Deep Performance Evaluation and Comparison of Cloud RDS MySQL (2024)
This article presents a detailed performance comparison of cloud RDS MySQL services across major providers, examining x86 versus ARM architectures, various storage options, and cost‑performance ratios, and concludes with practical recommendations for choosing the most suitable configuration.
On Huawei Cloud, x86 and Kunpeng (ARM) instances are priced the same, but the Kunpeng version shows a 15‑45% performance gap compared to x86. Given the strategic importance of domestically developed Kunpeng chips, this performance reduction is generally acceptable in many scenarios where x86 cannot be used.
On Alibaba Cloud, the economy (ARM) edition offers a 32% higher cost‑performance ratio than the standard (x86) edition. With 16 concurrent connections, the ARM version achieves 2,185 TPS versus 2,324 TPS for x86; pricing is 1.61 CNY/h for ARM and 2.52 CNY/h for x86, resulting in 0.74 CNY per 1,000 TPS versus 1.08 CNY, a 31.5% cost‑performance improvement.
On AWS, Graviton 2 instances provide almost no advantage over x86, contradicting the claimed 52% cost‑performance boost. Graviton 3, however, delivers a substantial performance increase over Graviton 2, and at high concurrency (e.g., 128 threads) the m7g instance outperforms the m6i, aligning with the advertised 27% cost‑performance gain.
AWS exemplifies consistency in CPU‑generation management, clearly labeling instance generations (m5.xlarge, m6i, m6g, m7g) and their architectures. In contrast, Alibaba and Tencent exhibit more noticeable performance inconsistencies due to CPU‑generation differences.
Overall, the cost‑performance balance between x86 and ARM varies by cloud platform: AWS favors x86 at low concurrency but ARM at high concurrency; Alibaba’s economy ARM tier is about 32% more cost‑effective; Huawei’s Kunpeng advantage lies in domestic chip autonomy rather than raw performance.
Architects' Tech Alliance
Sharing project experiences, insights into cutting-edge architectures, focusing on cloud computing, microservices, big data, hyper-convergence, storage, data protection, artificial intelligence, industry practices and solutions.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.