1

I am evaluating a multi-cloud setup and would like to find the closest latency efficient AWS region for GCP's us-central1 region, my options in AWS are US East (2) and US West (2). I was unable to find any service/guide that would give me this mapping.

A couple of Stack Overflow answers used manual scripts to find this mapping. Is there any resource or better way to find AWS region for a GCP region?

Amrish Kumar
  • 308
  • 3
  • 12
  • Start instances in each and see what the response times between them are. – luk2302 Jul 13 '21 at 10:28
  • 1
    Latency and bandwidth are very easy to measure at a single instance in time but very hard to predict. There are so many factors outside your control that will make most measurements useless tomorrow. Unless you know the actual routes the traffic will take now and in the future, and the details of capacity, bandwidth, and utilization, you can only generalize. – John Hanley Jul 13 '21 at 18:36
  • @JohnHanley I think the choice between US East & US West in AWS for US Central in GCP will still make a difference by 10s of ms in latency and the decision would matter. I am surprised there are no continuously monitored benchmarks for multi-cloud latency. – Amrish Kumar Jul 13 '21 at 20:44
  • Latency rarely makes a difference in real-world cloud services when you factor client -> service -> resources. There are special cases where latency is very important. However, you have not provided details whereby advice can be offered that is fact-based. At this point, I can only offer generalizations and recommendations. In other words, based upon the details in your question, either region is just fine. – John Hanley Jul 13 '21 at 20:54

2 Answers2

0

I’m not aware if there are some tool to measure multi-cloud latency, in GCP there is a tool called Performance Dashboard in the Network Intelligence menu, where you can choose up to 5 GCP regions to measure the latency among Google’s Network this could be from 1 hour to 6 weeks, the latency graph shows the median latency across all VMs deployed on Google Cloud. The vertical axis shows the median latency in milliseconds (ms), and the horizontal axis shows the change over time, I hope this information gives you a overview to make your decision and compare with AWS

arcabah
  • 71
  • 3
0

You could use the results in the aforementioned Performance Dashboard, and make a simplifying assumption that one of Google's "east" regions is close to AWS'east and similarly for "west." Additionally, if you don't have an account yet, there is a public dashboard which shows inter-region latency and throughput in Google Cloud, and you could use the same sort of assumption.

But ultimately, unless/until there is a public inter-cloud latency tool available your best bet is probably to fire-up some VMs and take some measurements. John Hanley's caveats about what you see "today" may not be what you see "tomorrow" are definitely worth keeping in mind.

Rick Jones
  • 166
  • 4