I know there are many tools to test for bandwidth and latency, but they all rely on having at least one host from which you can run those tests.

I wonder whether there's an online source or some other way to guestimate the latency or speed between two countries (in general).

For example, would a customer in Japan get lower latency if the server is located in Singapore or Australia? Is a user in India likely to get higher download speed from a server in the UK or in the US?

Are there any online resources or some clever ways to answer those questions with a reasonable degree of accuracy?

[UPDATE]: Thanks for the great suggestions from Raffael Luthiger. I didn't know about those looking glass servers. The submarine cable maps were also really cool to discover (Thanks to Jesper Mortensen). Also seems really wise if I could ask those network professional in the area for their experience, but obviously I don't have access to those. At least some of them are on SF :)

However, I'm still a little unsure how to combine those resources to give me some measurements. This is the information I have: Two countries (A,B). I do have IP addresses of customers in country A (I can obtain those from the web server log files for example). Presumably I can find some looking glass servers in country B and run a trace to those IPs. What's the best measurements to use? Are there any scripts that help automate at least some of this process?

  • 2
    I'd be sorry to see this one closed. Re-phrase it rather than close it. Guesstimating the network latencys between regions important for people who are building out webapps, hybrid phone-backend apps, and many other global services.
    – user2874
    Commented Feb 9, 2012 at 23:01
  • Guessing is pointless. Only real measurements will be of any value. Commented Feb 10, 2012 at 1:54
  • @JohnGardeniers I know guessing might not be accurate, but I wouldn't say it's pointless or of no value at all. There's something to be said about The Wisdom of Crowds. Hence the term guestimate :) In any case, I've updated the question to try to come up with something better than a guestimate (following from the great answers received so far).
    – Yoav Aner
    Commented Feb 10, 2012 at 10:09
  • 1
    "Estimation" is a more precise way of phrasing the question.
    – Fergus
    Commented Sep 19, 2012 at 15:33

5 Answers 5


You won't get this information just on one click. What you can try is to use many different Looking Glass servers and based on the information you get there you get some estimates. Here is a list of some looking glass servers. Another possibility is to use some traceroute servers. I recommend you to google for e.g. some famous newspapers in the regions you want to measure and then do some nslookup and with the IP addresses you get you can do some tests. This could give you some indications, but as Zoredache said it is not very reliable.

UPDATE: I doubt that you can automate those lookups. But maybe you can find a tool on this page.

  • so many tools but couldn't find any obvious candidate. Thanks for the update. I accepted your answer since it seemed the most useful. I still wish there was something more concrete, but maybe it's asking too much.
    – Yoav Aner
    Commented Feb 11, 2012 at 13:06
  • Maybe you need to learn more about BGP and routing protocols in general. Then the output of those looking glass servers will help you a lot more and it is then more clear what you have to look for. But as Jesper said. You will need Excel and some time to build up your own database. Commented Feb 11, 2012 at 17:59

There is probably no way to get an accurate estimate in a few minutes. Most downstream service providers are multi-homed, and everybody uses BGP based routing, so two individual customers who are physically very near each other may still get wildly different routing paths.

That said, you could probably create a rough guesstimate by:

  • Looking at the Submarine Cable Map. Almost all long-haul cabling is run under the sea, and this map gives you a good overview of which areas are well connected.

  • Sitting down with Microsoft Excel, a long list of looking glass servers, and building up a large'ish latency data set yourself.

  • Talking to other network professionals in the region you're interested in.


The Internet is not connected by country, almost every country has more then a single link. The link that is chosen depends heavily on both where you are connecting from, and where you are connecting too.

You could pick an ISP that is great for some users in a country, and horrible for others.

The issue is far to complex for a simple tool to give you some kind of guesstimate. The only way to be sure would be to test. You could try and get potential ISPs, to perform a trace-route for you to various network providers in the regions you are curious about.


I've been looking for similar information, but for private networks. I found this link which might be useful, it demonstrates the latency between sites on Private VPNs with Verizon. Verizon IP Latency Statistics

  • Welcome to Server Fault! Whilst this may theoretically answer the question, it would be preferable to include the essential parts of the answer here, and provide the link for reference.
    – Scott Pack
    Commented Sep 23, 2012 at 0:49

The verizon list can help or you can work with an ISP and get the SLA and MOS/MOP agreements from them.

A simple way to get a general Idea is if you can access 1 of the locations physically or remotely and I just ran a speed test from here to Dubai to see if it was close to what Verizon had in their SLA of 380ms with MOS's of 28xms and I got avg 262ms off comcast home service to Dubai's different server options.

Or if they have a public service running with an IP that isn't blocking ping's you can try that.

Or there are forumulas that you can use if you know your core network and are in control of it that take an avg ms per hop based on the lines being used or you could map it out and then the final hops from the ISP not purchase yet use the formula and take the data sheets of the firewall/ips and switching latency to calculate to the endpoint with the lowest margin of error.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .