Architecture for Interfacing Content Analysis Back Ends

This thesis will discuss the challenges of designing and implementing a complex system for the content analysis ecosystem. The system would consist of any number of client devices – both desktop and mobile clients – and of a single front-end service, which would utilize any number of external service providers and back ends designed for various content analysis tasks. These tasks could include, among others, text summarization, photo analysis or content-based search. The background research suggests that despite the extensive studies on content analysis in general, there is a lack of research on how to implement a usable system consisting of all of the required parts. In existing studies the priority is almost entirely placed on algorithm and analysis method design.

The purpose of the research presented in this thesis is to study how to define a generic, extendable, and maintainable system within the content analysis domain. This thesis describes the methods, technologies, and principles required in architectural design. Furthermore, the work is validated through several architectural iterations and proof-ofconcept implementations, which are also presented in this thesis.

Based on the results of the studies performed, this thesis will describe how to realize a feasible and practical generic architecture by following layered Application Programming Interface (API) and data models (presented in this thesis) in combination with commonly used architectural and technical solutions (Representational State Transfer (REST) and hybrid REST/Remote Procedure Calls (RPC)) and industry de facto representation formats (such as JavaScript Object Notation (JSON) and Extensible Markup Language (XML)).