74. Parallelism and Azure Batch

Показать описание

In this Betatalks episode, Jelle follows up on Betatalks #73, where he and Gerben achieved parallelism using Durable Functions. This time, he introduces Azure Batch, which is a great Azure service for parallelism problems of a more complex scale. Jelle shows you how to get Azure Batch to work as a developer, as it has more overhead than a Durable Function. Then, Jelle discusses the potential bottlenecks and the different ways you can set up a task that needs to run in parallel. He takes a look at the differences between a batch account, a batch pool, and batch nodes and the different ways to set them up. He also shows the different configuration options. Finally, with Azure Batch, there is a need to examine costs; Jelle talks about a few ways to save yourself from unnecessary spending. The code base is, of course, available to download; see the link below.

Timestamps:
00:00 – Introduction
00:55 – Choosing between Durable Functions and Azure Batch
02:40 – Setup using the portal: Storage Account
03:20 – Setup using the portal: Batch Account
04:05 – Application packages
04:38 – The goal of the application
05:40 – Accounts, pools, and nodes
06:42 - Resource creation in portal versus in code
07:03 – Exploring the application
12:04 – Configuration
13:41 – Consuming the Azure Batch API
15:07 – Costs
15:54 – Wrapping up

More links:

There is more to come!