AGT is passionate about its unique analytics platform that converts data into business value across various industries and verticals. Our focus is on using innovative IoT technologies to complement and extend established value chains, for example in the sports and entertainment sector. We provide the opportunity to drive ideas from conception to products which are deployed in the market. In order to meet the high expectations that are set into our system, we use different KPI to constantly evaluate and improve the quality of our platform. In the following we shortly explain three of them.
In sports and entertainment area the time aspect is very important. Fans, players, and coaches need to receive information about the event timely. Our software must satisfy strict latency requirements in order to being able to deliver high quality of service. We require the same low latency from third-party software we use. The HOBBIT project benchmarks would be a convenient and proven source of information to compare software latency.
Another KPI which is important for our software and the software that we integrate is throughput. Our typical installation includes hundreds of sensors measuring thousands of parameters that are streamed to our analytics platform. Many of the analytics we implemented require big amounts of data to produce precise results. Usually, the more data the better the results. So it is intrinsically important for our software to cope with high throughput. The HOBBIT streaming benchmarks would give us throughput measurements to make an informed choice.
Last but not least, the accuracy of the results is an important KPI. When basing crucial decisions on the outcome of our analytics, the customers must be able to fully rely on them. Especially in the sports and entertainment sector, quantities of interest are very often not simple physical quantities, such as temperature or air pressure, that stand for themselves. Instead, quantities of interest are derived from those basis physical measurements, e.g. recognizing applauding from hand movements or audio streams, with the best possible accuracy. The HOBBIT accuracy benchmarks give us the opportunity to select amongst the best tools.
Authors: Alexander Wiesmaier, Roman Katerinenko