← 返回项目详情
此成果详情页已在新标签打开。查看完成后，可关闭当前标签返回原项目页，也可以点击“返回项目详情”。网页不会强制关闭浏览器标签。
成果详情 / 项目说明与报告
云实验公开摘要
原始文件名public-summary.md
- 内容说明
- 用简洁叙述汇总网络、块存储和 PostgreSQL 的记录结果，并明确单区域、单配置、短时实验和证据缺口。
- 能够证明
- 支持公开包中实际保留的性能观察值及其限制条件。
- 不能证明
- 不证明 CPU、负载、队列和延迟之间存在单一已证实因果，也不构成 SLA。
- 许可说明
- CC BY 4.0 国际署名许可（自行编写的公开叙述）
- 来源和脱敏说明
- 来自 DigitalOcean 公开派生包；只复用已有摘要，本 Goal 未访问云端或新增 benchmark。
预览
站内查看方式
以 Markdown 纯文本预览，保留结论边界。
# Public summary
The recorded lab produced approximately 1.98–2.03 Gbit/s VPC throughput, up to approximately 452 MiB/s sequential block-storage throughput, approximately 6994 read and 3006 write IOPS in the selected random test, and PostgreSQL results of approximately 10272 TPS at 32 clients and 9937 TPS at 64 clients. Average latency was 3.055 ms and 6.249 ms respectively; failed transactions were 0 in the recorded aggregate.
The result is useful as an engineering pre-study: increasing client concurrency coincided with slightly lower throughput and higher average latency in the recorded points. The available evidence does not prove CPU, load, queueing, or latency as one single root cause.
The lab is single-region, single-configuration, and short-duration. It is not a complete performance audit, a production SLA, or a guarantee for other regions or sizes. Evidence archive gaps remain.
公开包完整性
SHA-256 摘要
7fa0b38038041b82a76d2fa941b91958ce4fa580a80b02ec0662e1a42bcabeec