Serverless? Where were
going we don’t need
Serverless Computing Architecture
Function as a
The Self Driving Car of Cloud Computing
Ok back to reality…. Things to consider with
Loss of control, vendors will put constraints on their services to deliver on reliability and performance goals.
If you are processing a large amount of data (i.e. millions of rows), you will want to break that up into smaller chucks so that
it can be processed within a cost affective timeframe.
Assume your code always has the potential to run in parallel as it can with dynamic scaling.
Its bleeding edge, the tooling is limited (Azure Functions local development is in Beta, No IntelliSense in VS)
Serverless is not magic – its just a way to have some experts build the parts of your app that you don’t specialize in.
Microsoft “Serverless” offering,
Lightweight, Makes composing
Cloud Apps easy
Pay only when your
Azure Functions at a Glance
Serverless (FaaS) Binding to Services Language Support
Serverless event based
processing, auto scales for you
based on your actual usage.
App developers focus on their
business logic / UX. Less on
boilerplate and scaling complex
Announced at Build 2016
Went GA November 2016
Azure Services (Document DB,
Logic Apps, Storage, Service
Github, Sendgrid, Twilio,
Dropbox, OneDrive, Box
HTTP & Webhooks
C#, Node.js, F#, Python, PHP,
Batch and more to come
• A Single application as a suite of small
• Each running in its own process and
communicating with lightweight
• The focus is on solving business
problems, adding capabilities,
competitive advantage – not writing
• Each Service can have its own
lifecycle, deployments and scaling