was also conducted.

Naoto Kai, Tomoki Yoshihisa, Toshiki Shimbaru, Hideto Yano, Hideyuki Tanushi

Can ChatGPT Outperform Other Language Models? An Experiment on Using ChatGPT for Entity Matching Versus Other Language Models

Abstract

In the era of rising AI, ChatGPT has become the most well-known chatbot, utilizing Large Language Models (LLMs), specifically GPT versions 3.5 or 4. It has been employed in various tasks, including text generation and text summarization. Entity Matching is one such task that requires the comparison of information in the records of interest. Traditionally, this work has relied on rule-based similarity measurements. However, in recent years, novel methods have emerged to combat this problem, including the use of word vectors, neural networks, and language models. In this paper, we will compare the results of the Entity Matching task by using ChatGPT and other language models, such as sentence-BERT and RoBERTa. Additionally, we will compare the results from zero-shot capable models like RoBERTa, DistilBERT, and BART. For the Blocking phase, we will use benchmark datasets that are available in ready-to-use formats, in conjunction with other novel blocking methods, if available.

Nontakan Nuntachit, Prompong Sugannasil

Blockchain and IoT Integration for Air Pollution Control

Abstract

Blockchain is a disruptive technology that enables untrusted parties to securely store and process data without the need of a centralised trusted party. In the context of data storage, blockchains offer a distributed database architecture that ensures data redundancy, fault tolerance and resistance to unauthorised alterations. The inherent immutability of blockchain technology guarantees the integrity of stored data, making them particularly suitable for applications where tamper-proof records are essential. These properties make blockchain the most natural infrastructure to build truly decentralised and secure IoT systems where IoT data can be shared and processed securely across untrusted entities without the need of a centralised trusted party. IoT and blockchain integration suffers of well known scalability issues. IoT systems can produce a high amount of data which can cause poor blockchain throughput and high fees. This paper presents an IoT application in the context of Industry 4.0 sustainability. In particular, a private blockchain is used in order to design and implement a system for air pollution control. Novel on-chain smart contracts for data storage and their aggregations are proposed. The approach has been implemented and evaluated in terms of its strengths, weaknesses and limitations with respect to the currently existing blockchain technologies.

Alessandro Bigiotti, Leonardo Mostarda, Alfredo Navarra

An AI-Based Support System for Left-Behind Children Detection in Vehicles

Abstract

In general, accidents involving school buses and parents leaving their infants in the car due to a lack of attention are increasing. Also, in-vehicle inspections are being neglected. In this paper, we focus on this problem and propose an AI-based support system for detecting left-behind children in vehicles. In the proposed system, we use YOLO which optimizes the hyperparameters when performing object detection. The proposed system can detect left-behind children by considering adults and dolls in a vehicle. Based on the evaluation results, we found that in Case #1, the accuracy was higher when the number of generations was 300. While, in Case #2, misidentification of adults was resolved by increasing the number of generations. Finally, in Case #3, some scenes were identified as dolls when were shown some body parts of the child.

Hibiki Tanaka, Naoki Tanaka, Shoei Sakano, Makoto Ikeda, Leonard Barolli

Design and Implementation of a Fuzzy-Based System and a Testbed for Selection of Radio Access Technologies in 5G Wireless Networks

Abstract

The transition to the 5th Generation (5G) mobile networks will rely on Ultra-Dense Heterogeneous Networks (UDHetNets). These networks introduce dense packed network configurations along with a diverse range of networks catering to user devices. The crucial challenge lies in establishing efficiently connections with an appropriate Radio Access Technology (RAT). This selection process involves the consideration of numerous parameters, which makes the problem NP-Hard. To address this challenge, in this research work, we design and implement a Fuzzy-based system and a testbed for RAT selection in 5G wireless networks. For the implementation, we consider three parameters: Coverage (CV), User Priority (UP), and Spectral Efficiency (SE). The output parameter is Radio Access Technology Decision Value (RDV).

Phudit Ampririt, Shunya Higashi, Ermioni Qafzezi, Makoto Ikeda, Keita Matsuo, Leonard Barolli

Classification of Steel Microstructure Image Using CNN

Abstract

The purpose of this study is to create a computerized system that can automatically evaluate microstructure images of steel materials, specifically focusing on ferrite, using a Convolutional Neural Network (CNN) model. Steel materials play a crucial role in our everyday lives, and their mechanical properties and reliability are determined by their microstructure, which is influenced by heat treatment and processing. It is essential to ensure the quality of steel, as problems can arise if the microstructure and mechanical properties are not adequately assessed before shipping. To accomplish this, the study involved preparing four different steel specimens with varying material properties and heating conditions, which were then photographed using a digital camera. The proposed CNN model was tested and validated to accurately classify the ferrite substances, and it was found that even a simple CNN structure could achieve high accuracy in image classification. The implementation of this system will alleviate the burden of human visual inspection. The paper provides detailed information on the preparation of the steel specimens, the method used to capture the images, the structure of the proposed CNN model, the experimental conditions, the validation methods employed, and the results obtained.

Shigeru Kato, Akiko Oshita, Tsubasa Kubo, Mitsuharu Todai

Enhancing Image Classification and Explainability with Object Isolation and Background Randomization

Abstract

In this paper, we address the problem of training a model on an image dataset that contains multiple objects that can introduce noise during the training of an image classification model. We propose a method for separating individual objects from the images and synthesizing these separated objects with a random background dataset to generate a new dataset in which each image contains a single, clearly defined object. We use the Attribution mask Compress-Semantic Input Sampling for Explanation (AC-SISE) method, a perturbation-based explainable artificial intelligence (XAI) model, to analyze the explainability of models trained on the previously generated dataset and the original dataset. The experimental results show that the ResNet50 model does not improve the explainability, but the VGG16 model improves the explainability somewhat.

Yongho Kim, Hyunhee Park

Energy-Efficient Role-Based Concurrency Control with Virtual Machines

Abstract

In order to realize energy-aware distributed applications, it is required to not only serialize conflicting transactions but also reduce the electric energy consumption of servers. In this paper, the EERO-VM (Energy-Efficient Role Ordering in Virtual Machine environment) scheduler is newly proposed to not only reduce the electric energy consumption of a server cluster equipped with virtual machines but also serialize transactions by significancy of subjects which issue the transactions. Evaluation results show a transaction issued by a more significant subject can be preferentially performed than transactions issued by less significant subjects in the EERO-VM scheduler. In addition, we show the electric energy consumption of a server cluster equipped with virtual machines can be more reduced in the EERO-VM scheduler than the RO (Role Ordering) scheduler proposed in our previous studies.

Tomoya Enokido, Dilawaer Duolikun, Makoto Takizawa

A Motion Analysis System for Pointing and Calling Considering Safety Checks for Soldering Work

Abstract

In Japan, pointing and calling are carried out to improve the safety, reduce human errors and prevent accidents. Via pointing and calling concept the safety is improved by pointing to the work object or tool and voicing the situation in order to predict the accidents during the work. In factories, soldering is part of the handicraft industry and one of the career options for people with disabilities. During soldering skill, it is necessary to memorize and repeat the same work, which takes a long time. Also, there are human errors caused by accidents and lack of safety checks or experience. In addition, ensuring worker safety requires continuous monitoring of worker motion, which is a significant burden for instructors. In this paper, we propose a motion analysis system for pointing and calling. The proposed system uses a depth camera to capture images of workers during pointing and calling. Also, the proposed system considers beforehand safety checks for soldering operations to prevent accidents and injuries. The experimental results of the pointing orientation show that the proposed system is effective for safety checks and can support beginners and people with disabilities to continue soldering work safely.

Kyohei Toyoshima, Chihiro Yukawa, Yuki Nagai, Yuma Yamashita, Tetsuya Oda, Leonard Barolli

Scalability Evaluation of Microservices Architecture for Banking Systems in Public Cloud

Abstract

The utilization of cloud computing and microservices architecture is of paramount importance in facilitating the advancement of information technology within the sector. In the present era, it is imperative for the banking sector to possess adequate preparedness to effectively manage fluctuations in transaction volumes during specific events. To assure the reliability of applications in always serving consumers, it is imperative for banking systems to possess scalability. This study aims to assess the scalability of the dual-layer security filtering approach in the microservices architecture designed for banking systems. The microservices architecture is implemented with the Kubernetes systems within the Google Cloud Platform. The dual-layer security architecture has been designed utilizing Web Application Firewall (WAF) and Next Generation Firewall (NGFW) technology. The scalability of the suggested architecture was assessed by the implementation of performance testing, which involved modeling certain scenarios and topologies that accurately reflect the transactional activities within banking systems. In the assessment of scalability, a series of 30 tests were performed, each involving the execution of 500 transactions per second (TPS) within a time frame of 100 s. The microservices design, as presented, demonstrates superior performance compared to the monolithic architecture in terms of vCPU utilization (48.81%), RAM utilization (8.22%), latency (19.09%), number of nodes (95%), and error rate (95.07%).

Amsal Maestro, Nico Surantha

Evaluation of Candidate Pair Generation Strategies in Entity Matching

Abstract

Entity Matching (EM) involves identifying and linking the given entities from various sources that pertain to identical real-world entities, serving as a fundamental element in data integration tasks. Such matching is assumed to play a pivotal role in enhancing the accuracy and reliability of downstream tasks in data analytics. Typically, the EM procedure comprises two essential stages: blocking and matching. This study focuses on the blocking phase, particularly the operations of candidate pair generation. Thus, the focal point of this study resides in the exploration of different techniques for generating candidate pairings from the sources during the blocking phase. The proposed work is evaluated by experiment on the benchmark datasets, which are DBLP-ACM and Amazon-GoogleProducts.

Kittayaporn Chantaranimi, Juggapong Natwichai

Evaluation of EV Performance by Battery Swapping Strategy

Abstract

According to the demand to reduce greenhouse gas, it is expected to deploy electric vehicles (EVs), widely. Especially, a type of battery swapping EV is highly expected because it can fulfill its electricity within significantly shorter time than a traditional plug-in EV. The battery swapping EV, however, is currently under development. So, it is also highly required to get examined. This paper reports the characteristic of battery swapping EV. Concretely, we discuss EV performance based on two strategies: 1) strategy deciding to swap battery only when an EV arrives at a destination PoI (Point of Interest), and, 2) strategy deciding to swap a battery whenever a battery gets lower level than a threshold. These strategies are called Check at PoI and Continuous Check, respectively. According to the results of evaluation, it is confirmed that the Continuous Check induces traffic jams around a battery swapping station (BSS) although the number of dead battery EVs is reduced. Hence, we propose a strategy to diversify the threshold for battery swapping, and clarify the advantage of the strategy.

Mayu Hatamoto, Tetsuya Shigeyasu

A Model of an Energy-Aware IoT

Abstract

Since a large number of devices and servers are interconnected, the IoT (Internet of Things) consumes a large volume of electric energy. In the FC (Fog Computing) model of the IoT, some processes of a sensor application for processing sensor data are executed on fog nodes and the other parts on servers. In our previous studies, the TBFC (Tree-Based FC) and FTBFC (Flexible TBFC) models are proposed, Here, application processes are replicated and distributed in fog nodes which are structured in a tree. In thee FTBFC model, the tree structure of fog nodes is changed to reduce the energy consumption. Here, the energy consumption of the changed tree is obtained by the simulation but it takes time to do the simulation, especially in a scalable tree. In this paper, we newly propose a mathematical model to estimate the total energy consumption of only nodes which are changed. By using the model, we discuss by which change operation on a target node the total energy consumed by the target node and changing nodes can be reduced.

Dilawaer Duolikun, Tomoya Enokido, Makoto Takizawa

Implementation and Optimization of Narrow-Band Internet of Things (NB-IoT) Nodes Coverage Using Doppler Effect Shift Chips

Abstract

Rapid growth of smart and other Internet of Things (IoT) network devices in urban areas has brought an increase in the demand for bandwidth, as well as efficient and accurate IoT coverage systems. Traditional methods such as conventional sensors or sensors that use wide bandwidth, or WiFi as sensors have limitations in terms of coverage, accuracy, and scalability. In this paper, we propose a novel approach for IoT node coverage using Doppler enabled networks. Doppler enabled networks by leveraging the principles of Doppler shifting enable real-time, energy efficient communication coverage, and monitoring of information flow. By deploying a network of Doppler sensors in convenient and well placed location, we can capture and relay comprehensive data on reduced amounts of data bloat and relatively narrow bandwidth. The relayed Doppler data is processed more easily using advanced signal processing and machine learning techniques to extract valuable coverage information including area, information density, and congestion patterns. Our experimental evaluation demonstrates the effectiveness of a Doppler network in accurately rendering coverage and providing real-time insights for IoT coverage systems. The proposed approach has the potential to significantly enhance coverage capabilities, leading to more efficient systems, reduced congestion, and improved safety on the network. The simulation was built and made over Cooja on Contiki.

Donald Elmazi, Fatjon Mehmeti, Elis Kulla

Performance Evaluation of FC-RDVM Router Replacement Method for Different Instances of WMNs Considering Subway Distribution: A Comparison Study Between UNDX-m and SPX Crossover Methods

Abstract

In this research work, we assess the performance of FC-RDVM router replacement method for small and middle scale WMNs considering Subway distribution of mesh clients. We carry out a comparison study of UNDX-m and SPX crossover methods. The simulation results indicate that for UNDX-m and SPX methods and both scales of WMNs, all mesh routers are connected so the size of giant component is maximized. Considering NCMC of UNDX-m and SPX methods, for small scale WMN, all mesh clients are covered. But in case of middle scale WMN, for SPX are covered all mesh clients, while for UNDX-m two mesh clients are not covered. The load balancing of SPX is better than UNDX-m. However, for middle scale WMN, the UNDX-m has slightly better load balancing than SPX.

Admir Barolli, Shinji Sakamoto, Leonard Barolli, Makoto Takizawa

An Integrated Energy Threshold and Priority Forwarding Approach to Improve Delivery Probability in Delay Tolerant Networks

Abstract

In this paper we propose an integrated energy threshold and priority forwarding strategy for Delay Tolerant Networks (DTNs). Simulations are conducted for Epidemic and Spray and Wait protocols and for different values of threshold to investigate the delivery probability and average remaining energy. Furthermore, the results of the proposed method are compared with another approach. The simulation results show the effectiveness of the proposed approach to improve delivery probability in DTNs.

Evjola Spaho, Orjola Jaupi, Frensi Muso, Esmeralda Shehu, Eri Kanani, Fatjon Zeqiraj

Development of a Piano Practice System for Beginners Using Mixed Reality Technology

Abstract

In this study, we have developed a piano practice system for beginners by fusing mixed reality and augmented reality technologies. This system allows piano learners to practice at their leisure. In this system, a virtual piano is superimposed on a real world piano, and users can practice the piano along with the virtual notes. The operability, visibility, functionality, relevance, effectiveness, and applicability of this system have been evaluated on 30 subjects. As a result, we were able to obtain high evaluations in many items.

Tomoyuki Ishida, Haruna Inutsuka

Development of a Traditional Craft Virtual Reality System for Mikawachi Ware/Hasami Ware

Abstract

In this study, we developed a virtual reality (VR) system that uses virtual reality technology to provide users with a simulated experience of Mikawachi ware (called Mikawachi yaki in Japanese) and Hasami ware (called Hasami yaki in Japanese). Wearing a head-mounted display (HMD) allows the user to experience space as if Mikawachi ware and Hasami ware ceramics were actually in front of them. In addition, the user can freely manipulate the virtual ceramics using a controller attached to the HMD. We conducted evaluation experiments with 24 subjects to evaluate this system on a five-point scale of operability, functionality, applicability, effectiveness, and relevance. Therefore, the subjects gave positive answers to all items, and we confirmed the superiority of this system.

Tomoyuki Ishida, Aya Deguchi

Development of an AR Application for Learning Traditional Patterns

Abstract

Traditional Japanese patterns, widely used in clothing, architecture, and crafts, are designs that convey various wishes expressed in the form of motifs which depict plants and living things. Moreover, these traditional patterns express the aesthetics of Japanese culture, which are highly appreciated both at home and abroad. In this study, we developed an augmented reality (AR) application that allows people around the world to learn the meaning and history behind these traditional patterns through AR contents by capturing images of the patterns on ceramics in real space.

Naho Kuriya, Tomoyuki Ishida

A Video Scene Segmentation Approach for Learner Monitoring

Abstract

Recently, for audio-visual communication are utilized many video conference systems and video on demand systems. Also, new live streaming and e-Learning systems are implemented and used for education and learning. However, during operation of these systems, the participants or learner do not watch the video because they may have also other tasks. In this paper, we discuss the video scene segmentation for monitoring a learner, where the video is divided the video scenes using a frame correlation matrix according to the learner’s behavior.

Kaoru Sugita

Portable Containerized MPI Application Using UCX Replacement Method

Abstract

In high-performance computing (HPC), using Linux containers can improve the convenience of system management and utilization, as well as facilitate the adoption of HPC DevOps and Clouds. However, utilizing containers with MPI applications poses some challenges in building portable MPI containers that can leverage various high-performance interconnect hardware available on different systems. A common solution is to replace the entire MPI stack in the container, but this may introduce compatibility issues between the host and container libraries. An alternative and more flexible solution is to inject host-shared libraries into the container and use them in conjunction with the container’s MPI stack, commonly referred to as a hybrid approach. In this paper, we propose a hybrid solution that uses the UCX communication framework as an intermediary between MPI and various high-end interconnects. This solution simplifies the container-building process while still achieving near-native communication performance. The flexibility of the proposed method was tested by performing a cross-version injection experiment. The results showed a good degree of forward/backward compatibility between versions, which can also refer to good portability as well. We evaluated the performance of our solution using communication micro-benchmarks and application benchmarks, and found that it achieved near-native communication performance. We also compared our solution with a similar solution using the Libfabric communication library, and confirmed that they have similar pros and cons with similar performance characteristics in most cases.

Sorawit Manatura, Kohei Ichikawa, Chantana Chantrapornchai, Chawanat Nakasan, Pattara Leelaprute, Arnon Rungsawang, Bundit Manaskasemsak

Selection of Reliable Best Peer in a Group of Peers by Using Correlation Coefficients and ns-3 Simulator

Abstract

In P2P systems, especially in mobile P2P, it is necessary to consider the mobility tolerance of nodes and the instability of the network. The goal of this research is to develop a system for selecting reliable and best peers in a P2P communication environment. Previous research has focused on improving the reliability of peers and evaluating the reliability of the network as a whole. In this study, we focused on the reliability evaluation of each peer by using correlation coefficients and ns-3 simulator. Foe evaluation, we consider three parameters: delay time, packet loss and throughput. The simulation results show that the proposed systems can choose the best reliable peers.

Yi Liu, Shinji Sakamoto, Leonard Barolli

An Intelligent Mixing System for Electric Guitar Using Fuzzy Control

Abstract

The Public Address (PA) system engineer performs mixing and adjust the sound in order that the live performance is comfortable for the performers and the audience. However, there are many cases where only one PA engineer is assigned and the burden of the PA engineer is very serious in the case of small live music clubs. Therefore, it is required to reduce the burden of PA engineers. In this paper, we propose an intelligent mixing system based on fuzzy control for electric guitar performance to simplify the mixing work of PA engineers and reduce their burden. The fuzzy control with low computational cost is applied to save computational resources and maintain real-time performance in volume control. From the experimental results, we found that the proposed system can automatically adjust the volume of the electric guitar during the performance.

Genki Moriya, Tetsuya Oda, Kyohei Toyoshima, Yuki Nagai, Sora Asada, Leonard Barolli

Construction of 1553B Bus Based on FPGA and Its Application

Abstract

A state control and data transform system was developed in FPGA based on 1553B data communication protocol between the command center and a remote terminal. In this system, two FPGA controlled 1553B Bus Protocol chips HI-6110 were configured as Bus Controller (BC) and Remote terminal (RT), respectively. FPGA drives BC to control BT by reading/writing Command and Status Register, and RT transfer data to BC by writing FIFO. Then message communication between BC and RT is achieved. The experimental results show that this method is practical. Moreover, this system can be extended to several RTs controlled by one BC to realize communication between command center and sub-devices in complex applications.

Zhang Jing

Evaluation of the Timber Internal Crack Using CNN

Abstract

When freshly harvested, cedar and cypress contain a high amount of moisture and must undergo a high-temperature drying process before we use them as building materials. However, a high-temperature drying process could cause internal cracks in the wood, and these defects reduce joint strength and buckling resistance. Therefore, human experts must visually evaluate the severity of cracks in the cross-section of timbers, which is highly labor-intensive and time-consuming. To address this issue, the authors have proposed to employ a convolutional neural network (CNN) to automatically evaluate the severity of cracks from cross-sectional images of timbers. Our previous study demonstrated that the proposed CNN could appropriately evaluate the crack severity. However, since the number of images was only 64, employing more images was required for further validation. Therefore, the authors added 140 images to validate the CNN in the present paper. This paper describes the experiment in detail and discusses the findings and future works in the conclusion part.

Renon Toyosaki, Shigeru Kato, Takashi Tamaki, Naoki Wada, Tomomichi Kagawa, Kazuki Shiogai, Hajime Nobuhara

Centroid Tuplet Loss for Person Re-Identification

Abstract

Person Re-Identification (Person Re-ID) is an important computer vision task in area surveillance, in which the goal is to match a person’s identity across different cameras or locations in videos or image sequences. To solve this task, Deep Metric Learning with the combination of different neural networks and metric losses such as Triplet Loss has become a common framework and achieved several remarkable results on benchmark datasets. However, Deep Metric Learning loss functions often depend on delicately sampling strategies for faster convergence and effective learning. These common sampling strategies usually rely on calculating embedding distances between samples in training datasets and selecting the most useful triplets or tuplets of images to consider, which makes these methods computationally expensive and may incur the risk of causing sample bias. Additionally, Triplet Loss also appears fragile to outliers and noisy labels. In this paper, we designed a centroid-based metric loss function, Centroid Tuplet Loss, which uses randomly selected mean centroid representations of classes in each mini-batch to achieve better retrieval performance. Experiments on two widely used Person Re-ID datasets, Market-1501 and CUHK03 dataset, demonstrates the effectiveness of our method over existing state-of-the-art methods.

Duc Viet Bui, Masao Kubo, Hiroshi Sato

Comparative Study of Metaheuristic Methods Inspired by the Prey House Mechanism

Abstract

Currently there is a set of techniques for solving complex problems known as NP-Hard, these problems have the characteristic that there are no exact procedures to solve them in polynomial time, so methods known as metaheuristics are used. Metaheuristic methods are methods that solve complex problems in a reasonable time and with useful solutions. There are different metaheuristics that can be used to solve problems, one category that has emerged recently are the algorithms based on swarm intelligence. Within these metahuristics it is possible to find methods inspired by the hunting behaviors of different animals in nature. This article evaluates the performance of three metaheuristic methods inspired by the hunting behavior of different animals. The metaheuristic methods considered in this work are: Cheetah Optimizer (CO) [1], Whale Optimization Algorithm (WOA) [2], and Gray Wolf Optimization (GWO) [3]. The performance was evaluated by means of 23 different functions divided into three types: unimodal, multimodal, and fixed dimension multimodal functions. The study shows a better exploitation capacity for the GWO algorithm while the CO algorithm has more capacity for exploration.

Jesus C. Carmona-Frausto, Adriana Mexicano-Santoyo, Pascual N. Montes-Dorantes, Jose A. Cervantes-Alvarez, Deysi Y. Alvarez-Vergara

A Tool for Solving the CVRP Problem by Applying the Tabu Search Algorithm

Abstract

The capacitated vehicle routing problem (CVRP) aims to optimize the delivery of products to different customers considering the capacity of the vehicles. This paper presents the implementation of a tool for findings routes that solve the CVRP problem. The proposed tool generates an initial solution by applying the Nearest Neighbor algorithm and then modifies the initial solution by applying two different versions of the Tabu Search algorithm. The differences between versions of the algorithm are in the generation of the neighborhood by interchanging two or four elements in the swapping process. To evaluate the results achieved by the algorithms, 27 instances acquired from the repository of CVRPLIB were tested, observing differences against the best-known value. Nonetheless, the tool allows to show graphically the initial instance, the initial and final solution routes and the performance of the algorithms with the aim of being a means for the analysis and improvement of algorithms.

A. Mexicano, J. C. Carmona, D. Y. Alvarez, P. N. Montes, S. Cervantes

A Comparative Study for Rate Allocation in Multi-source Systems with Same Rate Stream

Abstract

Demand for content distribution has increased significantly during last years. Content distribution systems are affected by scalability and dynamicity issues. This paper presents a comparative analytical study for rate allocation in a multi-source multicast framework based on the peer-to-peer (P2P) paradigm. This framework distributes multiple contents to all requesting peers exploiting full collaboration between the sources and the requesting peers. Each source distributes its own content blocks while additionally forwarding the content received from other sources. The requesting peers forward blocks of data received from the sources to the other peers. Our study compare the performance of two strategies to distribute same rate stream: Sources with independent rate allocation and sources with joint rate allocation. The rate allocation and redistribution assignments that maximize the overall throughput for both scenarios are analyzed. This study can help to decide which scenario is most suitable for the multi-source multicast framework.

Francisco de Asis López-Fuentes

Backmatter