added 2 research items
Motivated by requirements of Web 2.0 applications, a plethora of non-relational databases raised in recent years. Since it is very difficult to choose a suitable database for a specific use case, this paper evaluates the underlying techniques of NoSQL databases considering their applicability for certain requirements. These systems are compared by their data models, query possibilities, concurrency controls, partitioning and replication opportunities.
Big Data describes a recent trend in information processing whereby a huge amount of differently structured data is processed at a very high velocity. The given requirements this kind of information processing demands, tend to exceed the capacity and performance of relational databases at a more frequent rate, which had successfully been used in data handling for decades. Therefore, at the beginning of the millennium, leading e-commerce businesses felt compelled to develop new database systems, which are aligned to specific requirements of Big Data-use cases. The concept of these new database systems motivated the development of a multitude of non-relational database systems, which are nowadays known as NoSQL databases. NoSQL databases promise flexible data modeling, high processing speed and linear scalability. Therefore, they are the center of a currently ongoing hype around the topic of Big Data. However, there is hardly any neutral- and specialized literature of high quality for the above-mentioned topic. As a result, there is no qualitative evidence on the advantages and disadvantages of each database system and their correct application. Due to this difficulty, accessing the world of NoSQL databases is associated with a disproportionate effort and a high susceptibility to errors. The goal of this thesis is to facilitate the access to the complex topic of NoSQL and to reduce the high investment costs, which are currently associated with the implementation of NoSQL databases. Through a methodical reappraisal of the underlying concepts of these systems, universal statements towards advantages and disadvantages of the different database systems can be made and therefore, existing expert knowledge can not only be extended, but especially be enhanced qualitatively. Based on the insights of the conceptual reappraisal, urgently needed modeling patterns are created, whereby conceptual data models can be systematically converted into logical data models for each of the databases. The knowledge, which is gained within this thesis, is summarized conclusively in a set of criteria with which even inexperienced users are enabled to choose the ideal system for a specific use case from the multitude of currently available database systems. Through the clear emphasis on advantages and disadvantages of the different systems, this thesis serves as a valuable contribution towards the objectification of the NoSQL discussion.