AWS cost optimization and what thousands of accounts taught us en

Most teams run a cost optimization once, save some money, and move on. Nobody owns the follow-up. The findings age. The waste comes back.

We manage tens of thousands of AWS accounts globally as a distributor. Not a handful of large enterprises, but thousands of small and mid-sized businesses. That volume reveals what individual accounts never can: where cloud spend actually drifts, which fixes stick, and which ones quietly fail because nobody was responsible for keeping them alive.

This session maps cost optimization to four pillars and to the people who should own them. Operational hygiene belongs with developers and DevOps: outdated instance generations, unattached resources, right-sizing that never happened. Architecture belongs with architects: Graviton, network cost traps, compute model choices that shape your bill long before anyone looks at it. Commitment strategy belongs with FinOps and procurement: not just EC2. And the ecosystem layer belongs with leadership: what changes when you operate at distributor scale.

Real project examples throughout, savings from 11% to over 70%. What you take home is something you can map onto your own organization tomorrow morning: who owns what, where the biggest levers sit, and why repeatable process beats a one-off cleanup every single time.

Using Amazon Bedrock at scale: How we overcame quota limits without Provisioned Throughput en

Amazon Bedrock provides easy access to some of the best foundation models for GenAI projects. However, due to high user demand and limited hardware capacity, AWS unexpectedly reduced quotas in late 2024. Anthropic's models in regions like Frankfurt were particularly affected, causing serious problems for many customer projects due to very low quota limits on requests and tokens per minute.

In this talk, I'll share real-world experiences from several customer projects, highlighting practical solutions we've found to successfully scale the use of foundation models without relying on expensive provisioned throughput. I'll discuss specific techniques we've used to increase seemingly fixed quota limits by more than ten times. I'll also show how choices in foundation model selection, AWS regions, and architectural setups - such as multi-account structures and cross-region inference - can significantly improve capacity and reduce costs.

By the end of this session, cloud architects, engineers, and AI developers will have clear insights into how to optimise their Amazon Bedrock architecture. You'll learn straightforward methods for supporting cost-sensitive but robust GenAI workloads, even in demanding production environments.

Submitted for AWS Community Lounge at AWS Summit Hamburg 2025.

Empowering AWS Partners: Secure and Optimized AWS Account with Ingram Micro InstantFramework for AWS en

Ingram Micro Germany has developed the InstantFramework for AWS (IMIF). This framework focuses on optimizing single AWS accounts according to AWS/CIS Best Practices, with a strong emphasis on security. It is specifically designed for new AWS (Registered/Select) Partners who often lack AWS expertise and do not have their own landing zones.

Ingram Micro, as an AWS Distributor, enables AWS Partners to grow revenue opportunities by reselling AWS products and solutions. AWS Partners leverage our go-to-market support, offload billing and operations, receive funding benefits, and gain technical expertise. Since Q2/2024, all new AWS resellers (SPP) must select a distributor, the extended arm of AWS.

AWS partners typically need to set up AWS accounts quickly, but blank/empty accounts are not secure enough for production workloads. Common issues include auto-assign public IPs, open default security groups, lack of persistent CloudTrail, no AWS Config, and no default encryption...
The IMIF, developed with AWS CDK v2 for Python, addresses these issues, making the AWS ecosystem more secure for our partners. The IMIF MVP was successfully launched at the AWS Partner Summit Berlin 2024.

In this session, I would like to share the journey of building this framework, from gathering partner feedback (market needs) to delivering a ready solution and a first success story. I will show how innovation leads to more secure AWS environments.
The presentation will not only be theoretical but will also dive into the "meta-level", explaining why we chose IaC and CDK over Terraform. I will discuss the technical issues faced with CDK and the development (this is level 300), challenges with AI-assisted-coding, with releasing as CF templates, and more. Besides the presentation, the IMIF will also be demonstrated in a live demo (prepared accounts due to time constraints).

Submitted for AWS Community Day DACH 2024.

AWS European Sovereign Cloud and Digital Sovereignty en

In this session, we’ll explore the concept of Digital Sovereignty and how it is being realized through AWS’s upcoming European Sovereign Cloud (ESC) — a cloud infrastructure sovereign by design.
With a major investment of $7.8 billion, AWS positions the ESC as a key enabler for organizations that prioritize data sovereignty, regulatory compliance, and digital trust in a shifting geopolitical landscape.
The European Sovereign Cloud is built to meet the specific needs of public sector and highly regulated industries within the EU. It ensures that all customer data — including metadata — stays entirely within the EU, in full compliance with European laws. ESC will be completely separate from the global AWS infrastructure, featuring independent IAM, billing, and support systems.
Operations, technical support, and customer services will be carried out exclusively by EU-based personnel. The first region is scheduled to launch by the end of 2025 in Brandenburg, Germany.
We’ll also take a closer look at the legal and operational structure of the ESC, and what sets it apart from other sovereign cloud offerings.
This session will equip you with all the essential information to understand, evaluate, and prepare for the launch of the AWS European Sovereign Cloud.

First delivered on July 7th at the AWS User Group Nuremberg Meetup.
The session runs for approximately 45 minutes, but due to high interest and frequent Q&A — especially from attendees outside the AWS ecosystem — it's best to plan for a full hour.

Government Requests to AWS: Data Disclosure in Practice en

Government access to cloud data is one of the most debated topics in the public sector today. In this talk, we’ll explore how AWS handles official data requests in practice, clarify what the U.S. CLOUD Act really means, and separate common myths from legal and technical facts. You will learn how requests are processed exclusively through the Amazon Law Enforcement Request Tracker (LER), which types of authorities can submit them, and how formal international procedures—such as MLATs, Hague Convention channels, and emergency cooperation under the Budapest Convention — apply to requests coming from Germany and the EU. The session also explains the technical safeguards that protect customer data, focusing on AWS Key Management Service (KMS) and why AWS cannot access customer-managed encryption keys, and it highlights the AWS Digital Sovereignty Pledge as a framework for data locality, access controls, and operational independence. Finally, we’ll look at how the new AWS European Sovereign Cloud enforces strict data locality, EU-only access, and full independence from U.S. entities, ensuring compliance with European sovereignty and privacy expectations. By the end of this talk, you will have a clear understanding of how AWS manages government data requests securely and transparently — and what this means for digital sovereignty in Europe.

Level 200. Initial delivery at the pre-evening event for the AWS Public Sector Partner Executive Briefing DACH 2025 in Munich on November 5, 2025. The duration is around 20 minutes, but there are typically many questions, even during the talk. Best to aim for 45 minutes.

Wenn die Wolke wackelt - und die Multi-Cloud Diskussion von vorne beginnt de

Ende Oktober traf es gleich zwei Hyperscaler: Bei AWS (20.10.) und Microsoft (29.10.) kam es zu global spürbaren Einschränkungen – von Collaboration-Diensten bis Streaming. Was folgte: hitzige LinkedIn-Debatten über Multi-Cloud als Allheilmittel.
In diesem Impulsvortrag ordne ich die Vorfälle ein, entmystifiziere „Cloud A fällt aus, Cloud B übernimmt nahtlos“ und zeige, warum portable Multi-Cloud in der Praxis häufig mit hohen Kosten, Komplexität und Abstraktions-Trade-offs bezahlt wird (Stichwort: geringster gemeinsamer Nenner). Danke an Gregor Hohpe für die Grundlagen zu Multi-Cloud-Architekturen.
Stattdessen diskutiere ich einen pragmatischen Resilienz-Pfad: Multi-Partition auf AWS – inkl. AWS European Sovereign Cloud (ESC) – um Ausfall-Domänen zu trennen, ohne auf Managed/Serverless-Dienste zu verzichten oder doppelte Ops-Welten zu betreiben. Kurz: Resilienz erhöhen, ohne die üblichen Multi-Cloud-Nachteile.

Level 200–300. Erstvortrag beim Pre-Evening-Event des AWS Public Sector Partner Executive Briefing DACH 2025 in München am 5. November 2025. Dauer: 20–30 Minuten – abhängig von Publikum und gewünschter technischer Tiefe.