Master's degree works on grid technologies

Every year defenses of master's degrees and bachelor's diplomas are taking place at System Design department of IASA. Many of them are connected with grid technologies. The abstracts of the works of 2012, 2011, 2010, 2009, 2008 years are listed below.


2012

"Development of the accounting system for Ukrainian National Grid"
Svitlana Bevz

The purpose of work is development of the accounting system for Ukrainian National Grid taking into account features of construction of infrastructure and requirements to such systems, and development of effective mechanism for presentation of information about consumption of Grid Resources. Discovered and analyzed possibilities of development and deploying of mechanisms for realization of effective accounting within the Ukrainian National Grid's infrastructure for satisfaction of requirements in presentation of both general and detailed picture of the user's resource consumption. The tool for presentation of information of the resource consumption by users of Ukrainian National Grid was developed.


“The use of ontologies in data management in distributed systems”
Klim Burii

The aim is to study the use of ontologies for cloud brokers who work in distributed systems - clouds, to analyze of the current position of the cloud brokers and their vector of further development, to study models of service and event management architectures, to develop aggregator cloud services acting as a liaison between the consumer and provider services, which is based on a semantically structured data about providers.
The scientific novelty of the work: the possible implementation of cloud agents and their application in the cloud infrastructures were investigated, the logic of action set forth the broker based on semantically structured data was given.
The practical value of the work lies in the fact that the concept of the possible implementations of brokers was formulated and the trend from the grid to the clouds was studied. A system for semantic matching of cloud services on the basis of the criteria was developed.


"Relational data integration based on the semantic web technologies"
Eugene Gomolsky

Data integration involves combining data located in various sources and provide data to users in a unified manner. The aim is to study the architecture of data integration systems based on semantic web technologies and prototyping of data integration. The thesis presents theoretical information about the architecture of data integration systems, levels and challenges of data integration and semantic web technologies. The result is a research prototype data integration system based on semantic web technologies. The testing of the prototype and analysis of the results of the performance of the system was made.


"Research of decomposition problems in cloud computing"
Maksym Kreshchuk

The objective is to reserch the efficiency of decomposition problems in cloud computing. Analyzed the most popular cloud services. MapReduce paradigm in PaaS Google App Engine for decomposition problems was experimentally researched. API Task Queue in PaaS Google App Engine for decomposition problems was experimentally studied.


"Using the “cloud technologies” with an open source code"
Olga Muzhikova

The purpose of the work is a study of the models of services, of mechanisms of realization and standards of cloud calculations, an analysis of the existing cloud technologies with an open source code, and a practical use of analyzed technologies. The analysis of the cloud technologies Eucalyptus, OpenNebula, Google App Engine, Windows Azure, Stratos, OpenStack was made. During the analysis the main functional abilities, principles and mechanisms of the software deploying in the cloud environment were investigated. The Web applications were deployed in the cloud environment on the Google App Engine, Windows Azure technologies.


"GRID and Cloud Computing"
Oksana Petrova

Today there is parallel development of GRID and Cloud Computing used in modern e-Infrastructure. GRID has already become the usual form of construction of the distributed calculations for different industries of е-Science and business. GRID infrastructure is supported and developed by scientists, that know exactly, what they require for one or another scientific goal.
The main goal was to study the GRID and Cloud and the possibility of their integration. The study and comparison of their properties. Determination of safety, reliability separately in these technologies. As well as the impact of their association on the issue of safety and reliability. Generate possible scenarios combining Grid and cloud for further development.


“European Grid infrastructure (EGI) and the 7th Framework Programme grid projects”
Yaroslav Khramov

On December 22th 2011 a Memorandum of Understanding (MoU) between the EGI.eu and Base Coordination centre of UNG (BCC UNG) was signed. The agreement paved the way for the integration of UNG in the EGI, with the tremendous opportunities and advantages for UNG.
The aim was to study the European Grid infrastructure EGI and the 7th Framework Programme Grid projects, classify them by purpose and identification of the main EGI trends.
Analysis of the transition from EGEE to EGI and future directions of European grid infrastructure was provided. The integration analysis of UNG in EGI and the result of this integration for the end user was conducted. The 7th Framework Program grid projects analysis and their classification was provided. The 7th Framework Program grid projects in Ukraine was analyzed and the role of EGI in the formation of Framework programs was discovered. Recommendations for further development of UNG were provided.


“Research of possibility of realization graphical user interface for middleware ARC in Ukrainian national Grid”
Andrew Shukailo

The aim is to study of existent graphical user interfaces for Grid; to study their abilities and properties; to determine the necessary properties of GUI for middleware ARC in Ukrainian National Grid; to search for ways to create of it and argumentation of means of implementation of this graphical interface.


Bachelor's degree works

“Principles and architecture of autonomic computing”
Sergii Bykov

This diploma thesis is devoted to the principles of constructing large autonomous computational systems and their architecture. It considers structural features of the main system components, their interactions, conditions and limitations in developing mechanisms for self management. Thesis analyzes operational sequences of the management process, its main stages and MAPE-K control loop. It also reviews existing applications and solutions that use autonomous computing in military areas (DARPA), scientific research (NASA) and business (IBM, HP, Microsoft). This thesis also includes implementation of Parallel Virtual Machine (PVM) application which has elements of fault tolerance and adaptive behaviour.


"Grid technology in business"
Maksym Volodin

This thesis is dedicated to determination of the feasibility and features of Grid technology in business, identification of new business problems, which can be solved by using the Grid, to review and analyze of international Grid infrastructure projects and project implementation and application of Grid technologies in business, additional analysis of business decisions. The paper reviewed the middleware UMD, which is implemented in EGI, and analyzed existing Grid-business solutions, including Business Grid Computing products and IBM company products. The paper analysed the project «AgroGrid», that approvals Grid technologies into agribusiness. This paper made ​​recommendations on the use of Grid technologies in business, in relation to the needs and activities.


"Data integration and management in the semantic Grid"
Anna Doroshenko

Graduate work covers analysis of existing methods of data integration and management of data in semantic Grid. In this work the possible architectural approaches to implementing semantic Grid services including object-oriented and service-oriented were investigated, the advantages and disadvantages of each solution were given. The possibilities of existing software packages intermediate layer in terms of building a semantic Grid were covered. Different methods of data integration, described their features, implementation and benefits were analyzed.
A comparative analysis of grid-oriented integration of automated information was made. In this work OGSA-DAI software and its use as a basic platform for the integration of data in various scientific and engineering projects including medical research, geographic information systems, meteorology, transport, computer aided design, engineering and astronomy was studied. The ways of using this tool in the architecture of Semantic Grid were proposed.


"The database for the input data and results of GRID-system ALLTED"
Kseniia Karmaz

This bachelor degree work is devoted to the development of the database structure for storing both the input and output data required for the GRID-system ALLTED. The main results of the degree work are: the analysis of the existing database management systems based on the requirements which have been put forward in this degree work; developed database structure; practical implementation, which is expressed in the class writing using PHP programming language which is designed to communicate with the database; writing triggers for maintaining data integrity; experimental performance of the database, which is to be used in the GRID-system ALLTED.


"Study of the use of virtualization for high performance computing"
Ievgeniia Pavliuchyk

Degree work is devoted to research of features of the use of existing high-performance computing virtualization. The features of existing hypervisors and virtualization of parallel computing were examined. A comparative analysis of different technologies for their practical use, the conclusions about their strengths and weaknesses was made. Recommendations on the use of virtualization technology in the future for specific implementations are provided. Also a number of tests for high-performance computing, such as DGEMM, Linpack, STREAM, etc. was analyzed.


"Knowledge representation for the semantic grid"
Hatkova Irina

This diploma is devoted to analysis of existing tools for knowledge representation for the semantic grid. This work also analyzes the current project in the semantic grid. The comparative analysis of the latest grids was made with their categorization in four directions. The research and analysis being conducted can be recommented for futher work with semantic projects in the direction of system design.


"Instruments of distributed computing in system ROOT-CERN"
Volodymyr Yurchenko

This work is devoted to research of instruments for distributed computing that are based on ROOT system. An analysis of features and capabilities of software tools that are utilized for distribution of computing with using elements, methods and paradigms of ROOT system was made.
In order to investigate effectivity of mentioned instruments, a software package of test programs and examples was developed. On the basis of these investigations the evaluations of time parameters for execution of tasks for distribution of computing were provided.
The results of this work can be used as a tool for distribution of resource-demanding computations in local multi-core computers, cluster machines and in Grid environment as well.


2011

"Web service access to grid systems for modeling"
Zeleniuk Oleksii

The purpose of the research is the development of features of service-oriented architecture, analysis of existing web services and analysis platforms for their development on the example of building Web services access to the grid.
The research contains analysis and comparison of the basic approaches for creating Web services in the concept of a particular Web service; description of the features of registration and search (as well as automated search) of Web services; analysis of existing specifications, development and integration of Web services; the results of evaluating of basic research platforms and tools to quickly and easily create Web services and their subsequent registration and life cycle support; analysis of major libraries that provide access to the grid using Web services; description of the placement, development and testing of Web service.
Scientific novelty of the work is that: analyzed and identified different concepts of web service access to the grid; analyzed the main advantages and disadvantages of the existing platforms of web services; based on the analysis developed web service that takes into account the immediate shortcomings in the work of grid (in the context of integration with different middleware).
Work contains 92 pages, 19 pictures, 14 sources.


“Semantic Grid for science and engineering”
Ismagilov Evgeniy

The purpose of the master work is to review and analyze the use of Semantic Grid for science and engineering. The following tasks are solved: review and analysis of ways to use the Semantic Grid for science and engineering work on resource discovery, information retrieval and data integration, recommendations for using each method.
Scientific novelty of the work lies in the analysis of the use of Semantic Grid for science and engineering, as this technology is a powerful tool for organizing data integration, resource discovery and search of information. Material on the basic goals of the Semantic Grid: data integration, resource discovery and semantic search  was structured and integrated.
Work on 92 pages contains 18 illustrations. By preparation of work the literature from 22 resources were used.


"Grid and Cloud technologies"
Olenovych Evhenii

The objective is to study cloud technologies, trends and prospects in the future
transition to the virtualization of grid resources cluster network to network virtual
servers cloudy.
The work contains: analysis of the basic principles of the basic elements of cloud technologies; description of the features of security, privacy, data storage users and how
they connect to the clouds; analysis of existing types of cloud technologies; description of the basic principles of the Grid; research developments; identification of future direction of grid and cloud technologies.
The work contains 77 p.. 18 fig., 1 tab., 17 sources.


“Grid system for the Earth sciences”
Olga Bilobrovа

The purpose of the Master work is research of features of Grid systems and their application for the Earth sciences, by the study of existent initiatives and projects in the area of the Earth sciences and application of Grid systems in world practice and estimation of national situation of introduction of Grid technologies for recommendations in relation to introduction of Grid systems for the Earth sciences in Ukraine.
Attained results:
1. The modern state and progress of different projects of Grid trends is analysed, constrained with the use of geoinformation data, and also certainly basic features of their development and future applications for the Earth sciences.   
2. International projects are considered in the area of geoinformation researches and the Earth sciences, most which are oriented to such tasks, as global changes of climate, research of space, study of bottom of ocean and terrene, design of seismic processes.
3. The conducted analysis of constituents of Grid systems is for the Earth sciences, the row of general lines, which was observed for all considered Grid systems and row of features, inherent only some of them, was selected as a result.
4. The analysed situation of introduction of Grid technologies in Ukraine. As a result of consideration of reports in relation to realization of the «Government having a special purpose scientific and technical program of introduction and application of Grid technologies on 2009 – 2013 years» a few projects were selected from development of Grid- of technologies for the Earth sciences in Ukraine, they are analysed basic directions of development.
5. Recommendations are formulated in relation to application of Grid- of technologies for the Earth sciences in Ukraine.
Work contains 68 pages, 18 picture, 35 sources.


"Semantic grid for molecular research"
Osmanova Tetyana

The purpose of the given work is to research and analyze existing models of semantic Grid in molecular science. The work presents theoretical information on the existing Grid systems of the molecular science. The different types of ontologies describe metadata chemical structures in these models, effective ways to identify resources in the Semantic Grid, namely the process of finding resources that meet the needs of users.
The analysis of existing semantic Grid for molecular studies, including the purpose of each and list the tasks they perform with the current requirements of Semantic Grid. These requirements include the interoperability, scalability, decentralization and dynamism.
Results of investigations are to require the Semantic Grid in Computational Chemistry, consistent with the molecular studies in Ukraine and conformity to international standards and existing models, such as: Molecular GRID (World Wide Molecular Matrix (WWMM)), and a draft project CombeChem and Collaboratory for Multi-scale Chemical Science (CMCS).
The innovation of the executed work consists in analyzing and setting for the basic requirements for application of semantic grid for molecular studies. Now the world is very widely used and developed technology of Semantic Grid in many areas of science, but unfortunately does not have the rapid application of molecular science. Therefore, this paper can serve as a starting point in researching the development of semantic grid for molecular research in Ukraine.
Work contains 103 pages, 12 picture, 26 sources.


"Semantic grid infrastructure for biomedical applications"
Sergeyeva Larysa

The purpose of the given work is to research and analyze existing semantic Grid infrastructures in the biomedical field. Biomedical research requires integrated access and management of various information sources, analysis methods and applications. Research work also requires to investigate, characterize and obtain access to new sources of information. To ensure the integration, access, and research an appropriate Grid infrastructure is needed to create, which would better respond to the specifics of biomedical data and all biomedicine.
Results of the carried out research is to identify infrastructure requirements for biomedical Grid, taking into account the needs of the area that is rapidly changing, evolving and growing, as well as compliance with current standards of the world of biomedicine and existing infrastructures Grid for biomedical, analysis of the last, including: Grid Infrastructure for Research in biomedicine, caGrid, Semantic Grid infrastructure for integrated access and analysis of multilevel biomedical data for cancer research, ACGT, Semantic Grid infrastructure, SEAGRIN and project MIAKT (Medical Imaging with Advanced Knowledge Technologies).
An analysis of existing Grid infrastructures, taking into account the purpose of each and a list of tasks that they perform in accordance with the requirements of complex, modular nature of biomedicine are represented in this work.
The practical value of the work is to obtain a theoretical basis for the construction of Grid infrastructure for applications in biomedicine. The results can be used as a recommendation to build Ukrainian Grid infrastructure for applications in biomedicine.
Work on 140 sheets contains 13 illustrations. By preparation of work the literature from 15 different sources was used.


"Simulation of ukrainian grid-infrastructure"
Shinkaryuk Dmytro

The purpose of work is simulation of Ukrainian Grid-system using a GridSim toolkit, simulation of  loading , investigation of   Ukrainian grid-infrastructure and detection of "weaknesses" of  the system.
Obtained results:
1. Ukrainian grid-infrastructure was analyzed  and modeled using the latest data about configuration of computational resourses, network characteristics and processing elements evaluation.
2. Model of a network was created andits performance was tested under different loads: overloaded channels, site of an emerging large queues and other bottlenecks.
3. Task-grouping algorithm was experimentally investigated with different values of granularity time.
4. Comparison of time-shared and space-shared scheduling strategies was done.
As a result of researches three publications are done: theses of two presentations on International conference of SAIT -2010 and SAIT-2011 and the   article in an electronic magazine «System sciences and cybernetics». - 2010. - №1.
The results of the executed work  were passed in the Base co-ordinating grid- center (ITF NANU) in quality recommendations for the use in future Programs of introduction of Grid-technologies in Ukraine.
The work contains 125 pages, 43 images, 20 sources.


«Semantic data transfer in computational Grid»
Shugaiev Volodymyr

The goal of this work is overview and analys of semantic data transfer in computing Grid.
1. Analyzed methods of semantic data integration.
2. Analyzed methods of organization of semantic data exchange in computing Grid.
3. The recommendations given for further usage of approaches to semantic Grid-services development.
Work consists of 102 pages, includes 10 illustrations, uses 43 different sources.


2010

“Data and metadata in Semantic Grid"
Drozd Oleksii

The object of this research is semantic metadata in Grid environment: annotations, ontologies, semantic links and data and resources, that they describe. The subjects of research are properties of metadata and ontology, methods of their use. The aims of this work are systematization and generalization of approaches of dealing with semantic metadata, research of problems and “pitfalls” that can be faced on the way of development of methods and facilities to work with metadata in particular, and integral and effective infrastructure of Semantic Grid in the whole; research of existent implementations and architectures, determination of further vectors of development of subject.
The scientific novelty of this work lays in systematization and all-round consideration of topic of semantic metadata in Grid. Most researches on this question are concentrated round the separate aspects of life cycle or use of metadata, or pay attention to separate implementations and developments in this sphere. This work can serve as a starting point in research-and-development of architecture  of Semantic Grid in Ukraine.
The report based on work results was made at the conference «System Analysis and Information Technologies 2010» with topic «Data and metadata in Semantic Grid». The article «Organization of metadata in Semantic Grid» in is published in «Systems and cybernetics».
Work is on 118 pages, contains 24 illustrations and 2 tables, 23 different sources were used.


“Organization of Grid knowledge portals based on ontologies"
Khondar Vadim

The task of Grid technologies is to make the processing and storage of large volumes of data possible. The ways of combination of two technologies - Semantic Web and Grid - are of great interest. One of their common points is a portal of access to the resources. Therefore a problem of determination of ways of application of semantic technologies for expansion of the functionality of portal is urgent.
Development of semantic Grid portal is a part of the Government program «Creation of national Grid infrastructure for science and education».
The aims of work are: an exposure of parts of existent infrastructure that can be improved due to semantic technologies, analysis of necessary utilities needed for development of semantically enriched web applications, research of possible problems in this process and the requirements to the portal components.
The object of this research is semantic technologies for the Grid infrastructure on the whole and in a web-portal in particular: elements, data and processes that can be described semantically with the aim of improvement of work with them. The subjects of research are methods of use of semantic metadata, and also technological properties that can lead to practical benefits in case of their application.
The scientific novelty of this work consists of: systematization of scenarios of use of semantic technologies and development process of semantic applications. Existent researches  are concentrated on the separate aspects of work and application of semantics that complicates perception of overall picture. This work gives an opportunity to get all-round view of the basic moments of semantic technologies.
The report based on work materials was made at the conference «System Analysis and Information Technologies 2010» on topic «Ontology development tools». The article «Semantic technologies in Grid» is published in «Systems and cybernetics».
Work is on 124 pages, contains 33 illustrations and 1 table, 33 different sources were used.


“Modeling of Grid infrastructure with the GridSim package"
Kutz Vitalii (specialist diploma)

The aim of work is the development of the program, by means of which it is possible to design and analyze Grid infrastructures. To achieve the marked goal the next tasks were set and solved:
An analysis of Grid infrastructure of Ukraine, descriptions of its computing nodes, bandwidth of channels and other.
Dependency of modeling time and tasks execution time on the sizes of tasks and their number when using this infrastructure.
Comparative analysis of computing resources.
Queues. Mean task queue time, average and maximum number of tasks in a queue.
Work consists of introduction, four parts, conclusion, list of literature and one appendix, that contains the source code of the program developed with Java language in the NetBeans IDE 6.7.1. environment.
Project consists of 100 pages, including 79 pages of basic text, 29 pictures, 17 tables.


“Modeling of resource broker for Ukrainian Grid infrastructure with SimGrid package"
Yeliseyev Oleksii (specialist diploma)

In accordance with the topic of the work the goal is the modeling of the broker for allocation of resources of Grid infrastructure with the help of SimGrid package; analysis of modeling results, review of possibilities of SimGrid modeling package, advantages and shortcomings compared to other modeling packages. A broker is responsible for the discovery of resources, their selection, binding of software and data to the selected resources, start of calculations, adaptation to the changes in the system.  The main results of the work:
1. Possibilities of the modeling program of SimGrid are analyzed.
2. Important computer experiments on the modeling of parameters of the national Grid infrastructure with the different task scheduling heuristics.
3. The capacity of the developed software product is tested on the Ukrainian Grid infrastructure. Correctness of the proposed approach is confirmed by the results quality.
4. The state of computing resources of the Ukrainian Grid infrastructure is analyzed.
5. The analysis of dependencies of modeling time and task execution time on tasks size and quantity was conducted.
6. Mean queue time of tasks, average and maximum number of tasks in a queue were estimated.
7.The problem of optimization of distribution of tasks in Grid infrastructure was considered. It is shown that when tasks are arranged by width, task distribution gets better.
The experiments show that the offered broker system can carry out effective scheduling, even if it uses relatively simple algorithms of distribution.
Project consists of 113 pages, including 64 pages of basic text, 24 images, 18 tables.


 

2009

Software tools for simulation of Grid systems and estimation of their productivity
Igor Aronov

The object of research is most widely-used and effective program tools for simulation Grid computing environments, which are used for investigation of scheduling strategies, that are the key aspect of performance of the computational Grid.
Among the primary objectives of this work are: review, analysis and comparison of existing world simulation tools, which are focused on evaluation of scheduling mechanics in the computational Grid and estimation of their usability for the modeling behaviors of different types of the Grid and solving different kinds of tasks with different requirements of users.
The novelty of work consists in the analysis of the newest versions of Grid simulation tools and creation on the basis of this analysis of different classifications which will be very helpful for Grid researchers and users for optimal electing the necessary simulation tool for the decision of own
It should be noted that the accent done in this paper on simulation is based on the pulled assertion out, that in general simulation is the most effective approach for scheduling research and allocation of Grid resources, that are basic factors of influence on performance of the computational Grid. This is due to other known approaches for similar research, that are analytical computations and real experiments on the real computing platforms have the substantial failings, which interfere with the receipt of the controllable and repeatable results of scheduling experiments, that is the important requirement of high-quality research. At the same time, in this work marked, that unlike other researches communities (such as organization of networks, computer architecture and etc) in a community of research of Grid scheduling yet and until now an unique standard in relation to the simulation is not accepted, that imposes substantial difficulties on rapid progress of this branch of Grid computing. This absence of standardization could be caused by relative youth of this direction of computer science and wide range of application domain of the modern Grid.
The exactly noted feature stipulates substantial differences in architecture and realization of the simulation tools considered in work and does impossible their common direct comparison within the one test task. That is why the comparative analysis of these simulation tools is possible only in the context of mention of specifics of their use.
The results of this work and formulated propositions in relation to the ways of optimum usage of these program tools can be used by the wide group of users and researchers of the computation Grid/
Continuation of work is foreseen in the following directions: electing and implementation of the one of these simulation tools for estimation of performance of national grid infrastructure project.The thesis consists of 167 pages, 78 figures, 10 tables, 102 references.


The system of storage of data trajectories of molecular dynamics of albumens and nucleic acids in the MolDynGrid virtual lab
Glib Bochkarev

The object of research are biomolecular projects that use Grid to conveniently store, retrieve and
analyse biomolecular simulation data.
he purpose of work - review of existing biomedical Grid projects, analysis of a distributed  storage system trajectories of molecular dynamics of proteins to determine its functional characteristics,merits and feasibility of certain conditions, the creation and implementation of a studyl block in the Internet portal of virtual laboratory.
For many years in the biochemical research community, protein simulation data have been stored locally where they were generated; this severely limits data sharing within the biochemical community.
The solution to this problem is within the concept of grid computing as the Grid promotes the sharing of computational and storage resources. the scientific research world. In the concept of data grid, scientists are allowed access to geographically remote storage resources across a network in a uniform and efficient manner. With this data sharing capacity, biochemists are able to perform cross-simulation comparison to explore fully the functional dynamics of biomolecular simulations. Analyzed projects seeks to exploit the potential of the Grid on biomolecular simulation data and tries to solve some of the problems that hamper comparative analysis.
The novelty of work is in the analysis of developing biomedical projects which allow access to geographically remote trajectories in a coordinated way and provides a set of uniform analysis tools to facilitate comparative analysis on different simulation data types.
Projects observes in this work aim to provide a data abstraction layer from the user where scientists can concentrate on research without the concerns of ‘how’ and ‘where’ their data is stored. One of the challenges is dealing with the simulation data which comes in varying formats, as these analysis scripts are normally written to process a single data format, thus causing great incovenience in the comparison between simulation data of dissimilar formats. Also such problems as absence of desirable level of automation of the environment of a data storage and management of system of flat files are solved.
Theoretical results of this work was the description of ways more efficient sharing and post processing of biomolecular simulation data within the research community.
In practice the studing block within the limits of an Internet portal of the virtual laboratory
MolDynGrid, opening to users many additional possibilities has been created.
The thesis consists of 95 pages, 57 figures, 3 tables. By preparation of work the literature from 14 different sources was used.


Grid technologies and their application for the calculation of slow molecular dynamics of albumens
Vladimir Shapran

The ways of Ukrainian Grid-segment calculation power usage for increasing of molecular dynamics calculation efficiency are the object of current work research. Also aspects of virtual laboratories usage for resolving bioinformatics problems are also investigate
Primary objective of this work is creation of effective infrastructure for in silico molecular dynamics calculation of biological macromolecules (proteins, nucleic acids and their systems) in water-ionic environment in 100 nanoseconds time interva
The work novelty consists in usage of modern virtual laboratories creation methods. Created Internet-portal of MolDynGrid virtual laboratory is the first specialized portal for molecular dynamics calculations and doesn’t have analogues in Ukraine.
MolDynGrid internet-portal technical realization issues are resolving in this work. Including database, client application and required web-services creation tasks. MolDynGrid virtual laboratory internet-portal was created. It consists of calculation block, calculated molecular dynamics trajectories database, analytical block and educational block.
Educational block in internet-portal was created. It consists of educational materials block, library, students block and forum. Educational block was realized as rich internet application. Next technologies were used for its creation: Adobe Flex, WebORB and Java 2 SE. Relational database for user data and other information storing was created using MySQL.
Created portal was deployed on Institute of molecular biology and genetics server and is accessible via Internet.
The thesis consists of 85 pages, 12 figures, 1 table and 12 references.


Method of replics exchange under calculations of molecular dynamics of tiposil-тРНК sinteties in the Grid
Michael Tal'ko

The objects of research are replica exchange molecular dynamics methods, and their use for calculation of molecular dynamics of tyrosyl-tRNA synthetase in Grid-environment. Among the primary objectives of this work are: review, analysis and comparison of the most widely-used and effective replica exchange methods for calculation of molecular dynamics of proteins in Grid-environment, analysis of results, and analysis of their usability in Ukrainian Grid- environment.
The novelty of work consists in the analysis of the newest versions of replica exchange methods and application them to investigation of molecular dynamics of protein in Ukrainian Grid-environment on base of virtual organization.
Theoretical data on methods of an exchange by remarks for calculation of molecular dynamics of proteins have been presented in this work, also have been recommended value of parametres, the calculation of molecular dynamics tyrosyl-tRNA synthetase has done, the analysis of the received results in comparison with a traditional method of molecular dynamics is carried out, the analysis of the general algorithm of calculation of molecular dynamics is carried out and software for automation of its stages are created.
In this work the basic replica exchange methods for calculation of molecular dynamics - REMD, MREMD, REST, TEE-REX have been considered. Each of the specified methods has the advantages and lacks. The best for calculation in Grid-environment as computing among and for experience finding method MREMD is recognized, however in software product GROMACS its realization is absent also the objective answer concerning its efficiency in the given job to give it was not possible. However related method REMD has finished the suitability and a place among the best methods of an exchange of remarks. It is considered the most perspective method TEE-REX, considering a physical substantiation and efficiency of use of computing capacities. However use of method TEE-REX requires great volume of knowledge, abilities and experience in such areas of a science as biophysics and biochemistry as the choice of parametres for calculations is a determinative for reception of comprehensible results, and should consider features of selected model of research.
Job on application and adjustment of procedure of calculations of molecular dynamics by replica exchange methods such as REMD and TEE-REX in Grid-environment on the basis of virtual organization MolDynGrid with use of computing capacities of cluster NTUU “KPI” has been spent.
In the given work a number of software which automate following stages of calculation of molecular dynamics has been created: processing of files from protein data bank, creation of configuration files for the project of calculation of molecular dynamics by replica exchange methods, creation of the project for replica exchange method from a set of molecular dynamics projects. Automation of the specified stages releases scientists from great volume of routine job.
It has been practically applied two replica exchange methods - REMD and TEE-REX - to research of molecular dynamics of linker of tyrosyl-tRNA synthetase , in an interval amino acid bond chain tyrosyn341- isoleucine364. And the carried out primary comparative analysis with the results received by a traditional method of molecular dynamics which testifies to efficiency and expediency of use of replica exchange methods for calculation of molecular dynamics both their further studying and introduction in research process protein folding.
The thesis consists of 100 pages, 24 figures, 1 table, 11 references.


 

2008

'Grid computing environments'
Bogdan Bulakh

The object for research is Grid environment for the National Grid infrastructure, in particular, its main components: existing resources, middleware for their virtualization, special solutions for work and programming in Grid environment (programming tools, client programs for multiple tasks, Grid portals).
The purpose of work is analysis and comparison of existing solutions implemented in European and world Grid environment projects, formulations of propositions on multiple Grid environment components for the National Grid infrastructure.
Propositions on general Grid environment architecture were formulated, research on existing variants of its components was held, this being the basis for proposed solutions being based on the best existing approaches, the own realization variants were developed. The particular attention in the work was paid to the problem of maximally possible use of the best completed program realizations with open code because of the substantial economy of time and means for their implementation and maintenance.
The proposed Grid computational environment architecture for the National Grid infrastructure is based on existing and future computational, tool resources and data stores in scientific and educational centers of Ukraine and research-educational URAN net. It is proposed to install glite 3 middleware on recourses. In work, the main ways of providing its functional interoperability with the other software, in particular, Nordugrid, are considered. Tool resources are proposed to be realized through the development of specific Web-services for management and monitoring, and also the necessary additional software. User access to grid environment is proposed to be organized on the basis of Grid portal with the use of Gridsphere container and portlet libraries Gridportlets/vine.
The work results can be applied as recommendations on building the Grid environment at practical realization of National Grid infrastructure creating project. The course of work is looked in the following directions: detailed theoretical and practical development of separate Grid environment components, research of the issues which will rise at practical realization of applications given in the work.
The volume of work is 139 printed pages; it contains 14 illustrations, 14 tables, 77 foreign and native sources on the link list.


'Data Grid: Architecture and services'
Vitaliy Muratov

Purpose of the given work is exploration of DataGrid architecture and the main services, analysis of the existing projects and defining the possibility of Ukrainian Grid infrastructure participation in universal DataGrid.
First of all, DataGrid provides services and infrastructure for distrubuted programs which process the large number of information, which need to have access, to transfer and to change large data sets being stored in distributed data stores. That users could maximally benefit from work with infrastructure, there must be provided: possibility of search the needed data set among the available data and finding corresponding resources for data access; possibility to perform as fast data transfer between recourses as it is possible; users' possibility to manage data copies; possibility to choose corresponding computational resources and to process data on them; possibility to manage rights on data access.
In work, possible variants of the DataGrid architecture and basic services are explored, existing projects are analyzed and compared with each other, and possibilities of Ukrainian Grid infrastructure participation in universal DataGrid are defined.
Result of the given work is analysis of architecture solutions and contents of the basic DataGrid services. Conclusions concerning the possibility of Ukrainian Grid infrastructure participation in universal DataGrid are made. In work, basic components of DataGrid architecture and services are explored, and the analysis of its use is given.
The work is performed on 85 pages, contains 3 tables, and 16 illustrations. At work preparation, literature from 83 sources was used.


'Semantic Grid and semantic Web'
Vladimir Nechiporchuk

Purpose of the work is exploration of Semantic Grid, analysis of Semantic Web and the possibilities they can give to the modern scientist. Semantic Grid will be exceptionally important for eScience, because its technologies will give the possibility to work with knowledge which gives extraordinary possibility because earlier the engine could work only with information. It will give the possibility to work at fantastic for today scenarios.
For these possibilities realization, Grid programs must satisfy the line of requirements concerning storing the large information volumes, information content property, access to information content, simple to user, forming communities, information content matching, abstracts, informing users on events, resource reservation, security, reliability, knowledge service, evolutional growth, etc.
Semantic Grid can be considered as service-oriented architecture within which objects often exchange services in certain field on certain conditions. Within this field, establishing object interconnection rules, e.g. receiving and providing services, is possible. It is natural to present service owners and service consumers and autonomous agents.
Thus, each service owner will have one or more agents acting on behalf of him. These agents direct access to services, which they are reliable for, and will guarantee performing of all the coordinated rules (treaties). This activity attracts scheduling local actions according to the available resources and accounts, that corresponding services will be given in time according to the contract. Agents will also act on behalf of service users. Depending on the desirable automatization level, it can be adoption of treaties for service receiving and receiving or providing any results. For regret, agents and service system are only a kind of dream for today.
Now Semantic Grid is realized in rather simple but very powerful form: it supports metadata and ontology principles. Grid uses the method by the help of which information is represented, saved, available, distrubuted and processed. 'Information' here means knowledge, data with certain meaning. It is expected that the next generation will be developing methods by the help of which knowledge can be obtained, used, restored, published and processed to help e-scientists to achieve their concrete goals.Semantic Web realizations moved to Semantic Grid, in particular, RDF, OWL.
The Grid final purpose is knowledge discovery support, and Semantic Web is often considered as the global knowledge base, that is why symbiosis of Grid and Semantic Web is rather obvious, and perspectives of their development are connected. Progressive combination of Semantic Web and Grid will allow simplifying the mechanism of information search, choice and processing, especially in the light of semantic encyclopedias building projects, knowledge bases, spoken languages lexical bases, and the other projects recently having been begun in Web.
The volume of work is 98 print pages; it contains 5 illustrations, 54 foreign and native sources on the link list.


'Grid applications'
Vladimir Prokopchuk

The subject of research in diploma work are Grid technologies applications in fundamental sciences, industry and business, including engineer Grid application on the Chair of System Design, and also the structure of national and international projects and grid infrastructures. All these problems are very important and actual in connection with large-scale program on the National Grid infrastructure development being now developed in Ukraine. Participants of this program are science and research establishments and universities of the country.
Purpose of the work is comprehensive structural analysis of the existing applications for Grid infrastructures, with possible further using of analysis results at solving practical tasks which will appear at creating the National Grid infrastructure. Great attention is also paid to the exploration of existing Grid systems in technically developed countries of the world, to structural organization of these systems. In diploma work, it is tasked solving the problem of Grid infrastructure engineer application on the Chair of System Design of Institute for Applied System Analysis of NTUU 'KPI'. Result of the work is detailed analysis of Grid technologies application at solving the applied scientific and technical tasks, estimation of perspectives of possible scientific Ukrainian Grid applications.
After research results, practical steps for solving the given task are proposed, namely: modification and use of ROOT software for solving applied engineer tasks. The last software version and all the related documentation can be downloaded for free from the specialized portal http://root.cern.ch/. As ROOT is the open source software it will have to be modified to each concrete application, taking into account all the needed application specificity. Taking into account that software has open code and that it was designed in С++ programming language, the Chair will be able to solve this task by performing in-house programming.
Results of diploma work can be used at further more detailed research of each separate Grid application in different fields of sciences, at choosing package of services which will be given to user in future Grid infrastructure, at choosing projects which will be performed on this infrastructure.
It must be noted that tasks of concrete branch applications for Grid must be solved only after creating and complete setting-up the whole Grid infrastructure, setting-up middleware and creating the efficient system of user access to Grid resources.
The work contains 110 pages, 1 table, 6 figures, and references of 60 names.


'Exploration of Grid resources access tools'
Vitaliy Chekaliuk

Purpose of the given work is exploration of program models for Grid which enables creating users' interaction with Grid resources, choosing and analyzing corresponding software, and also estimating the possibility of its use in the National Grid infrastructure.
In work, modern program models for Grid are considered, choice and analysis of needed software is performed. Results of testing the chosen software on NTUU 'KPI" cluster, estimation of this software efficiency and suitability for the further use are given.
Result of the work is analysis of program models and software for Grid. Also, practical results were received, namely: installation on setting up Netsolve software on NTUU 'KPI' cluster; running experiments which proved the reasonability of the given software use. In work, system requirements to Grid program models are systematized, review of modern models is made, and the ways of their improvement are proposed. Analysis and comparison of generations of UNICORE software development is carried out.
Practical value of the work is in fact that testing Netsolve software on NTUU 'KPI' cluster was carried out, and connection with it was performed via Internet, not the local web, i.e., efficiency of work in distributed environment at real conditions was shown. Netsolve system is recommended to installation and use because it provides the convenient interface to computational resources, provides load scheduling and balancing.
The work is performed on 120 pages; it contains 3 tables and 16 illustrations. At work preparation, literature from 12 sources was used.”


'Generations in Grid systems development'
Alexey Shendrik

Subject of diploma work research are Grid projects of different fields of past and our days, and also generations in Grid systems development.
Purpose of the work is exploration of Grid systems, description of first, second and third generations in Grid systems development, beginning from early projects of the 90-ies and ending with our days, analysis of current state and perspectives of the future Grid development, revelation of weak and strong places of Grid systems during th evolution, exploration and connections between different Grid infrastructures, paying attention to Semantic grid and its connection with Semantic Web; areas which need further work are considered.Corresponding materials which throw light upon the ways and directions of Grid technologies and their role for modern science and society were used for research.
Results of the work are exploration of Grid systems possibilities and concepts, and of modern computational infrastructures state, description of huge Grid potential, review of pioneering Grid projects, namely: I-WАY, FAFNER, Legion, Globus, Jini technology, Cactus, DataGrid, UNICORE, WebFlow and the other integrated systems, analysis of such nets as Grid, American Abilene net, European GEANT backbone research net, British National backbone research and education net, Asian APAN net, Global Terabit Research Net (GTRN). Such Grid works as British eScience program and American TeraGrid system are considered.
The work is performed on 117 pages, contains 22 illustrations, and the list of references from 50 names.


'Data Mining in Grid for banking'
Vladimir Sorokoumov

In work, theoretical information on data mining, along with its application to business and science, is given. Analysis of system needed for data mining use in Grid, namely credit scoring in banking, is performed; examples of such systems building with the use of existing tools are given. Advantages and disadvantages of different methods used for credit scoring, together with problems and perspectives of development are considered. Possibility of credit scoring optimization with the use of SAS Institute software is also explored.
Result of research performed is practical part of the work representing the description of process of creating the model of natural person credit scoring with the use of SAS Institute Enterprise Miner software.
Novelty of the work is in using Grid for data mining in conditions of distributed heterogeneous sets, data sources, and different software. In theoretical plan the work gives method of creating model for credit scoring which can be easily ported in future from common computational platform to Grid.
Practical value of the work is in exploring the possibility of using Grid infrastructure and computations for data mining namely in analytical applications, such as credit scoring, in banking. Development of methodology for process of creating model for credit scoring is also valuable.
Work on 120 pages contains 5 tables, and 24 illustrations. At work preparation, literature from 19 different sources was used.


'Data Mining tools in GRID'
Nikolay Sergeev

Purpose of the given work is research of data mining systems which are able to process and load data from separate sources; load multiple toolkits for processing data from different servers, hosts, with different access interfaces; provide the user with convenient developments tools and tools for organization access to data and to developments tools. Multiple data mining systems and toolkits were analyzed in the work. LEAD was found to be the most appropriate.
LEAD contains data mining tools, ADaM modules, xBaya workflow composer toolkit, mechanism of creation services for workflow executing, the most powerful data visualization tool, IDV.
The xBaya workflow composer gives the possibility of workflow graphic building. But xBaya is only the convenient interface.
ADaM modules allow realizing the wide range of data mining functions.
Results of the work are architecture; practical work with workflow building and data mining tools; application start for providing interconnection. Gfac application, which is the toolkit for services setting for providing interface between the workflow composer (xBaya) and data mining (ADaM modules).
The novelty of the given work consists in using innovative tools for data mining systems building, using WEB applications for fast workflow creation and software for creating services for access to data mining tools. In theoretical plan, information on data mining tools is collected; different at purpose and functionality systems, represented on Ukrainian market, are considered. General classification of data mining algorithms is given.
In work, there are performed development of data mining system architecture; analysis of data formats used by these means; analysis and choice of equipment for workflow management; solution for the task of workflow building tool and data processing modules interconnection.
Practical value of the work is that the received set of components allows the distributed data mining systems implementation. Capabilities of such systems were studied in the work. In particular, such systems allow creating virtual organizations, i.e., any collective working in its field can provide its own data mining system, remote access to it, and collaboration on data streams.
Research results can be used at creating the national data store infrastructure, building the new, more perfect search systems which perform search in this or that object classes using association rules.
Work on 131 page contains 4 tables, and 22 illustrations. At work preparation, literature from 75 different sources was used.


'Information security in Grid infrastructure'
Vladislav Bachinsky

Purpose of the given work is exploration of information security organization in Grid infrastructure, analysis and choice of software, which would satisfy modern requirements for information security in swiftly developing Grid infrastructure, developing the model of security structure elements interaction with the purpose of their practical use at Grid organization.
In work, theoretical information on methods of information security organization in Grid infrastructure is given, modern software for Grid security infrastructure building are considered and analyzed. Processing of supra described information is performed, structure of chosen elements interaction at information security infrastructure organization is proposed; procedure of their installation and configuration is described.
Result of the work is developed model of components of information security in Grid infrastructure interaction, recommendations on installation and setting-up of software carrying out the Grid security and enabling keeping records of virtual organizations and their users. This model differs from the previous ones with using software, not interoperable until recently, but, as it is being swiftly developed, in last software versions the possibility of their integration appeared, and that enabled keeping records of users and their certificates, and cetralizingly demarcate their access to resources.
Also in work, the main technologies of information security organization together with optimal software the most suitable for modern Grid system, were considered; recommendations on their setting-up and installation were given.Conclusions and recommendations
Research performed in the given work showed the possibility of use and combination of chosen software at Grid information security organization. Model of security system elements interconnection carrying out the Single Sign On mechanism was chosen. Result of the given work can be considered as practical-theoretical model at organization of information security in Grid infrastructure.


'Grid portal programming'
Vitaliy Matuschenko, Roman Chepurnoy

Purpose of the work is creating Web portal able to work with computational Grid system, and also able to integrate with world Grid organizations and Ukrainian supercomputers, particularly, with NTUU 'KPI' cluster. Mechanisms of problem definition on computational system must be present. Also portal must have the convenient means for interconnection between users, attractive external interface, and possess information useful for portal setting-up and work with Grid systems.
1. The research of existing Grid portals and tolls for their creation is performed. On the basis of this research, decision to use GridSphere toll for Grid portals creation is made. The main advantages for the use of GridSphere are:• Portlet API fully interoperable with JSR 168• Java Server Faces standard full support• Practically full interoperability with IBM WebSphere 4.2• Single authentication system on the basis of JAAS• Possibility of new portlets fast development and implementation • Use of XML• Build-in access control support• Use of Hibernate JDO/OQL• Build-in setting-up and testing Junit/Cactus systems • Support of multilinguinity both for Gridsphere and API portlets.
2. Testing system for testing work with Grid infrastructure is created. It consists of the following sections:• Grid portal(http://sdgrid.org.ua/)• Authentification server• Computational server• Sattistics gathering serverFor interconnection with Grid, Globus Toolkit software is used.
3. Functionality of all the basic portal components is set up and described and detail. Mechanisms of portal set up are also described. The main portlet applications are connected and set up. User authentication and authorization system is set up.
4. Portal user interface is created: styles are remade, logotype is created, and links and headlines are corrected.
5. Portal is translated into Russian and Ukrainian for user comfort. For now, portal is multilingual; user can choose one of 14 languages.
6. Portal information filling is improved. The part that is seen by unregistered user is completely remade. Now unregistered user can receive information about portal without being involved into the registration procedure.
7. Exploration of user-oriented portals is performed, with the purpose to define portal components necessary for communication between the virtual organizations users. After research results, Collaboration portlet (bluesquid - 1.1) from NBIRN company(http://www.nbirn.net) was installed and set up. The given portlet has such user interconnection mechanisms as forums, chats. blogs, RSS reading, message exchange, virtual organizations creating for common work on projects, possibilities to place photos and multiple images on portal. Portal possesses al the basic user-oriented and social net-oriented portal attributes, which greatly simplifies general work on the projects.
8. Portal productivity testing is performed. As the result, it is defined that portal works stably at large loads (20 users at the same time). Narrow places in portal work aren't revealed yet.


'Middleware allocation and support, including repository creation and support'
Andrey Komarov

The work contains theoretical information on world Grid systems; the most well-known and widely-used of them are considered. More attention is paid to Grid being of interest in Ukraine, and also to the possibility of creating software mirrors for them.Technology of repository creation is considered, basic structure for its creation is developed.
Research performed in the work showed the possibility of creating local repository of Grid systems middleware, with the possibility of automatic software update according to manufacturer's software update. The repository created can be used as the basis for further development in this direction.
The work has recommendations for the further repository support and enhancing.


«Data Mining and GRID: large-scale computations management on the basis of new algorithmic approaches, distributed computational resources, and effective scaled transparent software'
Dmitriy Skachko

In work, theoretical information on Data Mining task, methods and means of achieving the needed results are given. General information on algorithms needed for solving each Data Mining task is given. Also in work, variants of middleware able to provide transparent performing of user-level programs are considered. Attention is paid to middleware able to monitor task performing and to provide user with needed software resources.Result of research performed is the practical part of the work representing the implementation of set of algorithms designed for multi-processor computational system. At software creating, the MPI technology for information transfer between processors was used. As the result, improvement of quality of data classification with the use of unconditioned number of processors was achieved.


«Grid service for exchange quotation prediction'
Boris Ponomarenko

In work, methods of technical analysis, used in exchange quotation prediction process, are explored. Grid systems architecture, basic Grid infrastructure levels, basic represented resources and means of their management were considered. The possibility of Grid infrastructure use for solving tasks of exchange quotation prediction was also explored.
On the basis of research of technical analysis methods and peculiarities of Grid computational environment architecture, explanation of choice of the methods and tools for implementation of exchange quotation prediction service is made. Grid service architecture of given domain for making predictions which are used as model by neuronic nets method, along with Globus Toolkit services are proposed.


'Electronic courses design technologies for e-learning Grid systems'
Sergey Kravtsov

The aim of the given work is research of electronic education systems, applying to Grids with the purpose of developing methodology and choosing software enabling fast development of electronic courses for the eLearning Grid systems.
The Grid concept is so embracive, that it enables integration even with those fields whose informatization is considered to be impossible, or possible in very distant future (multiple human-based grid services). The eLearning system isn't exclusion, too. The subject raising the researchers' sharp interest is using cheap Grid computational resources for building educational systems which aren't on principle implemented on separate PCs because computational or disc resources are limited, and also building systems which use virtual Grid laboratories. Grid collaborative possibilities, such as building virtual Grid organizations, are also perspective directions for building eLearning systems.
The concept of fast courses development is considered by many organizations as the means of returning control on eLearning process through content production standardization, content integration into LMS learning process.
Fast courses development also made corporative eLearning available for small and budget organizations which have no large budget for courses development.Exploration of fast courses development concept realization technologies for eLearning Grid systems and development of fast courses development methodic with the use of pioneering technologies are actual now.
In work, theoretical information on modern eLearning systems architecture, e-Learning GRID systems architecture, concept and methodologies of fast courses development are given. General classification of software applied at fast courses development is given.
Analysis of technologies and software used at creating courses for the existing learning systems, is performed, analysis of interoperability of courses developed for eLearning GRID, and modern standards of courses exchange, is performed.
The result of the research conducted is development of methodology for fast courses design for eLearning Grid systems, аnd also choice of the software for realization of this methodology.
Novelty of the work performed is in applying Grid technologies for eLearning. In theoretical sense, the work gives description of possible architectures of eLearning grid systems and methodic of learning content development. Work on analysis and systematization of these methodic into generalized methodic of fast learning content development was made. Also analysis of software being used for eLearning courses development is made, and software set suitable for fast courses development according to generalized methodic, was chosen.
Practical value of the work is in getting the work methodic and the needed software set for building eLearning courses in different eLearning systems, including eLearning GRID.
Methodics of eLearning development within the concept of fast courses development were systematized. The generalized methodic of development of learning content which takes into account the experience of pioneering companies, the developers of eLearning courses, was formulated.
Software being used for learning content development is analyzed. The list of recommended software used on every step of developed generalized methodic for learning content development, is formed.
The results of research can be used as practical guide at creating courses for modern eLearning systems

 

Ukrainian (UA)Russian (CIS)English (United Kingdom)