The Doppler Quarterly Summer 2017 - Page 56

at a Google Next conference two years ago and watched a colleague who knew very little about machine learning models train a model in less than 20 lines of code. My jaw hit the ground. Not only are the cloud providers making the technology simple to consume, but they are greatly reducing the time to market and cost of entry. In industries like genomics and space research, companies are able to perform research that was not even feasible or even possible in the on-prem and DIY model. Source: Google Cloud Platform Figure 2: Sample Big Data Architecture on Google Cloud Platform The Internet of Things The Internet of Things introduces a whole new set of technical challenges because of the exabytes of data that need to be processed live outside of our datacenters or clouds. Building your own IoT architecture requires Herculean efforts. Have no fear though. Public cloud platforms have provided a collection of APIs that interface with your devices and provide the necessary ingestion capabilities, security framework, workflow and infrastructure to simplify and streamline your IoT processes. AWS has even partnered with some of the major chip-makers to embed a SDK on chip for developers to perform processing out on the edge. Here’s another advantage of the cloud. Architects building IoT applications can leverage their existing security, networking and deployment architectures that are the result of previous work building other applications. This is the beauty of what the cloud providers are offering their customers. As companies mature in the cloud and move to the more advanced technology concepts like IoT, intelligent apps, blockchain, and others, they get to build on top of existing investments. Each new managed service comes fully integrated with existing security, logging and monitoring frameworks. Where do we go from here? IoT, 3D printing, robotics, virtual and augmented reality, drones and many other emerging technologies all leverage the cloud in some fashion. All of these technologies produce huge amounts of data or perform various algorithms 54 | THE DOPPLER | SUMMER 2017