AI, Cloud Strategy, & Data Control for Modern Cloud at Cloud Field Day 23 – Tech Field Day Takeaways

I'm Alistair Cook Tech Field Day event lead, and here are my takeaways from Cloud Field Day 23. This edition of Cloud Field Day was primarily focused on data and applications. We saw specialized vendors with multi-location data storage or protection technologies that target unstructured data as a strategic asset within your organization.

We also saw infrastructure management and the ability to run the cloud edition of SAP in your on-premises data center. As with every IT event, it was an undercurrent of IT enablement and IT agents and AI adding, uh, value to products across the board. My first takeaway was around cyber resilience and data sovereignty and how they're driving modernization across cloud architectures.

We saw quite a lot of products that enable applications to remain on premises and to be accessible from cloud or data to remain on premises and be accessible from cloud or equally data being created in in clouds at specific geographic locations to retain that sovereignty but still enabling access to that data from other locations. Another element in this was the need to integrate with existing legacy on-premises applications, and that was often what the driver was for having some sort of mobility story around getting that data in and outta the cloud, and particularly back to those on-premises applications. We particularly saw on-premises a applications as no boat tank is probably a little bit too nasty a term for it, but as, as a a central point for this conservative and compliance heavy environments to a remain premises because that's where the core of their application is and they've built huge amounts of knowledge and expertise around those applications.

This is why hybrid and multi-cloud is very much a standard kind of approach for organizations that are enterprise scale and particularly compliance heavy ones. We did see hybrid cloud and private cloud architectures as the default for enterprise applications. That combination of the legacy built on premises applications where a lot of business process still runs, connecting, backup to things that are more cloudy and cloud-like.

We saw the idea that the cloud is an operating model rather than a location. It's a way of approaching how you provide and consume IT resources within your organization rather than just having to be in somebody else's data center. Data mobility was definitely a central theme in this, uh, the ability to make data available to your on-premises applications and potentially have that same data available in cloud locations or remote locations even if the data itself remains in a central location.

So having some sort of high performance way of delivering remote access to data, particularly for unmodified applications. Another element we see is this idea of differentiated services that the hyperscalers are differentiated from one another and differentiated what from what you can deliver on premises. And each of these locations has its benefits and each of these differentiated services have their benefits, but they also have their limitations.

And this is a big driver for the hybrid multi-cloud theme that I see as as being the reality for enterprises. Often despite their best intention of standardizing on one solution, they find that they're pushed to using multiple solutions, possibly multiple clouds because they have different business requirements across the organization. Final theme was the ongoing theme of AI reshaping both storage designs as well as the operational models that we're seeing in large organizations.

So generative AI is starting to really show benefit to some organizations. As always, the future is here but it's not evenly distributed. There are people who are getting significant benefit out of generative AI applications.

Often it's some awareness that you experiment in the public cloud because it's low risk to experiment in the public cloud. The cloud is taking the, the risk on buying this large infrastructure and assuming that people will come. We see huge amounts of money being spent on these public cloud infrastructures, but often we see that the actual production use returns to on-premises, not always but often coupling that AI to the application that's on-premises actually enriching an on-premises legacy application with AI is the most beneficial element.

And so there's some interesting challenges there and around experimentation on a platform that's cloud and deployment on a platform that is actually on-premises. Uh, one of the overriding themes of course when you're stunned to go to the on-premises model is that you are buying these very expensive GPUs outright and making sure that you're actually getting value from those GPUs often means that you need the data infrastructure to feed the GPUs and particularly the multi-tenant, multiple applications running on these GPUs that we see for on-premises deployments. So AI is definitely changing the way we want to store and deliver data.

We particularly see a lot more object storage and use for AI applications, but also it's changing the way we operate. That combination of maybe experimentation, maybe production and and cloud and often production deployment for inference actually on Actually hybrid cloud and multi-cloud is the reality for enterprises. With all the complexity that comes from having hundreds of applications to over time, we do continue to see a blurring of the line between what's available on the public cloud and what can be delivered in a cloud-like manner in your actual private data center.

Another thing that struck me was the difference between the two HPE presentations. So we had both HPE, GreenLake, SAP, and HPE on ops ramp presenting and it's highlighting for me the breadth of solutions available from HPE, from software as a service to on-premises managed services and not just the the hardware that we originally think of for HPE. I was really glad to welcome TTI McGee as a first time delegate and glad to see how well she fitted into the whole delegate community.

Also delighted to have Mitch Lewis back with us as a delegate again, but even better, he was a presenter representing Signal 65 and presenting some of the work that he had done. Looking at the Intel Gaudy accelerators, really interesting cost benefit analysis in there, which we don't quite see with AI products. All the videos from Cloud Field Day are posted on the Tech Field Day YouTube channel, and we'll start seeing the delegates articles covering and their opinions of what they saw turning up on the cloud Field day webpage on the Tech Field Day website.

Thanks for watching this episode of the Tech Field Day takeaway series on the Tech Field Day plus YouTube channel. If you enjoyed it, subscribe like the video and share your thoughts on the cloud field day and the comments. Follow Tech Field Day on X, Twitter, blue sky and master it on for more updates.

And you can watch all of the presentation videos on the Tech Field Day website and YouTube channel. Thanks for watching and we'll see you next time.