Developer’s test report shows whether AWS Singapore or Japanese VPS is better in terms of response time differences

This article is a developer’s test report aimed at exploring the difference in response times between the AWS Singapore node and Japanese VPSs, and providing selection recommendations for different user groups. The report focuses on network latency, routing stability, and concurrent response performance, aiming to provide guidance for locating Web services targeting users in Japan and Southeast Asia, while also outlining the limitations of the tests and key optimization points.

The testing utilized a combination of multiple probing methods and common tools, including ICMP ping, TCP handshake time, HTTP(S) request latency, and concurrent connection stress testing, conducted from representative locations such as Japan, Southeast Asia, and Hong Kong, China. Each test set is repeated multiple times across different time windows to reduce the impact of transient fluctuations, and the number of routing hops, packet loss rate, and jitter are recorded to compare potential differences in experience compared to the actual production environment.

Overall, requests from Japan or neighboring countries have lower round-trip latency to Japanese VPSs, resulting in generally better perceived first-packet times and first-page load times for users ； Requests to access the AWS Singapore node perform better in Southeast Asia and nearby regions. The actual difference is affected by geographical distance and intermediate network transmission routes, resulting in latency differences ranging from milliseconds to several dozen milliseconds.

The interconnection relationship between routing policies and carrier providers has a significant impact on performance. Even with a short physical distance, the quality of peering between different ISPs can still result in significant differences. Tests revealed that certain paths experienced increased jitter and packet loss due to cross-border hops or congestion in border routing. This suggests that when selecting locations, attention should be paid to the network connectivity between the target users’ primary carriers and cloud/hosting providers.

Under concurrent connection testing, nodes with sufficient bandwidth can maintain stable throughput, but latency-sensitive requests (API calls, WebSocket, real-time communication) are more affected by jitter. AWS Singapore performs stably under high-concurrency cross-regional access due to its advantages in data center resources and backbone connectivity, while Japanese VPS offers more controllable latency and a lower packet loss rate for high-concurrency scenarios within Japan.

In conclusion, there is no absolute “which one is better”; instead, the choice is based on user distribution and business characteristics: If the target users are primarily in Japan, it is recommended to consider Japanese VPS for lower latency and a better local experience ； If the target audience spans Southeast Asia or a wide regional user base, AWS Singapore usually provides more balanced response times and availability. The key is to conduct real traffic testing and A/B testing.

It is recommended to conduct multiple tests before making a decision, and to combine CDN, Anycast, edge caching, and monitoring alerts. If necessary, use multi-region deployment and traffic scheduling to achieve low latency and high availability. Continuously monitor real user experience (RUM) and synthetic test data, and dynamically optimize node selection based on users’ geographical distribution, thereby achieving long-term and stable improvements in response speed.