Entegre bilgi sistemi modeli geliştirilmesi: DataOCEAN©

Abstract

Günümüzde veri çeşitliliğinin artması ve veri hacminin çoğalması ile güvenilir, düzenli ve güçlü sistem tasarımı gereksinimi giderek daha önemli hale gelmiştir. Merkezi erişim sağlayan yeni nesil bilişim sistemleri tasarlanarak kullanım yaygınlaştırılmıştır. Bu durum, özellikle kurumlar arası veri paylaşımı ve farklı yapıdaki, sistemler arasındaki entegrasyon çalışmalarını hızlandırmıştır. Dolayısıyla, merkezi erişim amaçlayan sistem tasarım çalışmaları artmış, veri erişimi de web servislerinin yaygın olarak kullanıması ile kolaylaşmıştır. Bu çalışma kapsamında, dağıtık sistemler için; hızlı, doğru ve güvenilir bilgiye erişimde kullanılabilecek yeni bir entegre bilgi sistemi modeli tasarlanmıştır. Tasarlanan model; birlikte çalışabilirlik ilkelerini koruyan, hibrid yapı temelli, birden fazla sistemi barındıran, entegrasyonu web servisleri aracılığı ile sağlayarak, çok yönlü veri akışına olanak tanımaktadır. Uzun vadede konum ve zaman kavramlarına dayanarak verileri analiz etmeyi, yönetmeyi ve işlemeyi de amaçlamaktadır. Model, bilişim sistemi çalışmaların kavramsal merkezini (çekirdeğini) temsil etmesi, ayrıca kurumlar arası veri tekrarı durumunda, en doğru verinin tespiti ve güvenilir bilginin alınmasını sağlayacak şekilde kurgulanmıştır. Veri tekrarının, tutarsız analiz sonuçlarına neden olduğu bilinmektedir. Bu nedenle, oluşturulan bu entegre bilgi sisteminde, uluslar arası standartlara uygunluk esas alınmıştır. Birlikte çalışabilirlik ilkesi gereğince, sadece verileri aynı standartta saklamak ve/veya kullanmak değil, merkezi bilgi/veri sistemi tasarımının benimsenerek, verilerin paylaşılması ve ortak kullanılmasının sağlanması çok önemlidir. "Entegre bilgi sistemi modeli geliştirilmesi: DataOCEAN©" konulu tez çalışmasında tasarlanan model ve sistem prototipi Servis Yönelimli (Servise Oriented Architecture – SOA) yapıyı temel alan ve Nesne Yönelimli (Object Oriented - OOA) mimari yapısına benzerlikler taşımaktadır. Kullanılan iskelet uygulaması (framework), MVC mimari tabanlı (Model-View-Controller) yazılmış, özgün bir çalışmadır. REST (Representational State Transfer) yaklaşımı ile tasarlanmış, RESTful ve SOAP (Simple Object Access Protocol) servis mimarilerine ilişkin web servisleri hazırlanmış, sistemin yönetim panelinin içereceği alanlar yazılmış ve test edilmiştir. Uygulama dili olarak PHP (Hypertext Preprocessor) tercih edilmiş, konumsal veri gösterimi için GoogleMAPs kullanılmıştır. Çalışma süresince oluşturulan web servislerinin entegre edildiği yapı DataOCEAN©-DO (Veri Okyanusu) olarak, oluşturulan web servisleri kütüphanesi ise OceanLIBRARY©-OLY (Okyanus Kütüphanesi) olarak anılacaktır. DataOCEAN© sistem modelinde veri yapıları web servisleridir. Web servisleri, sistem kütüphanesine entegre edilme aşamasında, nesne tabanlı programlamada kullanılan hiyererşik oluşumla tanımlanmaktadırlar (içerdiği veri, servis yapısı, v.b. özellikler ile). Veri sorgulaması, dinamik olarak uygulama kapsamında kodlanan web arayüzü aracılığı ile gerçekleşmektedir. Web arayüzünde özgün sorgu yazıldıktan sonra, sorgu işleme aşamasında veri havuzundan kullanılacak web servisleri tespit edilir. Sistemde anlık olarak erişilebilir olan, veri doğruluğu en yüksek servislerden bilgi toplanır, web arayüzünde ilgili formatta sonuç sunulur. Kullanıcı, verinin kaynağı olan kurum ve/veya yapıyı bilmeden, alınan sonucun güncel ve sorgulama anı için en yüksek doğrulukta olduğunu bilmektedir. Dolayısıyla, kullanıcının yüksek yetkinlikte olması beklenmemektedir. Sunulan modelin temel amacı, yazılım sürecini kısaltmaktır. Sistem, bilgi sistemlerinde bulunan verileri paylaşabilmeye uygun mimari ile tasarlanmış olması nedeniyle birlikte çalışabilirlik konusunda da iyi bir örnek olduğu düşünülmektedir. Tasarlanan DataOCEAN© mimarisinde birbirinden farklı sistemler veri alışverişinde bulunmaktadır. Özgün sorgulama platformu sunan ara yüzü ile sorgulama anında sistemde aktif olan web servisleri arasından doğruluk derecesi en yüksek olan servis/servisler seçilerek, işlemi kısa sürede hızlı ve doğru sonuçlandırmaktadır. Bu çalışmanın giriş bölümünde, çalışmanın özeti, kullanılan yöntemler ve sistemin yapısı özetlenmiş, veri ve veri çeşitliliği konularında yapılan çalışmalara değinilmiştir. Birinci bölümde çalışmanın temeli olan web servisleri, OOA, SOA (REST ve SOAP), MVC mimarileri incelenmiştir. İkinci bölümde araştırma yöntemi, geliştirilen DO mimarisi ve OLY kütüphanesi açıklanmıştır. Üçüncü bölümde yazılan prototip uygulama ve alınan sonuçlar irdelenmiştir. Sonuç bölümünde sonuçlar özetlenmiş, hedeflenen sonraki çalışmalar ve çalışmanın katkıları belirtilmiştir.

Today, with the increase in the variety and volumes of data, the need for a reliable, systematic and powerful system design has increasingly become more important. Next generation information systems facilitating central access were designed and their use has become widespread. This has particularly accelerated data exchange between institutions and integration work between systems with different structures. Thus, system design works aiming at providing central access have increased, and data access has become easier thanks to the spread of web services. In the context of this study, a new integrated information system model was designed to be used in providing access to fast, accurate and reliable information for distributed systems. This model protects the principles of interoperability, is based on a hybrid structure, accommodates multiple systems and enables multifaceted data flow by providing integration through web services. It aims to analyze, manage and process data based on the concepts of time and space in the long run. The model was built to represent the conceptual core of the information system works, and to identify the most accurate data and have the most reliable information in the case of data repetition between institutions. It is known that data repetition results in inconsistent analysis results. Therefore, this integrated information system is based on compliance with international standards. In accordance with the principle of interoperability, not only storing and/or using data according to the same standards, but also enabling the sharing and common use of data by adopting a central information/data system design is highly important. The model and system prototype designed in the thesis study "Development of an integrated information system model: DataOCEAN©" is based on service oriented architecture (SOA) while it shares similarities with object oriented architecture (OOA). The framework used in the design is an original work based on model-view-controller architecture. It was designed using REST (Representational State Transfer) style. Web services in relation to RESTful and SOAP (Simple Object Access Protocol) service architectures were prepared; areas to be included in the management panel of the system were written and tested. PHP (Hypertext Preprocessor) was chosen as application language while using GoogleMAPs for spatial data representation. The architecture the web services are integrated in the study is referred to as DataOCEAN© - DO while the web services library is referred to as OceanLIBRARY© - OLY. In DataOCEAN© system model, data units are web services. These web services are in the process of integration with the system library and defined with the hierarchical formation used in object based programming (with features such as data, service structure etc.). Data query is dynamically conducted through the web interface coded as part of the application. After the original query is entered in web interface, web services to be used from the data pool during the query processing are identified. Information is collected from the services that can be instantly accessed and have the highest reliability of data, and the result is presented in the relavent format in the web interface. Without knowing the institution and/or structure that is the source of the data, the user know that the result is up-to-date and has the highest accuracy in the query time. The user are not expected to be highly competent. This model mainly aims at shortening the software process. As it is designed with an architecture that is suitable for sharing the data in the information systems, it is considered setting an example for interoperability. In the designed DataOCEAN© architecture, different systems carry out data exchange. With the interface presenting an original query platform, it selects service/services with the highest accuracy among the web services active in the system during the query, and finishes the operation in a short time with quick and accurate results. The introduction part summarizes the study, applied methods and system structure, and mention studies on data and data diversity. The first part examines web services, OOA, SOA, REST, SOAP and MVC architectures as the basis of the study. The DataOCEAN© Information System Model was built on SOA architecture. This platform-independent system can be adapted to any web-accessing environment. In the DataOCEAN© structure, the user is not concerned with the resources and structures of the web services. However, the user can assume that the most accurate data available during the usage period is being presented. The web services defined in the system are saved according to the data they include and the query types to which they can respond. The second part explains the study method, and the DO architecture and the OLY library that were developed. The web services in the library (OLY) are defined according to a hierarchy similar to the structure of the OOP Namespace (Object Oriented Programming Namespace). The interface established on the DataOCEAN© servers facilitates query production by interconnecting with SQL (Structured Query Language). DataOCEAN© also operates as a RESTful web service set up according to the REST architecture. It is a service which processes a SQL sentence submitted to the service, and then it returns the data in the JSON (JavaScript Object Notation) structure. The model enables the web service library defined within the application to be used in the form of a table by means of SQL. The columns of the table are comprised of request and response parameters. In the service definitions located in the service library (OLY), the parameters are described together with their properties in the form of table columns. The web service registry is made in compliance with the structure of the object-oriented architecture. In the periods after the first web service registry, the alterations in the service structure do not require any code change in DataOCEAN© operations. Only the usage manner of the service structure is updated. The third part addresses the prototype application and results. One of the most significant contributions of this study is the logic of the service registry being generated at the software layer. In the DataOCEAN© structure, traditional URL reading and update transactions are required. The interlinked DataOCEAN© servers synchronize continuously with the service libraries. The changes made on one server are distributed simultaneously to the other servers registered in the system. The DataOCEAN© system is intended to make use of SOA structures more efficiently during integration processes, and to shorten the software process. This system was presented with software, middleware and the implemented working environment. The conclusion summarizes the results and indicates targeted studies and the contributions of the study. In the suggested model, every kind of system located on the web or with an output to the web can be integrated. The written application is a candidate integration layer, which could be considered as a compulsory feature of today's information systems. In this case, it is platform - independent, and enables multi directional data flow by increasing the scope of a given system. Any freshly integrated data can be added to the unified data in a simple, fast, and easy way. However, the added data is required to have a structure which fulfills the standards of the web service. The most important innovation introduced by the study is object-oriented definition logic and the algorithm followed while adding web services to the system's library. Thanks to the suggested object-oriented definition, web services would not only be defined easily, but also would be found easily during the transaction. In addition, local changes do not affect the functioning of middleware in the suggested structure. It was observed that the designed DataOCEAN© Architecture has the capacity to conduct transactions swiftly and correctly by selecting the service/services which has the highest correctness level among the web services active within the system at the moment of inquiry. This is possible due to the interface, which provides a distinctive query platform. In this architecture, different systems also interchange data. In the metadata, the section in which the structures are integrated, any updates can be carried out instantaneously system-wide. In the long term, archive data can be generated via this application. A service oriented, time dependent model which was designed to be convenient for fast data sharing with the last user, was also formed using the architecture. Since we preferred diversity in line with the purpose of the study and price policy in the course of thesis work, hardware design and data safety layer was not taken into consideration.