Add majority of slides and content

This commit is contained in:
Pavle Portic 2020-06-08 14:13:17 +02:00
parent 0b407351fb
commit a418e81949
Signed by: TheEdgeOfRage
GPG Key ID: 6758ACE46AA2A849
10 changed files with 102 additions and 8 deletions

Binary file not shown.

After

Width:  |  Height:  |  Size: 186 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 54 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 86 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 104 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 211 KiB

BIN
_static/say_agile.jpg Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 218 KiB

BIN
_static/serverless.jpg Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 71 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

BIN
_static/why_few_lambda.jpg Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 337 KiB

110
index.rst
View File

@ -10,7 +10,7 @@ The Serverless hype
What does serverless mean?
--------------------------
.. figure:: _static/empty_racks.jpg
.. image:: _static/empty_racks.jpg
AWS Compute startup times
-------------------------
@ -24,8 +24,8 @@ AWS Compute startup times
* Lambda - milliseconds (when warm and cozy)
* https://firecracker-microvm.github.io/
Lambda lifecycle
----------------
The Lambda lifecycle
--------------------
.. code-block:: python
@ -42,14 +42,108 @@ Lambda lifecycle
Edifice Corb
============
.. figure:: _static/corbusier.jpg
.. image:: _static/corbusier.jpg
* Named after some weird French dude
* Real time person detection
* User interface through SMS
Architecture
------------
Rapid iteration
---------------
.. image:: _static/architecture.png
:width: 80%
.. image:: _static/say_agile.jpg
Person detection
================
* Tensorflow model
* Inference on video frames is compute intensive
* EC2 GPU instances are neither serverless nor cheap
AWS to the rescue
-----------------
.. image:: _static/behold_elastic_inference.jpg
:width: 40%
Still not serverless
--------------------
.. image:: _static/elastic_inference.png
:width: 40%
The AWS SageMaker of six paths
------------------------------
* Managed end-to-end ML service
* Model training, deployment, and monitoring
* SageMaker studio, an "IDE for ML"
Processing pipeline
===================
.. image:: _static/processing_pipeline.png
Triggers and destinations
-------------------------
.. image:: _static/triggers_and_destinations.png
Serverless framework
--------------------
.. image:: _static/serverless.jpg
:width: 30%
True yaml engineering
Triggering events
-----------------
.. code-block:: yaml
functions:
VideoIngress:
handler: aws_lambda.run_detector.handler
events:
- s3:
bucket: video-bucket
event: s3:ObjectCreated:*
rules:
- suffix: .mkv
Infrastructure is hard
----------------------
* CloudFormation bad
* Terraform good
* Multiple environments are harder
Messaging pipeline
==================
.. image:: _static/messaging_pipeline.png
:width: 40%
Y tho?
------
.. image:: _static/why_few_lambda.jpg
More triggers and destinations
------------------------------
.. code-block:: yaml
CleaningReportGenerator:
handler: aws_lambda.request_cleaning.handler
events:
- eventBridge:
pattern:
detail:
responsePayload:
command_type:
- request_cleaning
destinations:
onSuccess: arn:aws:events:::event-bus/default