How software gave us green fingers

A fun talk about experimenting with innovative software products to keep our office plant in good condition.

Transcript

1. How software gave us green fingers Trying out tech innovations in a fun way
2. About us Arnout Rijpkema Rick Slinkman
3. • Distribution & Access (DT) • Handling orders and licensing to online study materials • Java • Designed and built in 2013 • Running in AWS • DT Platform in the works • Agile team: 11 colleagues + Arnout + Rick What we do
4. *not really our office
5. We need a plant! That will certainly make our office prettier
6. Our product owner gave us a nice plant Awesome! So how do these things work? According to biology teacher: 1. Water 2. Sunlight 3. Take out dead leaves
8. Then we found this Xiaomi Mi Flora Bluetooth enabled Plant Sensor • Moisture • Sunlight • Temperature • Soil fertility
9. Can we... • Show stats in our metrics dashboard? • Store the data in the cloud? • Run that cloud application in Kubernetes? • Be notified when the plant needs water?
10. DT PlantVorm A fun non-production environment to practice with Kubernetes, Lambda's, Angular, Istio, Android Things, etc.
11. Getting plant data to the cloud DT PlantVorm ???
12. Getting plant data to the cloud DT PlantVorm
13. PlantAdapter NodeJS Arnout • Open source MiFlora reader • Collects data periodically • Handles multiple sensors at once Android Things Rick • Custom BLE device scanning • Custom Xiaomi data decoding • Awaits device discovery • Retrofit REST client And then we’ll POST it to some cloud application...
14. Building cloud applications is what we do Don’t settle for a basic application Learn by practice Show cool new innovations to colleagues & friends Great way to show the boss what’s new Show our muscle! Arnout Schwarzenegger
16. So we came up with... Tech • Containerization • Orchestration • Modern web Fun • Alexa • Time lapse Learn about • Docker • Kubernetes • Angular Learn about • Lambda • RPi Camera All this for 1 plant...
18. PlantVorm RainCloud • Persisting of all sensor data • 2 rest end-points ◦ Save sensor data ◦ Fetch sensor data • Database ◦ 14385 measurements saved
19. Get the Rain to the cloud
20. Making the rain • Maven build • Build docker container • Save docker container
21. Taking it to the cloud
22. Kubernetes • Orchestration of docker images • Infra as code • Automating deployment • Self-healing • Service discovery and load-balancing • Docker vs. Kubernetes?
23. Rain cloud Code examples
24. If plants could talk Expectation Reality
25. Getting tired of running the query… Being able to ask the plant how it’s doing
26. Basic workings How your skill knows what the user said Alexa, ... Uploads audio Calls your skill Amazon Voice Services my custom Skill Calls your endpoint Alexa enabled device
27. Building your custom Skill Interaction model • What can Alexa understand? Skill endpoint • Respond to user’s calls Account linking (optional) • Amazon account ←→ myUser.userId
28. Interaction model The hardest part Invocation name How Alexa knows to target your skill Amazon translates speech to Intents • AMAZON.HelpIntent • AMAZON.StopIntent • PlantConditionIntent • LatestInfoIntent Variables in Intents are called Slots • AMAZON.Date, AMAZON.Time • AMAZON.Number, AMAZON.City • PlantDataType Alexa, ask OfficePlantStats how it’s going… Alexa, tell OfficePlantStats to get me the plant’s state Alexa, ask OfficePlantStats if it needs water Alexa, ask OfficePlantStats if it needs sunlight Alexa, ask OfficePlantStats if it needs {{type}} Alexa, tell OfficePlantStats I need help Alexa, ask OfficePlantStats to tell me it’s secrets
29. Skill Endpoint The easy part IntentRequest [Amazon → you] IntentResponse [Amazon ← you]
30. Handling IntentRequests Dealing with IntentRequests as JSON AWS Lambda OR • Easy setup • Security is taken care of • Super scalable • Alexa Skills Kit Custom HTTPS endpoint • Your language of choice • Full control • Could fit in existing application • Alexa Skills Kit
31. DEMO Alexa Simulator Amazon Alexa Developer Console
32. We’re still building! Photosynthesis & Weatherman
33. Current state Project overview Estimated billed amount: € 200 DT PlantVorm exists of.. 1x Kubernetes cluster 2x Spring Boot application 1x Alexa skill 1x AWS Lambda 2x Raspberry Pi 1x Angular application
34. But is it alive? Did all this technology help to keep the plant in good condition?
35. Meet the DT PlantVorm It’s doing great!
36. How software gave us green fingers Arnout Rijpkema Rick Slinkman