🐾Computing for Data and ML - AWS Lambda🐾

Advantages:
🔹 You don’t need to manage underlying resources and you pay only for what you use
🔹 You can start quickly - the usual start time is up to 5 seconds
🔹 It scales based on your concurrency limit automatically. Great article to understand AWS Lambda scaling by Julian Wood.
🔹 It can be triggered by scheduled job, S3 upload, HTTP request, SQS, etc.

Limitations:
🔸 Function and layers should take no more than 250 Mb in total (unzipped)
🔸 You have 10,240 MB memory and 10,240 MB storage limitations
🔸 Workload running time should be under 15 minutes
🔸 No support for GPU

Use cases:
🔹 Simple data processing for small files
🔹 Machine Learning inference
🔹 Data drift monitoring

AWS Lambda with Docker adds the following capabilities:
🔹 You get 10 GB for your libraries and other artefacts
🔹 You can use languages that are not available in Lambda, for example, R

---

If you like this post, you can share APAWS newsletter with friends: