The Grand Plan

Welcome to Mediaserver8, a place where I document my adventures in home automation and media serving - mostly for my own reference and future sanity!

I'm going to keep this post on top and maintain the overview schematic at Gliffy to offer an introduction to my HA system as it evolves. (Here's a direct link to a scalable version of this diagram).

Card

Read on for a detailed breakdown of this setup...

unRAID as a software dev. platform

unRaid continues to amaze.

I find myself in the middle of a spare-time software development project that quickly grew from a personal home automation idea to a quite public 'when's it ready' kind of thing. It's the development of an Alexa Skill to allow spoken commands to an Amazon Echo to control Squeezebox devices. (see Hab-Tunes.com for the project details).

Securing Skills

Having got a basic setup working with an Alexa Skill calling to Node-Red on my unRaid server, I was a little worried about the security of this setup long term. I was also unhappy with going through a Lamba function to bypass the HTTPS requirement for Alexa endpoints. I decided to set up better security.

Skilling Up

In anticipation of an Amazon Echo Dot arriving, I've been looking at ways of integrating voice control into home automation. There's a certain amount that Alexa can do by herself but to really tailor things, it's necessary to build custom skills. To get these integrated with my emerging openHAB / Node-RED infrastructure, I set about setting up MediaServer8 to act as a host for the skills logic, thereby allowing me to hook sensors and devices into the skills via Node-RED. (My design objective is to run as much code and services as I can locally to minimise reliances on web services and allow everything to be glued together on my local server).

The idea was to run the core skill logic on my Node-RED install running on my unRAID server. To acheive this, I largely followed this excellent guide which required me to set up an Amazon Developer account. Alexa skills can be configured to process on either the Amazon Lambda service or on a user-defined https endpoint.