DBIS

Data stream synopses are compact representations of large, potentially infinite data streams that enable efficient analysis and processing without storing the entire stream.
They consist of a corresponding data structure and an algorithm updating it.
Synposes can also be used to reduce the volume of a data stream, for example on constrained network links. Synopses can provide a tradeoff between the data volume and the accuracy of the resulting data stream.

In Cloud-Edge Settings, for example in Distributed Data Stream Processing on the Edge (DSPoE), it is especially important to reduce data volumes on the edge. Here, multiple, distributed data sources contribute to results of global stream queries and reduced result accuracy may be acceptable to reduce overall data volumes.

In the context of this thesis, you will explore how tradeoffs between accuracy and data volume can be configured at runtime through suitable data stream synopses in distributed edge environments.

Towards that goal, you will:

• Research existing approaches and frameworks for data stream synopses, such as Condor[1]
• Evaluate the suitability of existing approaches for hierarchical and distributed stream summarization
• Evaluate the suitability of existing approaches for dynamic accuracy tradeoffs at runtime
• Implement distributed, hierarchical stream synopses

Interested? Questions? Contact Us!

Liam Tirpitz, M.Sc. – tirpitz@dbis.rwth-aachen.de – Tel: +49 241 80-21542

[1] https://doi.org/10.14778/3467861.3467871