Businesses across many verticals are experiencing a "push to real time". Increasingly, value is being delivered by analyzing and acting on ever growing data streams in shrinking time frames. Often the window to act is a few seconds, sometime even less. In this talk we'll examine some of the technical challenges presented by "the push to real-time" and how Hazelcast, the open source, in-memory, stream processing platform addresses those challenges, enabling real-time decision making in a way that is not feasible with more conventional approaches.