[Bret McGowen] Serverless tools for machine learning, big data and app backends

[Bret McGowen] Serverless tools for machine learning, big data and app backends

Presentation from GDG DevFest Ukraine 2018 - the biggest community-driven Google tech conference in the CEE.

Learn more at: https://devfest.gdg.org.ua


Whether your app is mobile, web, or cloud native, Google Cloud Platform and Firebase have an entire suite of serverless tools that can help. What is serverless? It means cloud power available instantly and transparently, without having to create any virtual machines or manage scaling or application runtimes. See how GCP and serverless can make your app's backend more powerful, easier to build, easier to manage, and cheaper. We'll explore compute products like Google Cloud Functions, big data tools like Firestore and BigQuery, and machine learning tools.


Google Developers Group Lviv

October 12, 2018


  1. @BretMcG Serverless for apps, machine learning, and big data Bret

    McGowen Google @BretMcG
  2. @BretMcG What is serverless?

  3. Google's new serverless data center not ^ @BretMcG

  4. What is serverless? Serverless for apps Serverless for data Serverless

    for ML How to put them together Agenda
  5. @BretMcG Principles of serverless Invisible infrastructure Automatic scaling No paying

    for unused CPU cycles
  6. @BretMcG Two types of serverless Compute (often FaaS) Backend as

    a Service (BaaS) Application components as a service • Authentication • Databases • E-mail delivery Application runtime infrastructure as a service • Scaling • Networking • OS management
  7. @BretMcG Two types of serverless Stripe Payments Auth0 Firebase Auth

    Twilio Nexmo SMS/Text Sendgrid Mailgun Email Firestore BigQuery Database Cloud ML Engine ML Code that runs in response to a trigger event Cloud Functions Compute (often FaaS) Backend as a Service (BaaS)
  8. @BretMcG Middleware Compute Databases Cloud Functions App Engine Functions as

    a Service, event-driven Platform as a Service (standard environment) Firestore Cloud Datastore NoSQL document store & sync BigQuery NoSQL Data warehouse & analytics Machine Learning Cloud ML Engine Serverless Tensorflow training & prediction AutoML Training & prediction from examples Cloud Dataflow Stream & batch data processing Cloud Pub/Sub Global real-time messaging Frontend Data Studio Firebase Serverless with Google Cloud
  9. @BretMcG Serverless for apps Mobile, web, and cloud

  10. @BretMcG Serverless for apps and compute Mobile apps Functions as

    a Service App backends (e.g. APIs, web apps) App Engine Cloud Functions Firebase Fully managed, auto-scales Versioning, traffic splitting Java, Python, Go, PHP, Node.js Event-driven functions Microservices, cloud "glue" Node.js, Python (beta), Go (alpha) Real-time syncing databases Authentication, analytics, notifications, hosting, storage, ML Kit, much more...
  11. @BretMcG Cloud Function event triggers Cloud storage (incl. Firebase) Cloud

    Pub/Sub HTTPS Firebase realtime database & the new Firestore database Firebase authentication Google analytics for Firebase Firebase hosting Firebase Crashlytics
  12. @BretMcG Access 20+ Google services from GCF Cloud Storage Cloud

    Pub/Sub HTTPS Firebase Cloud Scheduler Cloud Functions as cloud glue
  13. @BretMcG New Cloud Functions languages!

  14. @BretMcG What's new in Cloud Functions Generally Available (with SLA!)

    Cloud Scheduler Python 3.7, Node 8, Go 1.11 (alpha) Regions: Tokyo, Belgium, 2 in US Environment Variables New! Ubuntu 18.04 with many packages (ffmpeg, imagemagick, headless Chrome) Security Controls: VPC, IAM Scaling Controls Cloud SQL Direct Connect New!
  15. @BretMcG Serverless containers on Cloud Functions Stateless Event/Response Auto-scaling Healthchecks

  16. Demo(s)

  17. @BretMcG Serverless containers on Cloud Functions Later this year: sign

    up at g.co/serverlesscontainers Provide arbitrary container images and run them "serverlessly" • Takes a pre-built Docker image • Use arbitrary base images • Use arbitrary system libraries • Use arbitrary language runtime • Same serverless execution environment ◦ No servers ◦ Pay only while code runs
  18. Google's platform for building mobile and web apps Backed by

    the power of Google Cloud Platform
  19. @BretMcG Improve app quality Crashlytics Performance Monitoring Test Lab Grow

    your app Analytics Predictions Cloud Messaging Remote Config A/B Testing Dynamic Links Build better apps Auth Cloud Functions Cloud Firestore Hosting ML Kit Realtime Database Cloud Storage
  20. @BretMcG Firebase Google's mobile platform Serverless databases Authentication Analytics Notifications

    And much more... Google Cloud Platform Google's public cloud Machine Learning Big Data IoT Virtual machines And much more...
  21. @BretMcG Supercharge Firebase with GCP Cloud Firestore Triggered by Firebase

    events Cloud Functions BigQuery Big data analytics Cloud Pub/Sub Integrate with other applications Cloud ML Engine Machine learning Firebase Google Cloud
  22. @BretMcG Serverless for machine learning Images, text, audio and more

  23. @BretMcG Fully trained ML as an API Cloud Natural Language

    Cloud Vision Cloud Translation Cloud Speech Cloud Video Intelligence What's new? New Cloud Text-to-Speech
  24. @BretMcG Generic task Someone else has solved this before Trained

    on common classes ML APIs TensorFlow “dog” What type of ML problem are you solving? “Miss Potato” Custom task Specific to your dataset
  25. @BretMcG Cloud ML Engine Fully managed platform for TensorFlow Distributed

    Training with GPUs (and TPUs) Fast and scalable online/batch prediction cloud.google.com/ml-engine
  26. @BretMcG AutoML Generic task ML APIs TensorFlow “dog” What type

    of ML problem are you solving? “Miss Potato” Custom task AutoML
  27. @BretMcG AutoML Vision Photo dataset Train Deploy Serve Generate predictions

    with a REST API AutoML Vision to the rescue
  28. @BretMcG AutoML Demo

  29. @BretMcG Learn more • Sign up for the alpha: cloud.google.com/automl

    • Blog post: bit.ly/announcing-automl • Intro video: bit.ly/automl-intro-video • Podcast: bit.ly/automl-podcast
  30. @BretMcG Raspberry Pi Android iOS TPU GPU CPU TensorFlow: many

  31. @BretMcG Serverless for data

  32. @BretMcG Serverless for data big and small Data for your

    application Analyze your data Process your data Firestore BigQuery Cloud Dataflow NoSQL, scale, real-time sync Data warehouse, SQL Transform and enrich data in real time or batches
  33. @BretMcG Serverless NoSQL database NoSQL: document-based database Realtime updates: synchronize

    changes across clients Offline support: Read, write, query even when offline Scales: Multi-region, high availability, consistency guarantees Serverless pricing: pay by database operation, storage, bandwidth Firestore
  34. @BretMcG Serverless big data analytics with BigQuery Fast: petabytes, terabytes

    with blazing-fast queries Simple: SQL; empower your data scientists Interoperable: Java, Python, Tableau, R… Serverless pricing: pay per TB of data queried (flat rate available) Instant data sharing Free monthly quota—1 TB of queries cloud.google.com/bigquery/pricing Managed big data analytics warehouse
  35. @BretMcG BigQuery Demo

  36. @BretMcG Qualities of a good RDBMS

  37. @BretMcG Qualities of a good RDBMS • Inserts & locking

    • Indexing • Cache • Query planning
  38. @BretMcG Qualities of a good RDBMS • Inserts & locking

    • Indexing • Cache • Query planning
  39. @BretMcG

  40. @BretMcG

  41. @BretMcG

  42. Examples

  43. @BretMcG Firebase Auth App Engine Cron Firebase Realtime Database Cloud

    Function Cloud Function Lottery Send Notification
  44. @BretMcG Merch Firebase Realtime Database

  45. @BretMcG • Comprehensive Smart Parking Solutions • Industrial grade IoT

    • More than just about parking Smart Parking
  46. @BretMcG Thank you! @BretMcG cloud.google.com/serverless cloud.google.com/automl firebase.google.com cloud.google.com/bigquery cloud.google.com/vision Free