AWS re:Invent 2017: Serverless Architectural Patterns and Best Practices (ARC401)

[@atif7865]

This dwarfs a lot of sessions. Well prepared and great flow!

[@alexroman]

One of the best sessions I've visited. Thanks a lot!

[@AlessandroBehling]

Thank you AWS awesome talk !!!

[@danielpapukchiev3754]

Awesome talk! We thought of this architecture a couple of months ago and we didn't know if it would work. Now with the confirmation we are more than excited to launch a fully serverless architecture in production!

[@abhishekarrawatia]

Very Informative, Thanks !!

[@worldexplorer796]

useful video. both guys explained well.

[@MsBeverlyDennis]

These guys are great!

[@lijie6431]

Can you have a file server in serverless? The drive shows up as a mapped drive for the user.

[@curiosull]

Great Value Bleach S3 is used for files in this scenario. If you need file servers on premise see storage gateway.

[@hpar121]

Hi , I have a question on asynchronous lambda call from one lambda. e.g. my main lambda function reads the number of line in the input csv file. and trigger the second lambda function which is my main API to process records ranging in 15K group. something like first instance will process 1-15K, second instance will process 15K-30K, third will process 30K-45K etc... now the API functions can be invoked in count of 1 to 1000 depending upon the total records in input file. what I want to do is catch or find out when all the child function instances will complete. so I can run another third lambda function to merge the results created by APIs. what should I use? One solution I thought to use AWS Dynamo DB to write the status of each child function and run the third lambda when all the records come to completed status. Other solution I thought was to use the Step function which basically call first lambda (Which intern call main API let say for 135 times) and than wait for 2min and run my final third API. any suggestion? How to achieve this in better way without doing lot of custom coding.

[@HEXX12341]

Pvt message me on this. It sounds familiar to what I am currently working with. I'd like to hear about your implementation of it and how well it worked.

[@zilupe]

Didn't they suggest looking at pywren.io for this?