Reduce duplicate code among stream processing engines

Besides Kafka Streams, we're working on support for Flink #20 (closed) and Beam #147 (closed), #21 (closed). With more engines, we experience a lot of code duplication.

Hence, for each/some of the benchmarks (UC1-4), commons projects should be introduced, which in particular should provide data classes theodolite.uc3.streamprocessing.HourOfDayKey.

added Nice To Have modul/benchmarks labels

Part of this should also be a Flink class which provides an adapter for Kafka Streams serdes to make them usable in Flink.

added modul/benchmarks/flink modul/benchmarks/kstreams labels

We also have duplicate code within the same engine and different benchmarks (e.g. UC2 and UC3).

changed milestone to %0.10.0