Near Real Time Analysis with Tabular without DirectQuery
Real-time analytical systems are challenging, and this makes them interesting. The classical solution for real-time systems is to build ROLAP partitions, so to have a MOLAP cache with consolidated data and a ROLAP
partition for fresh information. In Tabular, there is no option for this solution, you either go in-memory or with DirectQuery, making DirectQuery the only available choice for real-time. Is it the only option? By using partitions in a clever way and pushing the process time to its limits, it is possible to obtain good performance for near real-time systems. In this session we will show the techniques used and analyze how much we can go near to real-time by using in-memory databases.