RESUMEN
Response time measurement of a web-based system is essential to evaluate its performance. This paper shows a comparison of the response times of a Web-based system for Ophthalmologic Electronic Health Records (EHRs), TeleOftalWeb. It makes use of different database models like Oracle 10 g, dbXML 2.0, Xindice 1.2, and eXist 1.1.1. The system's modelling, which uses Tandem Queue networks, will allow us to estimate the service times of the different components of the system (CPU, network and databases). In order to calculate those times, associated to the different databases, benchmarking techniques are used. The final objective of the comparison is to choose the database system resulting in the lowest response time to TeleOftalWeb and to compare the obtained results using a new benchmarking.
Asunto(s)
Registros Electrónicos de Salud/organización & administración , Almacenamiento y Recuperación de la Información/métodos , Internet , Teoría de Sistemas , Humanos , Modelos Teóricos , Oftalmología , Factores de TiempoRESUMEN
Response times are a critically important parameter when implementing any telematics application. Hence, it is important to evaluate those times to check the performance of the system. Different database will get different response times. This paper presents a response time comparative analysis of the Web system of Electronic Health Record (EHRs), TeleOftalWeb, with the four databases used: Oracle 10 g, dbXML 2.0, Xindice 1.2, and eXist 1.1.1. Final goal of the comparison is choosing the database providing lower response times in TeleOftalWeb. Results obtained using the four databases proposed give the native XML database eXist an edge which, added to other features such as being a free software and easy to set up, makes us opting for it. TeleOftalWeb is being used by 20 specialists from the Institute of Applied Ophthalmobiology (Instituto de Oftalmobiología Aplicada, IOBA) of the University of Valladolid, Spain. At this time, there are more than 1000 EHRs and over 2000 fundus photographs of diabetic patients stored in the system.
