Slide 1
Slide 1 text
AWS Lambda
Layers FTW!
Serverless On Stage #20 - Milan, 12/02/2019
Slide 2
Slide 2 text
Working with AWS Lambda?
Don’t repeat yourself!
Slide 3
Slide 3 text
HELLO!
I am Francesco Lerro
I am a Solution Architect who loves the Cloud
Find me on Twitter @flerro
Slide 4
Slide 4 text
DEPLOY
Code
External Library 1
External Library 2
MyFunction 1
data.json
Function code Code
External Library 1
Custom Module 1
MyFunction 2
data.json
Function code Code
Custom Module 1
External Library 2
MyFunction 3
data.json
Function code
Non DRY!
Slide 5
Slide 5 text
DEPLOY (shared modules)
Code
MyFunction 1
Function code Code
MyFunction 2
Function code Code
MyFunction 3
Function code
External Library 2 data.json Custom Module 1 External Library 1
MyBundle1
Slide 6
Slide 6 text
Lambda Layer
A collection of arbitrary code, third party libraries and data
that can be referenced by many Lambda functions
Slide 7
Slide 7 text
Lambda Layer benefits
∎ Enforce separation of concerns
∎ Enable code reuse between functions
∎ Allow faster deployment, avoiding duplication
Slide 8
Slide 8 text
Using Lambda Layers
1. Put shared components and data in a ZIP package
2. Create a new layer version, uploading the ZIP package
3. Reference layer version from a function
Slide 9
Slide 9 text
Paths
Default paths for Lambda
Layers unzipped
Slide 10
Slide 10 text
No content
Slide 11
Slide 11 text
Lambda Layers limits
∎ Up to 5 layers can be used per function
∎ Max extracted size per layer is 250MB
Slide 12
Slide 12 text
Lambda Layers access
∎ Shareable across same account, different AWS accounts or public
∎ When deleted, a layer could not be referenced by newer functions
∎ AWS provides NumPy/SciPy layer for ML application
Slide 13
Slide 13 text
SAM
A SAM template with
Lambda and Layer,
use SAM CLI 0.9.0+
Slide 14
Slide 14 text
DEMO
Code available on Github: https://github.com/flerro/lambda-layers-FTW
Slide 15
Slide 15 text
THANKS!
Any questions?
Content released under Public Domain
Presentation template by SlidesCarnival