Sanjeev Mohan, Principal Analyst at SanjMo returns to speak with Vinay and kick off Sovereign DBaaS Decoded for 2024. The discussion touches on the evolving landscape of cloud, data sovereignty, the ongoing influence of geopolitical and economic factors driving IT strategies, as well as the technological undercurrent that is driving key decisions over where enterprises are running their data workloads.

They dive into the growing hyperscaler acknowledgement of data sovereignty through the lens of AWS’s and Microsoft Azure’s recent sovereign cloud programs and explore whether this is meaningful action of merely “sovereignty washing”. They reflect on key events of 2023, from the rise of generative AI and court rulings like Schrems II, to notable industry events like security breaches, and MariaDB’s challenges.

The conversation then looks to 2024 as they contemplate the shift from a “cloud-first” to “cloud-smart” approach, driven by economic uncertainties, evolving technology and the need for cost optimization. Sanjeev predicts a year of consolidation, speculating on how AI workloads could be brought to on-premises data. They highlight the renewed relevance of hybrid environments in potentially leveraging AI capabilities while navigating vendor lock-in concerns and evolving regulatory landscapes. The episode concludes on a positive note as Sanjeev predicts that enterprises will shift their focus back to innovation.

Key insights

⚡Hybrid cloud gets a second act

Are enterprises shifting from a “cloud-first” to “cloud-smart” approach, re-thinking hybrid models as a way to optimize for cost, compliance and control? Ironically, it is cloud technologies like containerization that are making this possible; and, although we’ll continue to see workloads move to cloud, it is likely that enterprises will take a more deliberate approach when deciding where their workloads need to live.

⚡Stop taking your data to AI, bring AI to your data

Given recent events such as the New York Times’s lawsuit against OpenAI and high-profile data breaches, the evolving privacy regulation landscape, and the commoditization of the “picks and shovels” of AI, it may be better and now more realistic to bring AI to your workload instead of the other way around. If your workload is already in the cloud, sure, leverage AI tooling there. If it isn’t, maybe not?

⚡The optimization vs. innovation pendulum swings again

A bold prediction from Sanjeev; if 2023 was a year of retrenchment and optimization, 2024 will, and in fact, needs to be a year of returning to innovation. It’s not a money problem of money; VCs have it, enterprises have it. They just need to decide where to make the bet and Vinay is bullish on a H2 return.

Episode highlights

💡 Hyperscalers release sovereign cloud solutions, is it “sovereignty washing”? [00:17:11]

AWS and Microsoft are the latest entrants in the data sovereignty space with their sovereign cloud programs. But, Vinay wonders whether this is merely “sovereignty washing”, causing unnecessary confusion when in fact, their validity is predicated on US / EU data flow agreements that are consistently challenged and overturned — as the latest most likely be.

💡AI bringing the sexy back to hybrid deployments? Yes and no… [00:26:21]

Vinay asks whether cloud-first is transitioning into cloud-smart, re-opening the door for hybrid deployments. Sanjeev agrees, caveating that we won’t be going back to the on-prem days and that there will continue to be a migration to the cloud. They discuss what is making this possible and how enterprises are seeing signs that bringing AI to their data may be a better move.

💡Reserved instances — transforming OpEx into the worst of CapEx? [00:30:52]

Vinay and Sanjeev discuss the muddy waters of committed spend. Vinay questions whether or not this transforms the OpEx model into the worst of the CapEx one, because you don’t own the asset while Sanjeev speaks to some of the difficulties he’s heard mentioned in properly estimating that spend, as well as how some comparisons between it and on-prem aren’t honest.

💡Reconsider workload control vs. benefiting from integrated, all-in-one solutions [00:37:41]

Using Microsoft Fabric as an example, Sanjeev explores the pros and cons of going with best-of-breed vs. all-in-one solutions, and vice versa. Like IaaS, the latter provides integrability, “one neck to choke” if something goes wrong, and cost savings. But again, you’re facing  lock-in and you’ll likely forgo some specialty database-specific advantages — so, it’s a trade-off.