talk

Developing IoT with Zephyr is a journey from hardware all the way to application. It involves multiple teams and expertise, from hardware to cloud and application development. This talk will cover the options for getting a Zephyr app connected (WiFi, Ethernet, Cellular), selecting the right data encoding (JSON/CBOR), securing the data transfer (DTLS/TLS), and choosing a protocol (HTTP/MQTT/COAP). But that’s not the end of the story, the cloud needs to manage devices allowed to connect, consume the data being received, open up options for using that data, and be aware of the continued state of the hardware. And once you have the data you need to build a user-facing application on top of it. Understanding this lifecycle will help us as developers to make good choices on what Zephyr provides, helping ensure successful IoT projects.

Developing APIs over a RESTful interface with JSON payloads is kind of the de-facto standard nowadays, but it still lacks an easy way to build it with a well-defined interface and document it to be used by others. What if we can leverage gRPC’s fast, type-safe, and modern way of building APIs and still be able to provide an interface over REST/JSON ? Check this talk to find out how.

Muitas vezes quando desenvolvemos aplicações de escaláveis hoje em dia, acabamos atrelando parte deles a uma nuvem especifica. Porém isso traz problemas para como vendor lock-in, dificuldade de rodar em ambiente local, falta de portabilidade e entre outros problemas. E existem serviços em comum que são acessados em nuvem como Banco de Dados, Filas de Mensageria, Armazenamento de arquivo, logs, tracing e que poderiam ser abstraídos e preferencialmente intercambiáveis entre nuvens. Nessa palestra quero mostrar como desenvolver apps mais portáveis e ainda assim mantendo as vantagem de se rodar em nuvem.

Vamos ver o passo a passo de como configurar um ambiente de integração continua e deploy continuo usando ferramentas gerenciadas no Google Cloud, sem se queimar tendo que gerenciar servidores. Demos Demo em Flask https://github.com/alvarowolfx/flask-demo Demo de Múltiplos Ambientes https://github.com/alvarowolfx/gcloud-ci-cd-demo Use Case em IoT https://medium.com/google-cloud/serverless-continuous-integration-and-ota-update-flow-using-google-cloud-build-and-arduino-d5e1cda504bf https://github.com/alvarowolfx/gcloud-ota-arduino-update Referencias Deploy de aplicativos móveis Android APK https://cloud.google.com/community/tutorials/building-android-apk-with-cloud-build-gradle-docker-image Flutter e Cloud Build https://medium.com/@lidemin/flutter-ci-cd-with-cloud-build-android-9cd12ade8306 Outros ambientes de execução Google App Engine https://medium.com/google-cloud/continuous-delivery-in-google-cloud-platform-cloud-build-with-app-engine-8355d3a11ff5 Cloud Functions https://cloud.google.com/cloud-build/docs/deploying-builds/deploy-functions https://medium.com/swlh/how-to-ci-cd-on-google-cloud-platform-1e631cded335 https://cloud.google.com/devops https://github.com/GoogleCloudPlatform/github-actions/blob/master/get-secretmanager-secrets/README.md https://docs.github.com/en/actions/reference/context-and-expression-syntax-for-github-actions#github-context https://cloud.google.com/cloud-build/docs/configuring-builds/substitute-variable-values#yaml_2 https://cloud.google.com/cloud-build/docs/building/build-go#building_using_go_modules Cloud Run Quickstart - https://www.youtube.com/watch?v=3OP-q55hOUI https://fireship.io/lessons/ci-cd-with-google-cloud-build/

A junção das áreas de Machine Learning em ambiente embarcado/IoT tem crescido bastante, sendo atualmente chamada de TinyML. Já temos modelos robustos e pequenos o suficientes para rodar até mesmo em micro controladores com 16kb de memória. Nessa palestra vou mostrar as diferentes formas de se trazer modelos de Machine Learning para ambiente embarcado usando o ecossistema do Tensorflow.

Developing APIs over a RESTful interface with JSON payloads is kind of the de-facto standard nowadays, but it still lacks an easy way to build it with a well-defined interface and document it to be used by others. What if we can leverage gRPC’s fast, type-safe, and modern way of building APIs and still be able to provide an interface over REST/JSON ? Check this talk to find out how.