High-performance, highly reliable data storage: CBS efficiently supports hot migration of cloud servers, preemptively avoiding business interruptions caused by physical failures. CBS provides three copies of data redundancy, complete data backup, snapshot, and second-level data recovery capabilities. CBS is suitable for high-load, core critical business systems.
Elastic scaling: CBS can be freely attached and detached within the same availability zone without shutting down/restarting the cloud server. The capacity of CBS can be elastically configured and scaled as needed.
Massive data analysis
The typical Spark-HDFS offline data analysis framework performs sequential I/O for RDD read/write and shuffle write, while only shuffle read I/O is random I/O, with sequential I/O accounting for up to 95%. CBS's multi-threaded concurrent throughput performance is excellent, efficiently supporting Hadoop-Mapreduce, HDFS, Spark, and TB/PB level data offline processing capabilities.
In multi-disk concurrent scenarios, the throughput performance of a single HDFS cluster can reach 1GB/s.
Large enterprises such as Xiaohongshu, Giant Network, Ele.me, Yoho!BUY, Weipiao, and others have extensively carried out data analysis, mining, business intelligence, and other big data practices on CBS.
Deployment Environment: Five cloud servers with 12 cores and 40GB RAM each, simulating offline data analysis of 1.5TB data volume.
Performance Testing:
Each CVM has one 1TB HDD cloud disk mounted. Five HDD cloud disks provide a read speed of 500MB/s, allowing data to be read to memory in 50 minutes.
Each CVM has one 1TB SSD cloud disk mounted, allowing data to be read to memory in 25 minutes.
Core databases
SSD Cloud Disk is suitable for scenarios that require high I/O performance and high data reliability. It is particularly suitable for medium to large relational database applications such as PostgreSQL, MySQL, Oracle, SQL Server, core business systems with high I/O intensity and high data reliability requirements, and medium to large development and testing environments with high data reliability requirements.
Balancing data reliability and high-performance, SSD Cloud Disk has provided reliable support for large enterprises such as Blade & Soul, Ask Tao, Yoho!BUY, WePiao, and Xiaohongshu.
Deployment Environment: Four cloud servers each with 4 cores and 8GB RAM, each mounted with an 800GB SSD CBS, running MySQL version 5.5.42.
Performance Test: Using sysbench to simulate OLTP performance test, with a test dataset of 10 million records, TPS can reach 1616, and QPS can reach 29,000. A single disk is sufficient to support tens of thousands of online transactions per second.
