Big data streaming in Web-based Testbeds
The web-based architecture adds to testbeds key advantages like results sharing, remote test execution, and collaboration. However, complexity of the testbed’s configuration data might undermine the performance due to big data amounts. As a result, the tremendous amounts of data generated by a testbed need to be processed and sent from the testbed server to the web UI, which is non-trivial. Sending all data at once as a bundle is not feasible as the frontend should be portable, lightweight, and simple. In addition, it may take very long, up to several hours, to download large amounts of data depending on the internet connectivity. A long download time or handling big data mount within a user client brings performance issue and thus less ease-of-use for the user, which breaks the idea of bringing a testbed to the web. To eliminate all hurdles and to preserve the simplicity, data streaming is considerably a feasible approach to overcome all shortcomings in terms of data transmission and to work with real-time data transferability at the same time.
Within this thesis, the focus will be on finding a better way to manage a large amount of throughput between server and client. Real-time data transmission technology like WebSocket may help achieve this thesis's goal. For data streaming, the WebSocket protocol will be used to transfer while data is created. As a result, users will be able to observe priority-based logs and results in real-time and be able to see historical information on demand. The challenging task will be finding an appropriate approach to prioritizing generated data in real-time and searching historical information from stored data on the server. As a solution to those problems, numerous mechanisms need to be defined and showcased in a prototypical implementation within the existing testbed aTLAS.
The objective of this master’s thesis is to find an approach or a combination of approaches to solve the previously mentioned problem in the context of web-based testbeds and big data streaming. This particularly includes the state of the art regarding streaming big data to web clients. The demonstration of feasibility with an implementation prototype of the concept is part of this thesis as well as a suitable evaluation with exemplary use cases.