Taking a system from 66 million to over 25 billion records is not for the faint of heart—especially when the system in question is financial, where precision is non-negotiable and data latency isn’t an option. In this talk, I’ll walk through the real-world journey of scaling such a system while staying sane and keeping the numbers right. You’ll get a look at how to juggle the need for ultra-high accuracy and low-latency performance, streamline your app logic, and sidestep the usual traps that come with database scaling. This session isn’t about deep-diving into internals—it’s about sharing pragmatic strategies that help you scale without drowning in complexity. Ideal for engineers and system architects eager to tackle serious scale challenges with clarity and confidence.</p> <h4>Abstract

We will explore how e-commerce projects prepare for the busiest time of the year, which key aspects to focus on, and what to expect. We’ll share our experience in setting up auto-scaling, load balancing, and discuss the loads that Silpo handles, as well as the solutions that help us navigate this season without failures.

Today, Redis is the de facto standard for caching in application services. It’s incredibly fast and easy to use, giving it a significant advantage in the market. However, as system complexity grows, Redis has several limitations that affect the reliability and availability of your system, especially under high load. In 2025, we are gradually moving away from Redis and rethinking the caching approach as a whole. In this talk, I will share my perspective on the problem and possible solutions. We will explore the challenges associated with Redis and look at alternative solutions for building more resilient and scalable systems.

<p>In this talk, we will get acquainted with TechEmpower Web Framework Benchmarks, consider generalized (programming language-independent) approaches to optimizing a web application and its environment to achieve extreme loads, and most importantly, how some of these things can be applied in practice in your projects. </p>

<p>Прямі втрати від даунтайму в 1 хвилину = $5-$10 тисяч доларів. Репутація – безцінна. </p> <p>У рамках доповіди розглянемо архітектурні стратегії, що необхідні для розробки високонавантажених фінтех-рішень. Зосередимось на використанні черг і стрімінгу для ефективної роботи та керування великими обсягами даних у реальному часі та заради мінімізації затримок. </p> <p>Особливу увагу зосередимо на архітектурних патернах, використаних при проєктуванні фінтех-системи, мікросервісах і event-driven архітектурі, що забезпечують масштабованість, відмовостійкість та узгодженість усієї системи. </p>