Best of LinkedIn: Cloud Insights CW 43/ 44

00:00:00: This episode is provided by Thomas Allgeier and Frennis, based on the most relevant LinkedIn posts about cloud in calendar weeks, forty-three and forty-four.

00:00:09: Frennis enables enterprises with market, technology, and competitive intelligence for portfolio and strategy development.

00:00:16: Welcome back, everyone.

00:00:17: Today we're doing our deep dive into the cloud conversation happening across the ICT and tech world on LinkedIn these past couple of weeks, our goal, to really cut through the noise and highlight the key trends.

00:00:29: And there's definitely been a shift, hasn't there?

00:00:30: It feels like the focus isn't just on getting to the cloud anymore.

00:00:34: It's much more about, well, security control resilience once you're there.

00:00:37: Exactly.

00:00:38: So today we'll focus on three big interconnected themes.

00:00:42: that urgent need for resilience, especially after some recent outages, the whole geopolitical side of digital sovereignty, and how all this affects the big modernization projects like ERP and getting ready for AI.

00:00:53: It's right.

00:00:54: And the common thread really is control.

00:00:56: The whole industry seems to be moving past just cloud-first thinking.

00:01:00: Now it's about being cloud smart.

00:01:01: Cloud

00:01:01: smart.

00:01:02: I like that.

00:01:03: Yeah.

00:01:03: It means demanding, you know, more flexibility in your architecture, knowing exactly where your data sits legally, and definitely keeping a much tighter grip on the costs.

00:01:14: The days of just lift and shift without thinking too hard, those seem over for big companies.

00:01:19: Okay, let's start with where the rules are getting tightest.

00:01:22: Sovereign cloud and compliance.

00:01:24: It feels like sovereignty is no longer just a concept, it's becoming something you can actually measure.

00:01:29: That's a great way to put it.

00:01:30: And a big driver is the European Commission.

00:01:33: They've just put out a concrete blueprint version, one point two point one of their cloud sovereignty framework.

00:01:40: Okay.

00:01:41: This is pretty foundational because it lays out clear requirements.

00:01:45: It takes the whole discussion from sort of vague ideas to actual defined mandates you can put in a contract.

00:01:51: So how practical is it?

00:01:52: Does it really change how you'd pick a cloud provider?

00:01:54: Oh,

00:01:54: absolutely, because it introduces something measurable.

00:01:57: the Sovereignty Effectiveness Assurance Level, or SEAL.

00:02:00: That's a huge step for procurement teams.

00:02:02: SEAL.

00:02:03: Yeah.

00:02:04: People tracking this, like Max Kerbecher and Stanislaw Jarkudstynowski, they've noted this SEAL model forces you to assess providers on specific things, legal sovereignty, supply chain, even data and AI sovereignty.

00:02:18: There's even a sovereignty score that can seriously impact a tender's quality score.

00:02:22: Right.

00:02:22: That immediately brings up the whole geopolitical angle, doesn't it?

00:02:25: Especially around data access by foreign governments.

00:02:28: Precisely.

00:02:29: Michael Justina highlighted the ongoing issue with the US Cloud Act.

00:02:33: Even if your data centers are physically in Europe, relying on non-EU providers can, well, compromise sovereignty because of potential US legal access, especially for sensitive government data.

00:02:44: It's always that question.

00:02:45: who ultimately holds the legal case.

00:02:47: Exactly.

00:02:47: And it's not just Europe.

00:02:48: Lauren Zhang pointed out that Canada's sovereign cloud plans have similar challenges.

00:02:52: They need clearer rules to guard against U.S.

00:02:55: legal reach, even for data inside Canada.

00:02:58: So are organizations actually acting on this?

00:03:00: We saw a really striking example.

00:03:02: Professor Dr.

00:03:03: Dennis Kenji Kipker and Marcel Warkaftig discussed the International Criminal Court, the ICC.

00:03:09: They decided to drop some services from a major non-European provider.

00:03:13: Wow.

00:03:13: OK.

00:03:14: Yeah.

00:03:14: And they're moving towards German softwares and a socket desk.

00:03:18: That's a very clear high-profile move to boost their autonomy and cut down that risk of outside legal reach.

00:03:25: That definitely sends a message.

00:03:27: But the big cloud providers, the hyperscalers, they must be reacting.

00:03:30: They have to.

00:03:31: Ashok Kumar confirmed AWS is responding directly with plans for a European sovereign cloud, the ESC.

00:03:38: The whole point is to guarantee data stays strictly within the EU, complying with GDPR, and also the newer regulations like Dorar, the Digital Operational Resilience Act.

00:03:47: Right, Dorar.

00:03:48: It just shows how much regulation is now driving actual product development in the cloud space.

00:03:52: So

00:03:52: sovereignty covers the legal control.

00:03:55: But what about physical control and just things working?

00:03:58: The last few weeks, gave us all a bit of a mandatory lesson in resilience, didn't they?

00:04:02: That's the perfect segue.

00:04:03: Walter Zizavi put it really well.

00:04:05: Those recent AWS and Azure outages weren't just blips, they were, in his words, a mandatory resilience audit for everyone's operating models.

00:04:13: The forced

00:04:14: audit, yeah.

00:04:15: The key thing is outages are becoming something you almost have to expect to plan

00:04:20: for.

00:04:21: And the reasons for the failures are interesting too.

00:04:23: David Lenthicum and Barbara Cresty both pointed out.

00:04:26: These weren't usually sophisticated cyber attacks.

00:04:29: They were often internal issues, right?

00:04:30: Like configuration mistakes.

00:04:32: Mari Candela detailed that big AWS, US East One outage.

00:04:37: It came down to internal config problems, DNS resolution issues causing chaos.

00:04:41: Yeah,

00:04:41: and Barbara Cresty's point was pretty stark.

00:04:43: Three major outages in just ten days.

00:04:46: It shows that the cloud backbone, while powerful, can have these internal cracks.

00:04:50: Reliance on one massive system means when it hiccups internally, the impact is huge.

00:04:55: centralized failure points.

00:04:57: So given that, what's the consensus for architects now?

00:05:00: What should people be doing?

00:05:01: It sounds simple, but it's tough in practice.

00:05:03: Relying on just one cloud provider is a major strategic risk now, not just a technical one.

00:05:09: Lenny Kay made a really crucial warning.

00:05:12: Just using multi-region or multiple availability zones within one provider might not save you if the provider's own core network has a problem.

00:05:22: The failure domain can be bigger than one region.

00:05:24: And

00:05:25: there's still that basic misunderstanding sometimes, isn't there?

00:05:28: That the cloud just magically handles all your backups and security.

00:05:31: Absolutely.

00:05:32: Steve Eschweiler made a point that always needs repeating.

00:05:34: Your data is not automatically secure or backed up in the public cloud just because it's there.

00:05:39: Availability zones, regions, that's the provider's job.

00:05:42: But protecting your data, setting recovery points, that's firmly your responsibility.

00:05:48: You have to actively configure it, plan for it, can't just assume.

00:05:52: So the answer seems to be diversification.

00:05:54: Abbasad was saying true multi-cloud is basically essential now maybe keeping an anchor on-prem or in a private cloud for the really critical compliance heavy stuff.

00:06:03: And Tariq Malik had a neat frame for this, designing for resilience within sovereignty.

00:06:09: So you don't ditch sovereignty for resilience.

00:06:11: You build resilient multi-zone setups inside a sovereign cloud region.

00:06:16: You meet the legal rules and protect against local failures.

00:06:19: Okay, let's shift gears a bit.

00:06:20: How are these big strategic pressures, sovereignty, resilience, hitting the core enterprise systems, things like SAP modernization, getting ready for AI?

00:06:29: Yeah, that's where it gets really complex.

00:06:31: In the SAP world, R.W.O.N.

00:06:33: noted that architects now have this much bigger list of concerns.

00:06:36: It's not just cost and risk anymore.

00:06:38: Now they have to factor in data sovereignty, climate responsibilities, things linked to CSRD, the sustainability reporting directive, and even geopolitical tensions, as Bob Busey also pointed out.

00:06:48: it's a much wider scope.

00:06:50: Let's get technical for a sec.

00:06:51: Isabel Martin shared some really specific advice on tuning SAP H&L VMs.

00:06:56: Something about an S&C II option.

00:06:59: Why is that small detail so important?

00:07:01: Because on these big high-performance systems like ANA, picking the wrong virtualization setting, even if it seems like a small detail, can silently kill your performance.

00:07:10: Or, worse, make your setup unsupported by the vendor.

00:07:14: It just shows, even in the cloud, you still need that deep system-level technical expertise to get it right.

00:07:20: And looking ahead in SAP, Trongvantran... flagged a huge shift.

00:07:25: It sounds like IDocs are on the way out in this for on a public cloud.

00:07:28: That's right.

00:07:29: IDocs, the integration documents that have basically powered SAP integration for what, thirty years, they're disappearing in the public cloud version being replaced entirely by APIs.

00:07:39: Wow.

00:07:39: That's massive for integration teams.

00:07:41: It's a fundamental change.

00:07:42: Companies need to adapt their integration platforms and skills pretty quickly because decades of processes built on iDocs are facing obsolescence in that cloud native future.

00:07:51: Okay, switching to AI platforms.

00:07:53: Dr.

00:07:53: Mary Ann Janik observed AI is moving past the initial wow factor.

00:07:58: Now it's about the now delivering actual measurable value.

00:08:02: Better decisions, more efficiency.

00:08:04: And we're seeing ambitious timelines too, like the UK Ministry of Defense aiming for really fast tech adoption cycles, including AI in like three months.

00:08:13: And AI infrastructure is also getting tangled up with sovereignty goals, isn't

00:08:17: it?

00:08:17: Definitely.

00:08:18: Vincent Caldera highlighted India's national AI mission.

00:08:21: They're deliberately using open source collaboration, specifically things like Red Hat OpenStack and OpenShift AI to build their AI capabilities.

00:08:30: It's a strategy for technological autonomy.

00:08:33: avoiding getting locked into one vendor.

00:08:35: That's smart, controlling the foundational tech.

00:08:38: And just quickly on the pure tech side, Lars Malow mentioned some interesting research on new models like XLSTM, maybe outperforming transformers at large scale.

00:08:46: Yeah, it shows the underlying AI models themselves are still evolving really fast, which brings us neatly to the final theme, the money.

00:08:53: Ah, yes, the cost.

00:08:55: All this complexity, sovereignty, resilience, new tech, it adds up.

00:08:58: It does, and Paul Appleby noted that the need for cost predictability is making hybrid and private clouds look attractive again, especially for those steady, heavy workloads like AI training or big data analytics.

00:09:09: All right, so that's the massage point again, multi cloud helps manage risk, but also gives you negotiation leverage on cost.

00:09:16: And the spending is huge.

00:09:18: Forty percent of enterprises apparently spend over twelve million dollars a year on public cloud.

00:09:22: Now you need control.

00:09:24: Absolutely.

00:09:25: Which leads to a really critical development.

00:09:27: Nicholas Fundrini discussed embedding specific phenops clauses directly into cloud contracts.

00:09:32: Finops clauses, so more than just budget targets.

00:09:35: Much

00:09:35: more.

00:09:36: Think of it as turning financial best practices into actual contractual rights.

00:09:40: These clauses demand transparency from the provider, set clear reporting standards, and crucially they make optimization obligations enforceable.

00:09:49: Things like the provider having to help you right size instances or find idle resources.

00:09:53: If they don't help you optimize, they could be in breach of contract.

00:09:56: That's a powerful tool for cost management.

00:09:58: That is a game changer.

00:09:59: Makes phenopsis principles legally binding.

00:10:02: Exactly.

00:10:03: And architecturally, this cost-aware hybrid world is shifting where computing happens.

00:10:10: You're going to know why.

00:10:10: I noted enterprise Kubernetes isn't just one cluster anymore.

00:10:13: It's spreading multi-cluster hybrid environments and out to the edge.

00:10:17: The edge is becoming the new data center, he reckons, by twenty thirty.

00:10:22: And the ecosystem around all this is busy too.

00:10:24: Yeah, lots of activity.

00:10:26: Martin Merz emphasized that Europe's digital future needs partnership for real sovereignty, not just isolation.

00:10:33: And we see market consolidation supporting that, like Vodafone buying Scaling, Capgemini acquiring Cloud Four C. It highlights the need for specialized partners and sovereign solutions.

00:10:43: So if we pull all this together from the last couple of weeks on LinkedIn.

00:10:47: It really feels like the dominant theme is control, doesn't it?

00:10:50: Absolutely.

00:10:50: Control over data jurisdiction driven by regulations and frameworks like SEAL.

00:10:54: Control over resilience forced by those outages.

00:10:57: And financial control through things like fine ops clauses.

00:11:00: It's like the Wild West days of cloud are ending.

00:11:02: The conversation has definitely matured.

00:11:04: It's less about the rush to migrate, more about strategic governance, and really deliberate architecture.

00:11:09: Precisely.

00:11:10: If you enjoyed this episode, New episodes drop every two weeks.

00:11:14: Also check out our other editions on ICT and tech.

00:11:17: Digital products and services, artificial intelligence, sustainability in green ICT, defense tech, and health.

00:11:24: And maybe a final thought as you consider this shift from cloud first to cloud smart.

00:11:29: How are you actively testing your own cloud architecture?

00:11:32: Not just simulations, but really testing its legal and operational resilience for when that next inevitable failure does hit.

00:11:39: Something to think about for your next strategy meeting.

00:11:41: Thanks for joining us for this deep dive.

00:11:43: Don't forget to subscribe.