The paper centers around big data processing and analytics, as well as big data compression when doing numerical simulations of hydrodynamic and acoustic processes in large industrial applications. One of the typical cases is the simulation of supersonic turbulent jets and high-intensity acoustic loads in rocket lift-off process. We give some estimates of the memory amount for correct simulations of gas dynamic and acoustic processes in this case. Moreover, we describe one of the most common technologies of big data compression and analytics — Proper Orthogonal Decomposition. This technology was implemented using the Apache Spark framework for large-scale data processing. The necessary memory for the storing of the results of simulation of acoustic pressure propagation can be reduced by 30 times for 1D case and 160 times for 2D case.
