Серверная и облачная инфраструктура

Уровень 0 · материалов: 5

В кластер включены документы об аппаратном обеспечении серверов и использовании облачных ресурсов для масштабирования, но исключены материалы об общих бизнес-моделях, инженерной культуре компаний и логистических стратегиях.

Общие признаки: масштабирование ИТ-инфраструктуры, архитектура серверов, облачные вычисления, аппаратное обеспечение дата-центров

Группа выше: Развёртывание и настройка серверов

Материалы

• #20446 — Let's Encrypt Database Server Migration to AMD EPYC — Представитель этого кластера

  Let's Encrypt перевел серверы БД на AMD EPYC — ITSumma

  Смысл: The main idea is that Let's Encrypt successfully scaled its database infrastructure by upgrading to AMD EPYC processors and NVMe storage, resulting in a ten-fold decrease in API latency and significantly increased system headroom without needing to change their software architecture (sharding).

  Let's Encrypt upgraded its database servers to AMD EPYC and NVMe SSDs, reducing API response times from 90ms to 9ms.

• #22511 — Replacing RISC with x86 Architecture in Bank Processing

  Тихая революция: внедрение x86-архитектуры вместо RISC-машин для процессинга банка — Shvydchenko

  Смысл: The main idea is that modern x86 server architectures, specifically HPE Superdome, have reached a level of performance and scalability that makes them a viable and far more cost-effective replacement for traditional RISC machines in mission-critical banking core processing.

  A major bank successfully tested replacing expensive RISC servers with x86-based HPE Superdome systems to drastically reduce total cost of ownership without sacrificing processing power.

• #3908 — Scaling Pinterest Architecture via Amazon Web Services

  Архитектура Pinterest – 18 миллионов посетителей, 10-кратный прирост, 12 сотрудников, 410 ТБ данных — globik

  Смысл: The main idea is that the use of cloud infrastructure (AWS) enables startups like Pinterest to achieve massive scalability and rapid user growth with minimal human resources and hardware overhead.

  Pinterest achieved massive scale and handled millions of users with a tiny team by leveraging AWS cloud services for storage, computing, and database management.

• #10247 — Introduction to Blade Servers and Their Advantages

  Сервера Лезвия на замену обычных. — kudzu

  Смысл: The main idea is to explain the concept, architecture, and economic benefits of blade servers over traditional rack servers, focusing on density, simplified cabling, and energy efficiency.

  Blade servers offer a high-density, modular alternative to traditional rack servers, significantly reducing cabling and energy costs while improving infrastructure flexibility.

• #11557 — Eco-Friendly Cardboard Server Chassis Development

  Экологичный серверный корпус из картона — 1mdm

  Смысл: The main idea is the introduction of an ultra-low-cost, environmentally friendly alternative to metal server chassis using recycled cardboard, aiming to lower the cost of data center services.

  A company has developed and implemented recycled cardboard server chassis to drastically reduce hardware costs and improve ecological sustainability.