Practical Examples of Serverless Architecture using AWS Lambda and PyWren as SQS Python Workers - Data Science and Engineering Club Meetup - 2018.06.27
Similar to Practical Examples of Serverless Architecture using AWS Lambda and PyWren as SQS Python Workers - Data Science and Engineering Club Meetup - 2018.06.27
Building your own calendly using amazon app syncDhaval Nagar
Similar to Practical Examples of Serverless Architecture using AWS Lambda and PyWren as SQS Python Workers - Data Science and Engineering Club Meetup - 2018.06.27 (20)
Practical Examples of Serverless Architecture using AWS Lambda and PyWren as SQS Python Workers - Data Science and Engineering Club Meetup - 2018.06.27
1. Practical Examples of Serverless
Architecture using AWS Lambda and
PyWren as SQS Python Workers
by Juan Pizarro
Data Science and Engineering Club Meetup 2018.06.27
https://www.meetup.com/Data-Science-and-Engineering-Club/events/251202735/
2. Juan Pizarro jpizarrom@gmail.com https://www.linkedin.com/in/jpizarrom/
● Half-time student (2018-2019)
● previous
○ Lead Software Engineer (2015-2018) at https://www.admetricks.com
https://www.youtube.com/watch?v=IvbiDE-lY_k.
○ Developer (2013-2015) at http://a-dedo.cl (now https://www.allrideapp.com).
○ Co-Founder (2009-2012)
○ Computer Science Engineer from University of Talca (2011)
3. Agenda
● Serverless/AWS Lambda
● PyWren
● PyWren as SQS worker
● PyWren use cases (maybe demo)
○ Hyperparameter optimization, e.g. scikit-learn distributed GridSearchCV
○ Web Scraping
○ ...
4. Serverless
Data centre: Hardware as the unit of scale. Abstracts the physical hosting environment.
IaaS: Operating system as the unit of scale. Abstracts the hardware.
PaaS: Application as the unit of scale. Abstracts the Operating System.
Serverless/FaaS: Functions as the unit of scale. Abstracts the language runtime.
Ref: https://read.acloud.guru/iaas-paas-serverless-the-next-big-deal-in-cloud-computing-34b8198c98a2
6. AWS Lambda
AWS Lambda Run code fully managed without thinking about servers
Pay only for the compute time you consume
Scales automatically to the incoming event rate
Runs your code in parallel
Ref: https://www.slideshare.net/AmazonWebServices/massively-parallel-data-processing-with-pywren-and-aws-lambda-srv424-reinvent-2017/5
Ref: https://gfycat.com/gifs/detail/FlawlessFlusteredCony
11. [...] there is little or no dependency or need for communication between those
parallel tasks, or for results between them.
Ref: https://en.wikipedia.org/wiki/Embarrassingly_parallel
Ref: https://www.youtube.com/watch?v=p0aLQLkzP24&t=32s
Ref: https://www.slideshare.net/AmazonWebServices/massively-parallel-data-processing-with-pywren-and-aws-lambda-srv424-reinvent-2017/7
Embarrassingly parallel
43. Ref: MapReduce on AWS Lambda
https://github.com/bcongdon/corral
https://github.com/d2si-oss/ooso
https://github.com/awslabs/lambda-refarch-mapreduce
44. PyWren Design Goals
1. Very little overhead for setup
2. As close to zero overhead for users as possible
3. Target jobs that run in the minutes-or-more regime
4. I don't want to run a service
5. It has to be from a cloud player that's likely to give out an academic grant
Ref: http://pywren.io/pywren.html