1 views collected around this technical thread.
This article explains Redis's LFU eviction algorithm, its implementation details, code-level mechanics, configuration steps, and how it improves upon LRU by using frequency-based counters and decay to make smarter cache eviction decisions.
This article explains the principles, implementation details, and trade‑offs of Redis’s LRU and LFU cache eviction algorithms, including their data structures, code snippets, configuration parameters, and practical guidance on choosing the appropriate strategy based on workload characteristics.
Redis implements approximate LRU and LFU eviction policies by sampling keys and using a compact 24‑bit field to store timestamps and counters, where LRU evicts the least recently accessed items and LFU evicts those with low, decay‑adjusted access frequency, each with trade‑offs for different workloads.
This article examines the storage bottlenecks in large‑scale AI training, evaluates local‑disk and Alluxio‑based distributed caching strategies, proposes uniform cache eviction and replica‑aware global policies, and introduces the SiloD framework for coordinated compute‑storage scheduling to dramatically improve GPU utilization and overall cluster throughput.
This article explains Redis eviction policies, compares strategies like allkeys‑lru, volatile‑ttl, and noeviction, and provides practical solutions for cache penetration, breakdown, and avalanche to maintain system stability under high concurrency.
This article explains Redis eviction policies such as allkeys‑lru, volatile‑ttl, and others, discusses how the eviction process works, and provides practical solutions for cache penetration, breakdown (stampede), and avalanche, including Bloom filters, mutex locks, and randomized TTLs.
The article analyzes a weekend outage of a Spring Boot microservice where all HTTP requests timed out, tracing the root cause to Caffeine's synchronous cache eviction lock being blocked by a long‑running compute operation, and proposes using AsyncCache with a dedicated thread pool to avoid similar contention.
The article explains the benefits, costs, and various update, penetration, no‑hole, avalanche, and hot‑key optimization techniques for caching in high‑concurrency distributed systems, providing practical guidance on choosing appropriate strategies based on consistency requirements and system load.
This article explains the high‑performance Caffeine Cache library, its W‑TinyLFU eviction algorithm, various loading and eviction strategies, statistics collection, and how to integrate and configure it in Spring Boot using annotations and programmatic beans.
This article provides an in‑depth overview of caching, covering its definition, when to use it, core concepts, various cache types (client, server, CDN, reverse‑proxy, in‑process, distributed), eviction policies, multi‑level cache architectures, common pitfalls such as cache avalanche, penetration and breakdown, and practical mitigation techniques.
This article explains Redis memory eviction strategies, including volatile and allkeys policies, details the approximate LRU algorithm Redis uses, and provides Java implementations of an LRU cache using LinkedHashMap and a custom doubly‑linked list, complete with code examples and configuration settings.
This guide explains caching fundamentals, purposes, and optimal use cases, details hit rates and eviction policies such as FIFO, LRU, LFU, TTL/TTI, compares client‑side (HTTP, browser) and server‑side caches (CDN, Redis, Memcached), explores multi‑level architectures, common pitfalls like avalanche, penetration and breakdown, and best‑practice strategies including pre‑warming, update patterns, and consistency management.
The article explains why caching is essential for high‑concurrency distributed systems, analyzes its benefits and costs, and then details various cache‑update, penetration‑prevention, bottom‑hole, avalanche, and hot‑key‑rebuild optimization techniques, offering practical guidance for reliable and performant backend architectures.
This article examines the role of caching in high‑concurrency distributed systems, outlining its benefits and costs, various update policies such as LRU/LFU/FIFO, timeout eviction, active refresh, and advanced techniques like cache penetration protection, no‑hole, avalanche, and hot‑key mitigation.
This article explains the role of caching in high‑concurrency distributed systems, detailing its benefits, costs, various update strategies, and advanced optimizations such as penetration protection, bottom‑hole mitigation, avalanche prevention, and hot‑key rebuild handling.
This article explains the benefits, costs, and various update, penetration, bottom‑hole, avalanche, and hot‑key rebuilding strategies for caches in high‑concurrency distributed systems, offering practical guidance on choosing eviction algorithms, using Bloom filters, and improving overall system performance and reliability.
This article explains the benefits, costs, and various update, penetration, hole, avalanche, and hot‑key optimization techniques for caches in high‑concurrency distributed systems, providing practical guidance on choosing appropriate strategies such as LRU/LFU/FIFO, timeout eviction, proactive updates, Bloom filters, and concurrency‑aware batch operations.
The article discusses the importance of caching in high‑concurrency distributed systems, outlining its benefits and costs, various update policies such as LRU/LFU/FIFO, expiration, active refresh, and advanced optimizations like penetration protection, hole mitigation, avalanche prevention, and hot‑key rebuild techniques.
The article explains why caching is essential in high‑concurrency distributed systems, analyzes its benefits and costs, and then details various cache update, penetration, no‑hole, avalanche, and hot‑key rebuilding techniques, offering practical guidance for backend engineers.
This article examines the role of caching in high‑concurrency distributed systems, outlining its performance benefits and associated costs, and detailing various cache‑update strategies—including LRU/LFU/FIFO, expiration, active refresh, Bloom‑filter protection, and solutions for penetration, avalanche, hot‑key rebuild, and the “no‑bottom‑hole” problem.