Wrap up!
So that’s a wrap on Microsoft Ignite 2024, and for me this week and been rather special, for two reasons mainly.
1 - It’s my first ever “actual” Microsoft Ignite event, not counting the “In/On Tour”
2 - On Thursday (21st), It was my Birthday marking the ripe old age of 30 and there has to be no better way of spending it in Chicago, with some amazing work colleagues and getting to experience the flagship event that is Microsoft Ignite! 🥳
So this is the preliminary report from Microsoft Ignite 2024, Next week, I will be hosting an after-party review with Wesley and Richard to cover all the exciting announcements which have been released at Ignite.
This will be hosted on the BuiltWithCaffeine Youtube channel: https://www.youtube.com/@thebuiltwithcaffeineblog
Community
The community vibes of this week has been awesome, it’s been super fun to finally meet the amazing #CloudFamily people from the various social media platforms I’ve met over the years.
I’ve even managed to make it into a could of vlogs and videos this week.
I had an interview with Kenny Lowe, asking me what was the take away from the Keynote from Day One
I sat down with Wesley Haakman and discussed Ignite and Day One
Talks and Events
For the sessions I’ve seen in person this week have been the following
Day One
Microsoft Ignite - Opening Key Note
Copilot is the UI for AI And we should treat it as such. This allows to open up to way more use cases than just copilots. With Agentic AI and Multi Modal AI we see new use cases arise. Instead of just chatting with our data, we now ask AI to automate things for us, based on data but also based on our requirements.
What’s next is that we need to invest more into reasoning when it comes to AI, so that we can trust AI, put in the safeguards around it and make AI work for us. I would highly recommend checking out the session Mark and Scott presented about responsible AI
We need to treat AI as something that helps us be more efficient and not look at is a job replacement. AI makes mistakes, a lot. As Scott Hanselmann rightfully put it: AI is that eager intern that wants to help you, but they still need to learn and make mistakes.
Building Cloud Native Applications with Dapr and Radius on Azure
In the afternoon of Day One, I attended a theatre session hosted by Will Velida showing how Microsoft doubling down on it’s investments within the open source community with tooling such as dapr and radius. Will showed in real time how you can platform engineer on Azure - I’ll be sure to check in with Will soon to get some more recorded content on how you can use dapr and radius to effetely platform engineer.
If you want to check out the full recording, Here is the link: Building Cloud Native Applications with Dapr and Radius on Azure
Day Two
Inside Azure innovations with Mark Russinovich
During this session there was a lot of cool new technology being showed. It started with a new network card which can support up to 100Gbps per interface, and what was more impressive was the fact that it’s running Azure Linux 3.0 on the card directly. During the demo, Mark showed how in real time you can update the firmware configuration on the network with millisecond downtime, This is a huge improvement from where were years ago, where you we would have to down a whole server to patch a network driver and then bring it back online.
Hyperlight is a lightweight Virtual Machine Manager (VMM) designed to be embedded within applications. It enables safe execution of untrusted code within micro virtual machines with very low latency and minimal overhead.
As we know Microsoft loves their open source contributions, This is where Hyperlight. What Mark was explain was that using Hyperlight you can create micro sized virtual machines, which can run specific code snippets in real time. If your curious about the run time facts.
< 0.03 milliseconds: The time it takes to start a new Wasmtime sandbox
1-2 milliseconds: The time it takes to spawn a new Hyperlight micro-VM
> 120 milliseconds: The time it takes to spawn an optimized, traditional VM
Based on these metrics, this would effecting allow you to build your own Azure infrastructure and automation platforms. allowing you to scale on demand.
Next up we have Azure boost and as Microsoft puts it’s:
“Azure Boost is a system designed by Microsoft that offloads server virtualization processes traditionally performed by the hypervisor and host OS onto purpose-built software and hardware. This offloading frees up CPU resources for the guest virtual machines, resulting in improved performance. Azure Boost also provides a secure foundation for your cloud workloads. Microsoft’s in-house developed hardware and software systems provide a secure environment for your virtual machines.”
If you want to find out more Azure Boost, you can check out the official docs page
If you want to check out the full recording, Here is the link: Inside Azure innovations with Mark Russinovich
Day Three
Entra Id Private Access and Zero Trust
Day Three started with a session on Entra Private Access and Zero trust, For the love of me at the moment, I can’t remember the session name, But when I do I’ll come back and update it. The TLDR was around using Entra Id Conditional access and limiting access via Zero Trust based on IP address ranges.
Scott and Mark learn responsible AI
After that session we then had some fun with Scott and Mark talking about building responsible AI and creating guard rails around the data uses and input and output parameters. One super important point they kept hammering home was to always test the AI model that you’re wanting to use. - This demo was based around the Azure AI Foundry. During the session Scott and Mark also showed how the current AI Models have been trained and gave was real use cases of what they coined as “AI Hallucinations”. This in short is where the AI tries to Gaslight you into thinking that the responses provided are correct, when in reality they are so far from the truth, however the AI model will try and confirm that what it’s provided is given truth.
If you want to check out the full recording, Here is the link: Scott and Mark learn responsible AI
Scale, secure, and optimize Azure Kubernetes Service (AKS)
Finally to wrap up day three, I joined Richard and Wesley for all thinks AKS. Which was pretty cool, As some of the features I’d already heard about from a high level, So it was really cool to get some more deeper context and some live demos. Touching back on the AI track, Most of what we’re seeing at the moment is all being powered by Kubernetes, the reasons for this is that using Kubernetes you can scale super fast from either a vertical or horizontal point.
If you want to check out the full recording, Here is the link: Scale, secure, and optimize Azure Kubernetes Service (AKS)