Best of LinkedIn: Cloud Insights CW 41/ 42

00:00:00: This episode is provided by Thomas Allgaier and Frennis based on the most relevant LinkedIn posts about cloud in calendar weeks, forty one and forty two.

00:00:07: Frennis enables enterprises with market technology and competitive intelligence for portfolio and strategy development.

00:00:14: And for this deep dive.

00:00:16: Our mission is pretty clear.

00:00:18: We're looking at how the whole cloud conversation has really matured.

00:00:21: We've pulled together the key insights, the actionable stuff that ICT and tech pros were sharing.

00:00:27: Exactly.

00:00:28: It feels like the focus shifted.

00:00:29: You know, it's not if you should use the cloud anymore.

00:00:32: Right.

00:00:32: It's how you actually manage the complexity, the architecture, the governance.

00:00:36: Yeah,

00:00:37: how you master it.

00:00:38: Absolutely.

00:00:39: Yeah.

00:00:39: I mean, the chatter online suggests cloud isn't just some optional extra now.

00:00:43: It's foundational.

00:00:44: It underpins everything, right?

00:00:46: Critical services, AI.

00:00:47: Everything.

00:00:48: So we're gonna dig into four big themes from those weeks.

00:00:51: And we'll start with, well, kind of a painful lesson in resilience.

00:00:54: Okay, let's unpack this.

00:00:55: Theme one, architecture and resilience.

00:00:58: Yeah, resilience is a good place to start because honestly, for a lot of companies, it went from theory to like urgent reality almost overnight.

00:01:06: There was a big incident then that really showed how fragile things can be if they're too concentrated.

00:01:11: Oh yeah.

00:01:12: That AWS Assist One incident, people like Elliot Green and Nankishore Kander were all over it.

00:01:19: And it wasn't small, was it?

00:01:20: No, not at all.

00:01:21: It caused massive internet problems, hit big names like Snapchat, Venmo, Fortnite, Reddit.

00:01:27: but also, you know, Delta Airlines, even U.S.

00:01:29: government services.

00:01:30: It was a huge wake-up call.

00:01:32: And what struck me was how fast the discussion shifted.

00:01:34: It wasn't just complaining about downtime.

00:01:36: It was immediately diagnosing the architecture.

00:01:39: The consensus was, look, relying on one single region, even a giant like us East One, it's just, it's fragile.

00:01:45: Folks like Rob Meyer were really hammering home the need for better design.

00:01:49: Cellular architecture, multi-region failover, that became the focus.

00:01:53: That cellular design point is interesting.

00:01:55: It's about labeling the damage.

00:01:56: right?

00:01:57: So one regional issue doesn't crater your whole global setup.

00:02:00: Exactly.

00:02:01: And Alexander Apolikov added something crucial too about planning your deployments properly.

00:02:06: Specifically, making sure your backup telemetry, the stuff you need to debug, isn't stored in the same place that just went down.

00:02:14: Yeah.

00:02:14: That makes sense.

00:02:16: You need to be able to see what's wrong even when the main system is blind.

00:02:19: That's the kind of detail that matters.

00:02:21: But it raises that question, you know, who's responsible when the provider has a regional outage?

00:02:27: Is it all on them?

00:02:29: Well, the post suggests the thinking is definitely shifting.

00:02:33: Ray Krueger made a great point.

00:02:34: Enterprises have to check the vendor's cross-region recovery capabilities before anything happens.

00:02:40: You can't just blame the provider after the fact.

00:02:43: Good business continuity planning, that's on you.

00:02:44: It's part of the design now.

00:02:46: Exactly.

00:02:46: And that's where multi-cloud really starts to make strategic sense, like Mario Candela and Cernschmans pointed out.

00:02:52: It's not just about price anymore.

00:02:54: It's a hedge, a real hedge against this concentration risk.

00:02:57: Avoiding that single point of failure is just... core business continuity now, full stop.

00:03:03: Which

00:03:03: explains why vendors like Snowflake, Emery Esdemarlar mentioned this, are actively promoting their cross-cloud resilience features.

00:03:10: It's a selling point.

00:03:11: Definitely.

00:03:12: And how do you manage these complex multi-region setups without going crazy?

00:03:17: Automation.

00:03:17: Right.

00:03:18: Elkhan Y brought up infrastructure as code, mentioning plume.

00:03:22: If you need fast, standardized recovery, you have to automate it.

00:03:25: reduces human error, shrinks that blast radius in a crisis.

00:03:29: Automation enables resilience.

00:03:31: And resilience, well, that allows you to tackle the next big thing.

00:03:35: Global governance and control.

00:03:36: Okay, yeah, theme two.

00:03:38: This is where it gets really interesting, moving beyond just tech into, like, geopolitics.

00:03:43: Sovereignty.

00:03:44: Governance and global deployment.

00:03:46: Cloud sovereignty isn't just a compliance checkbox anymore, is it?

00:03:50: It's strategic.

00:03:51: Darryl Meistow even mentioned Google.

00:03:52: Cloud sees it as a competitive advantage.

00:03:55: It's fascinating because sovereignty now means more than just data residency, you know, where the bits and bytes physically sit.

00:04:01: It's fundamentally about control, decision rights.

00:04:04: The conversation is demanding open APIs, real portability, ways to avoid vendor lock-in.

00:04:11: Walter Zerzavi talked about frameworks like German... So

00:04:18: wait, if sovereignty means control, how does that work when the big U.S.

00:04:23: hyperscalers start offering quote-unquote sovereign cloud solutions in Europe or elsewhere, isn't that?

00:04:30: Contradictory.

00:04:30: It's a really smart strategic move.

00:04:32: Actually, it shows they understand they have to meet these national requirements for operational autonomy for data jurisdiction, even if they technically still own the infrastructure.

00:04:41: Okay, so they're adapting.

00:04:43: They

00:04:43: have to.

00:04:43: And we saw a ton of activity showing this.

00:04:45: Barbara Cresty and Nolan Goddard highlighted that perceived sovereignty gap in Europe, this feeling of being too dependent on non-European providers.

00:04:52: And the providers responded fast.

00:04:54: Catherine Renz and Ashad Kumar were talking about the new AWS European Sovereign Cloud.

00:05:00: And Oracle's pushing hard, too.

00:05:02: Jesper Kongsdahl Sörensson pointed out the specific architectural guarantees their EU sovereign cloud offers.

00:05:08: They're all scrambling to meet those tough European demands.

00:05:11: This is not just the giants, right?

00:05:12: Fabian Poonan mentioned scaleway, aiming to be a sovereign leader in Europe, but focusing on transparency.

00:05:19: not just scale, a different angle.

00:05:21: Absolutely.

00:05:22: And look at the Gulf region, they're treating Sovereign Cloud like a national priority.

00:05:26: Miguel Vega mentioned Oracle launching the UAE's OneCloud.

00:05:30: And Core Forty-Two launched that self-serve Sovereign.

00:05:33: platform, right?

00:05:34: Tug Bikati noted that.

00:05:35: And then Carrington Mullen highlighted, they immediately landed a big deal to buy taxi company, moving fast.

00:05:40: Super fast.

00:05:41: Asia's doing the same.

00:05:42: Balaji Rajagopal shared news about Datta Samudra launching India's Sovereign Cloud.

00:05:46: Nicholas Routu noted SoftBang partnering with Oracle for Japan's AI and Sovereign Cloud strategy.

00:05:52: It's everywhere, even Canada.

00:05:53: Pippa Norman mentioned their push for data independence.

00:05:56: So the takeaway is sovereignty is tied directly to sensitive sectors, defense, government, public health, keeping critical national data under national control.

00:06:04: Yeah.

00:06:04: And we saw proof of that focus, like Becca Coombs noting the rising demand for GCP skills specifically within UK defense and government cloud projects.

00:06:14: It's very targeted,

00:06:15: which leads perfectly into how all this sovereign resilient infrastructure is actually being used because it's all getting.

00:06:22: tangled up with AI.

00:06:24: Right.

00:06:24: Theme three, AI on cloud from pilot to production.

00:06:29: Feels like the AI infrastructure talk got serious really quickly.

00:06:32: We're past the what if

00:06:33: statement.

00:06:33: Oh, yeah, it's now.

00:06:34: How do we run this stuff?

00:06:35: Agenetic AI, generative AI, reliably and at scale.

00:06:39: In production.

00:06:40: Exactly.

00:06:40: And because the data involved is often so sensitive, you need secure, compliant environments.

00:06:45: Walter Zizavi made the point about Agenetic AI in Germany.

00:06:48: People prefer private or sovereign clouds for that because compliance is absolutely critical for those real-time decision systems.

00:06:54: So it's about building these specific AI ready stacks now.

00:06:57: We saw launch is reflecting that, like Edvinia Group's multi-tenant sovereign AI cloud that our Peta Chowdery and Hedgester mentioned aimed right at Northern Europe's data residency needs.

00:07:08: And Core Forty-two and the Gulf really went for it.

00:07:10: Mark Dominic and Carrington Malin talked about their pay-as-you-go sovereign AI cloud using NVIDIA H-One-Hundred GPUs.

00:07:18: That's serious compute power designed for scalable AI in the UAE.

00:07:22: The H-One-Hundred mention is key.

00:07:24: Yeah, that's not just playing around.

00:07:26: That's production grade AI infrastructure.

00:07:28: And look at the partnership Satyam Dash highlighted SAP and open AI launching sovereign open AI for Germany

00:07:35: tailored for the public sector secure compliant AI.

00:07:38: the market is really specializing

00:07:39: and the tooling is catching up to.

00:07:41: so McBeharra mentioned SuperMicro's SuperCloud software suite software designed just to manage these super complex AI data centers.

00:07:50: So you've got the sensitive AI running on expensive, sovereign, maybe multi-cloud setup.

00:07:55: Yeah, sounds like a recipe for huge bills if you're not careful.

00:07:58: Exactly right.

00:07:59: The thinking now is that AI performance isn't just speed, it's reliability, it's evaluation, oversight.

00:08:05: Things like our rag architecture is choosing the right model.

00:08:08: Those aren't just academic exercises anymore.

00:08:10: They're massive, costly system design choices.

00:08:13: Built on that foundation of resilience and control we talked about earlier.

00:08:16: Precisely.

00:08:17: Which brings us neatly to the final theme, managing all this complexity and cost.

00:08:22: Theme four, fine ops, data hygiene, and operating discipline.

00:08:27: Yeah, with everything getting so complex and expensive, the basics suddenly matter a whole lot

00:08:33: more.

00:08:33: We're definitely seeing a renewed focus on just good operating discipline, driven by cost mostly.

00:08:39: I mean, when you're firing up H- one hundreds across sovereign regions, small mistakes get very expensive very fast.

00:08:45: That's why you see people like Sanjit Raj constantly reminding teams, shut down your test environments, clean up your AWS resources, those surprise bills from forgotten dev instances.

00:08:56: They're way bigger now.

00:08:57: Basic hygiene isn't optional.

00:08:58: And the platforms are trying to help.

00:09:00: Nicholas Fondrini noted IBM Cloud is building and fine ops tools to help manage cost across hybrid setups.

00:09:06: You know, where you've got on-prem stuff mixed with public cloud, central cost control becomes vital there.

00:09:11: And it's not just cost discipline, it's data discipline too.

00:09:14: Eric Clark had that stark warning about S three backup practices.

00:09:17: Yeah,

00:09:17: it seems basic, but

00:09:18: but if you neglect it, your cyber resilience just evaporates.

00:09:23: Especially with critical data now sitting in these new sovereign zones, you have to get the fundamentals right.

00:09:28: It really all comes back to culture, doesn't it?

00:09:31: Max Gull had this great insight, thinking of cloud transformation as a learning discipline.

00:09:35: Oh, so.

00:09:36: Well, the idea that teams do best when they kind of embrace feeling like they don't know everything, when they challenge assumptions constantly, that continuous learning culture is what you need for a mature cloud operation.

00:09:48: That makes sense.

00:09:49: And you see that learning happening across the ecosystem.

00:09:52: too, right?

00:09:53: Like that Virtuoso Weymere Link partnership in Japan that Ahmed Omni shared, focusing on regional delivery.

00:09:59: People are sharing best practices on phenops, multi-cloud compliance.

00:10:04: Because they have to.

00:10:05: so if we wrap this all up for you the listener the cloud conversation just in these two weeks It's clearly gotten way more mature.

00:10:11: We're way past just migrating stuff.

00:10:13: Yeah, we're deep into mastering complex architecture dealing with serious geopolitical governance, especially sovereignty and treating cost and just basic hygiene as critical operational needs particularly with demanding AI workloads coming online.

00:10:28: It's all about control, cost, capacity, all at the same time.

00:10:32: The bottom line for strategy is, you can't just let your cloud run on autopilot anymore.

00:10:37: It needs active, really disciplined management.

00:10:40: Which leaves us with that final thought from David Lenthigam.

00:10:43: He noted that a lot of cloud predictions people maybe scoffed at years ago, they're coming true now.

00:10:48: and they're fundamentally changing enterprise strategies.

00:10:52: So the question for you is, how quickly are you challenging your own past assumptions about the cloud?

00:10:56: Are you integrating these new architectural realities these governance needs fast enough?

00:11:01: If you enjoy this deep dive, new episodes drop every two weeks.

00:11:04: Also check out our other editions on ICP and tech, digital products and services, artificial intelligence, sustainability and green ICT, defense tech and health.

00:11:13: Thank you for joining us as we unpack the insights from the industry.

00:11:17: Make sure you subscribe so you don't miss our next deep dive into the source material.