Uploaded April 2026 | Updated September 2026, 2 weeks ago
This talk addresses a common misconception in Oracle application development: that larger connection pools in the backend automatically lead to better performance.
While this belief is widespread, experienced DBAs understand that oversized connection pools can actually harm performance.
The presentation provides an introduction to connection pool sizing for backend developers of all seniority levels, explaining the reasons why excessively large pools negatively impact system throughput and resource efficiency.
A key part of the discussion focuses on mathematical principles behind connection pool behavior, including Kingman's law, which models queueing delays,
to illustrate how adding more connections can increase waiting times rather than reduce them.
The talk also covers the mathematical foundations for sizing connection pools optimally based on workload and resource constraints.
By the end of the session, attendees will be equipped with practical formulas and guidelines to calculate the appropriate size for their connection pools.
This knowledge will enable them to guide application teams towards resizing their pools in a way that maximizes performance, stability, and efficient resource use,
rather than simply increasing pool size in hopes of improving it.
This talk addresses a common misconception in Oracle application development: that larger connection pools in the backend automatically lead to better performance.
While this belief is widespread, experienced DBAs understand that oversized connection pools can actually harm performance.
The presentation provides an introduction to connection pool sizing for backend developers of all seniority levels, explaining the reasons why excessively large pools negatively impact system throughput and resource efficiency.
A key part of the discussion focuses on mathematical principles behind connection pool behavior, including Kingman's law, which models queueing delays,
to illustrate how adding more connections can increase waiting times rather than reduce them.
The talk also covers the mathematical foundations for sizing connection pools optimally based on workload and resource constraints.
By the end of the session, attendees will be equipped with practical formulas and guidelines to calculate the appropriate size for their connection pools.
This knowledge will enable them to guide application teams towards resizing their pools in a way that maximizes performance, stability, and efficient resource use,
rather than simply increasing pool size in hopes of improving it.
![[VDBUH2026] Magda Miu & Alin Miu - Engineering Leadership in the Age of AI](https://i.ytimg.com/vi/DAgVdXO4NFM/mqdefault.jpg "[VDBUH2026] Magda Miu & Alin Miu - Engineering Leadership in the Age of AI
Engineering management isn’t a promotion, it’s a career shift from solving technical problems to navigating human complexity. In the AI era, this shift becomes even more challenging as leaders must make high-stakes decisions under uncertainty and guide teams through rapid change.
This talk is a condensed “management lab” for senior engineers and new managers, offering practical frameworks for building psychological safety, running effective feedback systems, coaching for autonomy, and making AI-aware decisions.
You will leave with actionable tools, not theory, to help you lead with clarity, confidence, and strategic impact from day one.")
![[VDBUH2026] Abdel Sghiouar - Optimizing LLM Inference for the Rest of Us](https://i.ytimg.com/vi/G58PbxBXC8c/mqdefault.jpg "[VDBUH2026] Abdel Sghiouar - Optimizing LLM Inference for the Rest of Us
Not every organization operates with the hyperscale resources of Anthropic, Google, or OpenAI. For the majority of businesses integrating Large Language Models (LLMs) into their critical paths, the high costs and scarcity of GPU/TPU accelerators present a significant challenge. Striking the balance between performance, availability, scalability, and cost-efficiency is a must.
While Kubernetes is a ubiquitous runtime for modern workloads, deploying LLM inference effectively demands a specialized approach. This session dives deep into practical strategies for optimizing your Kubernetes clusters and LLM Inference workloads to run efficiently and cost effectively. We will explore:
– Container and Model Optimization
– Accelerator Management
– Data & Storage
– Network & Load Balancing
– Observability
Attendees will leave with practical techniques for maximizing cost/performance for LLM inference for their AI-powered applications on Kubernetes.")
