En jämförelse av prestanda mellan centraliserad och decentraliserad datainsamling

Sammanfattning: In the modern world, data and information is used on a larger scale than ever before. Much of this information is stored on the internet in many different shapes, like articles, files and webpages, among others. If you try to start a new project or company that depends on this data there is a need for a way to efficiently search for, sort and gather what you need to process. A common method to achieve this is called Web scraping, that can be implemented in several different ways to search and gather data. This can be an expensive investment for smaller companies, as Web scraping is an intensive process that requires that you pay for a powerful enough server to manage everything. The purpose of this report is to investigate whether there exist other cheaper alternatives to implement Web scraping, that don’t require access to expensive servers. To find an answer to this, it was necessary to research the subject of Web scraping further along with different system architectures that are used in the industry to implement it. This research was then used to develop a Web scraping application that was implemented on both a centralised server and as a decentralised implementation on an Android device. Finally all the summarized research and results from performance tests of the two applications were used in order to provide a result. The conclusion drawn from these results was that decentralised android implementations is a valid and functional solution for Web scraping today, however the difference in performance means it’s not always useful for every situation. Instead it must be handled based on the specifications and requirements of the particular company. There is also a very limited amount of research done on this topic, which means it needs further investigation in order to keep developing implementations and knowledge on this particular subject.