Fred George

IoT, the prevalence of specialized interconnected devices, has invaded business and home. The myriad of devices, hubs, and APIs has created a Tower of Babel that makes the Android phone scene seem mild. Such devices are inherently less network ­reliable than previous products, primarily due to the casual attachment (convenient outlets). Add to this heterogeneous soup the security exposure that has already been exploited in DDoS attacks, and you have an environment begging for technical solutions. We propose Asynchronous MicroServices as a solution. Borrowing a page from historical J2ME thinking and more recent implementations using IFTTT, we introduce small bridging MicroServices for the various devices. Further, also consistent with MicroServices, we keep these services very simple: They broadcast device status to an event bus, or listen for action commands to relay to the devices from the same bus. Interaction among the various IoT devices is delegated to yet another set of MicroServices. One exemplary composite application would accept signals from motion detectors to turn on appropriate lights, or alternatively alerting the owner of unauthorized intrusion. Using the same motion detectors, lights can be dimmed and eventually turned off. Room temperatures and even audio/video gear could be adjusted as well. Using Asynchronous MicroServices, we can also set up controlled access from the outside world, rather than exposing each device with its own vulnerabilities. Finally, we wrap up talking about the challenges of implementing this in my own flat using Hue lights, Amazon Alexa, and 4th generation Apple TV. We run these MicroServices in local containers attached to the same home network. Docker support of ARM devices enables low-cost redundancy as well.

Hands-on development of asynchronous micro-services using a bus architecture. While “micro-services” is becoming a heavily over-worked term, it has been interpreted many ways. In this hands-on workshop, we will explore a particular style, tiny asynchronous services, that was the core of one of the most successful implementations. Given an animated specification of flow, we will embark on delivering the proposed services, wrestling with the issues common to this new paradigm. Participants have a choice of languages, with new starting code in Java, C#, and Ruby. At the end of the workshop, we will discuss MicroServices challenges and organizations that are optimized for such architectures.