GearPump was created at Intel®, growing out of real time streaming engine requirement that we wanted to compliment Apache Hadoop* with. At the time GearPump started (June 2014) Spark and other frameworks like Flink were just beginning to get popular. There was also a growing consensus amongst a number of ‘Big Data’ developers that akka / scala provided the tooling to build these types of systems. Most recently we’ve been talking with Typesafe (creators of akka) to collaborate on GearPump and akka-streams which is Typesafe’s reactive framework. We’re very excited about a joint roadmap with this key company.
This new Trusted Analytics Platform (TAP) frees me from those dependencies by exposing a highly flexible set of interfaces that I can use to access data sources and analysis tools. Most of the tools I need, such as an iPython or R interface, message queues, data stores and machine learning libraries are already “loaded” and ready to bind into my resulting application. When I do need help from an application developer, we are working in a common framework, so collaboration is easy and immediate. As a result, when my models are ready, my scoring application is straightforward to deploy and update into production environment.