Azure Weekly

Issue #433 - 27th August 2023

Lots of interesting articles in a very mixed bag this week. First up: Transforming Text to Video: Azure Open AI, Cognitive Services and Semantic Kernel Python, and a look into Enhancing Recommendation Systems with Large Language Models, and big news for any users of Excel: Python in Excel: Combining the power of Python and the flexibility of Excel, the item of note is that the Python workload is being run in Azure in ACI. This might cause a few issues of concern for some users, and I would expect this to support local MicroVMs (like Firecracker) in the future, so that you don't need to go outside the firewall to execute cells in your spreadsheet.

In the analytics space is also packed this week: Designing Microsoft Fabric Workspaces, creating Metadata Driven Pipelines for Microsoft Fabric, the 4th part of Barry Smart's "Microsoft Fabric for Technology Leaders" focuses on Exploring Strategies Enabled By Microsoft Fabric (using Wardley Maps), Advanced scenarios with private endpoints to Azure SQL Managed Instance, and Jessica Hill provides a deep dive into Notebooks in Azure Databricks

For orgs who want to automate onboarding / offboarding from their HR systems, this is exciting news: Introducing a New Flexible Way of Bringing Identities from Any Source into Microsoft Entra ID. Introducing Azure Artifacts support for Rust Crates is also a welcome addition.

Finally, a post close to my heart; code = energy (carbon footprint) = cost; anything we can do to write more efficient code benefits everyone. We love .NET because we can write incredibly computationally efficient code, and every year the investments by the .NET Product Group deliver performance improvements that we can benefit from for free. We document the improvement benchmarks in How .NET 7.0 boosted AIS.NET performance by 19% In the last 4 years, for no extra effort, we've seen a net gain of an extra: 1,799,570 messages / core / second for "inspect", and 1,254,421 messages / core / second for "read" scenarios.