This article summarizes the InfiniFS paper, detailing how its access‑content decoupling, speculative path resolution, and optimistic metadata caching enable efficient metadata handling for data‑center‑scale file systems supporting billions of files.
This article presents a comprehensive overview of Zilliz’s cloud‑native vector database ecosystem, detailing Milvus’s distributed architecture, indexing and query capabilities, related tools such as Towhee and GPTCache, storage challenges, tiered storage designs, performance metrics, and real‑world AI use cases like code‑assist and RAG‑based Q&A systems.
The article examines how long it would take to transfer a 1,000 TB dataset from Nanjing to Beijing using various network options—home broadband, enterprise data lines, and internet lines—calculates the required time and cost, and ultimately proposes shipping hard drives by high‑speed rail as the fastest, cheapest solution.
This article examines the challenges of storing massive DMP data in Redis, analyzes memory fragmentation, key‑value explosion, and latency constraints, and presents practical solutions such as eviction policies, bucket hashing, key compression, and fragmentation reduction to enable efficient in‑memory storage.
This article examines the challenges of storing and querying billions of DMP identifiers in Redis, analyzes data characteristics and memory fragmentation issues, and presents practical solutions such as eviction policies, bucket‑based key hashing, and fragmentation reduction techniques to achieve low‑latency, large‑scale caching.
This article examines the challenges of storing massive DMP ID mappings in Redis—including memory fragmentation, expansion, and latency constraints—and presents eviction, bucket‑hashing, and fragmentation‑reduction techniques to achieve efficient, real‑time, large‑scale key‑value storage.
This article examines the challenges of storing massive DMP data in Redis, analyzes memory fragmentation, key‑size issues, and latency constraints, and presents practical strategies such as TTL eviction, bucket‑hashing, custom key compression, and fragmentation‑reduction techniques to enable scalable, real‑time querying.
The article analyzes the massive storage and performance challenges of a real‑time DMP cache built on Redis, outlines data characteristics and technical obstacles, and proposes eviction policies, bucket‑based hashing, and fragmentation‑reduction techniques with Java code examples to achieve billion‑scale in‑memory key‑value storage.
58.com tackles the challenge of storing and searching billions of heterogeneous posts by employing a unified post center, a category‑attribute service, and an external search engine, using vertical table splitting, JSON‑based extensible fields, compressed keys, and horizontally sharded indexes to achieve massive scalability and high throughput.
The article recaps the 2017 Pika community salon in Beijing, explains Pika’s design as a high‑capacity Redis‑compatible storage, showcases real‑world deployments at 360, Huanxin and Weibo, and shares insights from four technical talks along with event highlights and resources.
QQ Album, China’s largest photo service with over 2 trillion images and 300 PB of storage, achieves massive scale through a layered architecture that combines MySQL indexing, a custom TFS‑based KV store, SSD‑optimized TSSD, aggressive compression, regional zones, CDN acceleration, and near‑upload strategies for low latency and high availability.
