Azure Weekly

Issue #445 - 19th November 2023

Well, what an interesting week it has been. .NET 8.0 was released, there was .NET Conf 2023, then Microsoft Ignite (AKA "Copilot ALL the things!"), and then some unexpected news from OpenAI. This issue is an absolute beast. As ever the Ignite Book of News 2023 is a great place to start for all the official announcements. It's worth going through with a highlighter and marking the things that might be relevant to your organisation. I'm not going to try and summarise everything here, just pull out a few things I found interesting. YMMV!

First up .NET 8.0, the next LTS version of .NET, was released as was Visual Studio 17.8. The .NET Conf 2023 sessions are now available on demand. And a special shout out to Ian Griffiths who presented a session on Modernizing Reactive Extensions for .NET. The Azure Product Groups have put in huge effort to ensure Day 1 support for new releases of .NET; Azure Functions being one of the main use cases: GA: Azure Functions supports .NET 8 in the isolated worker model, they also announced Azure Functions Flex Consumption: sign up for the early access preview, which looks like the Consumption Plan everyone has been asking for. The .NET ❤️ AI session is worth a watch too.

There were so many announcements at Ignite, but the highlights for me were: Microsoft Fabric workloads are now generally available (we've been giving our opinions about Fabric since it was first announced, so are very excited to have reached this milestone). I wanted to give a shout out to Tom Peplow (who you may recognise from our "Decision Makers Guide to Microsoft Fabric" series), who featured in one of the Microsoft Fabric case study videos "Milliman enables self-service actuarial modeling with Microsoft Fabric".

Microsoft announced they have developed two custom chips, Maia 100 and Azure Cobalt to deal with the demands of AI workloads. More detailed covered in Microsoft Azure delivers purpose-built cloud infrastructure in the era of AI. Azure Boost (which offloads traditional CPU workloads to custom hardware) also became GA. Another long-gestating innovation has made it to Public Preview: Announcing the preview of Azure Managed Confidential Consortium Framework which is a framework for building multi-party confidential workloads (Party A brings the algorithm, Party B brings the data) that run on Azure Confidential Computing infrastructure.

The last three articles of note; Azure OpenAI Architecture Patterns and implementation steps, and two bits of architectural guidance Deepening Well-Architected guidance for workloads hosted on Azure and Azure Well-Architected Review Assessment Updates. I hope you enjoy the rest of this edition!