Skip to main content

Ignite 2023 - Day 1

·1178 words·6 mins

Day one of the Microsoft Ignite conference, the word of the week is “Copilot”. (Or word of the year.)

Keynote Presentation #

You knew within 5 minutes that the theme of this conference was going to be AI. And if you’ve been paying attention to Microsoft, you know that the company has fully oriented itself towards being a leader in everything AI. They are putting heavy emphasis on their Copilot assistants in every product and service, and making the development of customized AI tools or solutions accessible to businesses. This is not the place to be if you are sick of hearing about AI.

The keynote went over a lot so I’m just going to bullet-point some of the items here:

  • They announced hardware they’ve designed and built specifically for AI workloads in the form of a CPU and special datacenter compute rack.
  • They talked about a whole series of new AI-oriented Azure services:
    • Models as a Service (Llama2, Mistral, Jais)
    • Small Language Models
    • NVidia AI Foundry Service
    • Azure AI Studio
  • As expected, Microsoft Fabric is now Generally Available
  • They spent a whole section on the new Copilot products:
    • Bing Chat/Enterprise is now just Copilot and is free for anyone. When a work user signs into the Edge browser it now defaults to protecting any questions or results so they aren’t logged or tracked.
    • Copilot for Microsoft 365 is now the paid, licensed version of Copilot that allows administrators to provide custom, internal data sources for the AI service to learn about for internal queries
    • Copilot for 365 also allows you access to Copilot studio where you can package and control those different data sets for Copilot to use and publish the chat instance to other applications
  • Lastly they discussed a number of other 365-related application features:
    • Windows 365 (the AVD as a service license) now supports Windows 365 Boot which allows a computer to boot directly to your Azure Virtual Desktop
    • Sharepoint Premium
    • New Desktop version of Teams, along with Teams Premium and Mesh for virtual meeting rooms
    • A new Teams meeting transcription feature which tells teams not to retain the content of transcribed meetings. Allowing for talk-to-text transcription without recording the meeting.
    • Something called Loop? This is one of those Microsoft Products that looks like they are designing it to do everything project-related while presenting it in a way that it looks so flexible you are not really sure what you are supposed to use it for. They seemed really excited about it.

Breakouts #

I tried to focus on Azure infrastructure or platform-related breakouts, as others on my team who have more AI and Analytics backgrounds were going to be prioritizing those types of sessions.

Entra Identity Governance #

This demo walked through some of the automated account provisioning features in Entra ID, which allows integration with existing HR system to automatically provision group access for new users. They have a feature where you can do attribute mapping based on SCIM to load user data from the source system, and then you can use filters based on those attributes to automatically provision birthright access.

They also discussed that they are making improvements to the access reviews feature based on customer feedback. Allowing the ability to apply filters to access reviews so you are not constantly prompted about users who’s roles are not likely to be changing, allowing you to focus on employees who may merit closer scrutiny.

What’s new in Azure IaaS #

They recapped the new AI-oriented data center hardware like the CPU. And then talked a little further about the new Azure Boost which can be used to get improved performance on Virtual Machines.

They also discussed:

  • Azure Storage Actions
  • Autoscaling for Azure ExpressRoute connections
  • A new Cold tier in blob storage, which resides between “Cool” and “Archive”, so it has less cost than the Cool tier but allows instant read access where the Archive tier requires a rehydration delay.
  • Azure Storage Mover

Azure VMWare Solution #

I was curious about this product because I’d never looked into it and we have a significant on prem VMWare presence. I missed the beginning of the session but it sounds like up until this point, this product was basically utilizing existing VMware hosts that were configured in an Azure data center and there have been some setup/management difficulties related to using the product.

They talked quite a bit about moving to a Azure fleet(?) integration which sounded like running the VMware products more natively on existing Azure data center equipment which allows for simplified networking, management and provisioning. Because I missed some of the early discussions I’m not exactly sure if this is accurate.

New Applied Skills Badges #

I went to a couple sessions about the new Applied Skills badges out of curiosity. It sounds like there was some demand for the ability to learn and demonstrate proficiency with individual Azure products or solutions outside of a full role-based certification. To earn these new badges, it only requires you to complete a Microsoft Learn module about the product and then you are given a situation-based lab challenge where you need to use the skill to satisfy the requirements.

It sounds like people are a little unsure of how these will really fit in relation to existing certs as a lot of questions seemed to be centered around. These appear to be more focused learning modules which allow you to a single service you may be unfamiliar with, and they are just giving you a badge so you can demonstrate that skill even if the knowledge is less of a resume builder and more of a personal achievement icon.

Applied skills credentials do not expire (currently). They also seem like a really great free tool that gives you access to a lab environment for hands on learning.

Platform Engineering QA #

Platform Engineering seems to be the new name for DevOps, or a combination of DevOps, DevSecOps, GitOps, etc. I enjoyed the natural irony of the opening question being “What is Platform Engineering?”, it seems that the discussion has picked up exactly where DevOps has left off. It seems the answer to the question can still be answered by the Supreme Court Justice Potter Stewart’s definition of obscenity; “I know it when I see it.”

Joking aside, this was a really good discussion on having a team dedicated to providing tools or systems for developers or app owners or any Azure internal customers to provision resources. I liked the statement by one of the experts that the “platform” itself often is different depending on the customer of the platform and their needs. He made a point that it’s the job of the platform team to work with customers, identify their pain points and they build a solution for that problem. Those solutions or interfaces may vary quite a bit.

The other big point of emphasis was that platform engineering by necessity is a function that involves multiple technology teams like Infrastructure, Security, and Enterprise Architecture to come up with unified, simple and scalable platform solutions.

Related

Azure, Terraform, GitHub and My Last Year
·1034 words·5 mins
I prefer to write posts that are topical, however I haven’t been writing anything in over a year so I don’t think it’s going to be a problem to spend some time writing a personal catch-up post.
Terraform Comparing Count and For_Each
·1068 words·6 mins
When I was looking at building the module that could attach additonal disks to servers in azure I know I needed to support the ability to add and potentially remove data disks without the risk of losing data.
Terraform - Azure Managed Disks with For Each
·1204 words·6 mins
We have a use case with Azure where we want to support the ability to add, remove or resize additional data disks on a VM.