| dc.description.abstract | Medical images are used daily to support diagnosis in different areas of Radiology throughout the world. These images follow an international standardization defined by ISO 12052, known as DICOM (Digital Imaging and Communications in Medicine) standard. Each institution that claims compliance with this standard has its own storage services, visualization and processing systems specific to that data. However, there is a great need for these images to be analyzed by different specialists, so that each case can be discussed in a broad way, in the search for the best treatment for each pathology. The unavailability of real-time data for specialized medical evaluation has a direct and profound impact on therapeutic success. The cloud computing model has the necessary characteristics to ensure that these images are within the reach of the professionals most recommended for each case, able to offer the best service. The large amount of resources available in the cloud to handle this data in a scalable way facilitates the creation of an infrastructure to support remote diagnosis through Telemedicine resources. Based on the computational paradigm Publish/Subscribe, we can establish real-time communication to solve situations in the field of health, such as communication between hospitals or clinics and between doctors, nurses and experts involved in the diagnosis. In clinical environments that deal with massive transmission of high resolution images in the DICOM standard, as well as in environments with network performance problems, transmitting these images in a timely manner, storing and making them available securely is a problem without a spontaneous solution. In this way, this work proposes a computational cloud-based architecture to collect, compress, store and retrieve data using the Publish/Subscribe paradigm and two levels of scalability. The PS2DICOM model is established as a middleware that provides infrastructure resources in the IaaS (Infrastructure as a Service) layer, supporting the tasks of transmitting and storing files within the DICOM standard. The system offers data compression with different intensities, depending on available bandwidth. The PS2DICOM model also has two levels of load balancing and the reactive elasticity offered by the infrastructure. The research contributes to presenting an efficient architecture to optimize network tasks, capable of being adopted as a solution when developing applications focused on computational clouds applied to health in future real situations. The architecture was tested using a prototype with distinct modules, developed for each specific service offered and proved to be efficient as a solution to the problems in question. Its details are described in the following chapters, as well as its implementation, which corroborates the viability of the model. | en |
