摘要: RAM3S (Real-time Analysis of Massive MultiMedia Streams) is a framework that acts as middleware software layer between multimedia stream analysis techniques and Big Data streaming platforms, so to facilitate the implementation former on top latter. Indeed, use platforms can give way efficient management large data amounts, but they require user concentrate issues related distributed computing, since their services are often too raw. The greatly simplifies deploying non-parallel like Apache Storm or Flink, fact demonstrated by four different cases we describe here. We detail lessons learned from exploiting implement detailed cases.