When you order a server, you really want one thing: that it feels fast from where you are and where your users are. A simple Looking Glass latency test and network speed test lets you check that before you commit to any hosting plan.
In the hosting industry, this is one of the easiest ways to compare locations and providers without touching your production setup. With just a browser (and maybe wget), you can see how stable, how fast, and how consistent the network to your future server actually is.
Think of a Looking Glass page as a public test bench for a hosting provider’s network.
You get:
One or more test IP addresses
Buttons or links to download test files of different sizes
Tools like ping, traceroute, sometimes MTR
You sit at your laptop, open the page, pick a server location, and start poking it from your own network. No logins, no contracts, just raw numbers.
Most serious hosting companies, including GTHost, provide some kind of network test or Looking Glass so you can see real latency and speed to their data centers before you deploy anything.
You do not need much:
A web browser
Optional: wget or a similar command line tool if you like terminals
That is it. No special software, no complicated setup.
First, grab the Looking Glass page from your hosting provider. You will usually find it linked in their footer or docs as “Network test,” “Looking Glass,” or “Speed test.”
On that page you typically see:
A dropdown or list of server locations (e.g., US, EU, Asia)
A field to type a domain or IP
Buttons for tools like ping and traceroute
Links or buttons to download test files, like 10 MB, 100 MB, 1 GB
Pick the location you care about most, for example the data center closest to most of your users. We will test against that one.
Latency is “how long a message takes to go back and forth.” Lower is better. It is what makes things feel snappy or sluggish.
On the Looking Glass page:
Find the “Domain name or IP address” field.
If there is a “My IP” shortcut, click it to auto-fill your public IP.
Or type the domain/IP of a server you already have.
Click the ping button.
You will see a few lines of output scroll by: each line shows how many milliseconds (ms) it took for a packet to go there and back. At the end you get an average.
If you are doing this locally on your own machine instead of the web UI, it might look like this:
bash
ping -c 10 203.0.113.10
Look at the avg value. That is your typical latency to that location. For example:
~10–30 ms often feels almost “local”
~50–100 ms is usually fine for web apps and APIs
~100+ ms you start to notice lag with interactive stuff like SSH or gaming
If your ISP and your hosting provider both support IPv6, it is worth testing that too.
On the Looking Glass:
Enter your IPv6 address or a server’s IPv6.
Choose ping6 or the IPv6 ping option.
Run the test and compare averages.
From the command line, it could be:
bash
ping6 -c 10 2001:db8::1234
Many modern networks give slightly better paths over IPv6. If the numbers differ a lot, that tells you something useful about how the provider’s IPv4 vs IPv6 routing works.
Now we care about throughput, not just delay. That is your download speed.
On the Looking Glass page you will see test files like:
10 MB
100 MB
1 GB
Pick one based on your connection:
Small file: just a quick “is it alive” check
100 MB or 1 GB: better for seeing if speed stays stable over time
Click the file size button and watch your browser’s download speed. If your home connection is 100 Mbps, the maximum you should expect is about:
100 Mbps / 8 = 12.5 MB/s (roughly)
So if you see around 10–12 MB/s sustained from that test server, you are pretty close to your line speed. If it jumps between, say, 2 MB/s and 12 MB/s, that tells you the route or the network is not very consistent.
This kind of test is simple, but it answers the question everyone cares about: “Will my users actually get the speed I am paying for?”
If you prefer terminals (or you just like accurate numbers), use wget or a similar tool instead of the browser.
Take the URL of the test file from the Looking Glass page and run:
bash
wget -O /dev/null "http://example.com/test-1GB.bin"
Here -O /dev/null means “download the file but throw it away,” so you do not fill your disk.
While it runs, wget prints the current and average speed, often in MB/s. You can let it run for a bit and see if the speed:
Quickly ramps up and stays flat (good)
Keeps dropping and climbing (not great)
Never gets anywhere near your line speed (also not great)
You can repeat the test for different data centers to see which region gives you the best combination of latency and throughput.
If you do not have a server yet and just want to test a provider’s network properly, it helps to have something you can spin up fast and discard later. That is exactly the kind of situation where instant-deployment providers shine. 👉 Try GTHost’s instant dedicated servers and test their low-latency network from your location in minutes. After that, rerun the same ping and wget tests and see how it feels in real use.
When you are done, you will have a little list in your head:
Location A: 25 ms, 11 MB/s, very stable
Location B: 90 ms, 8 MB/s, a bit jumpy
Location C: 150 ms, 5 MB/s, sometimes spikes
From there, you can decide:
Which data center is best for your main users
Whether you want a second location for another region
If the provider is good enough, or if you should look at alternatives
You are not guessing any more. You have real numbers from your own connection and tools you understand: ping, browser downloads, wget.
Do I need special software to use a Looking Glass?
No. A browser is enough. wget or curl just give you more detailed speed readings.
Is one test enough?
Not really. Run tests at different times of day. Evening peaks can look very different from early morning.
What is a “good” latency?
It depends on your app. Under 50 ms feels great for most interactive work. For general web traffic, anything under 100 ms is usually fine, as long as it is stable.
A Looking Glass page is a simple way to test latency and network speed to a hosting provider’s servers before you move any real workloads. With a few pings and some download tests in your browser or with wget, you can see which data center actually feels fast from your location.
If you need low-latency, stable hosting you can spin up quickly just to test, 👉 why GTHost is suitable for low-latency global hosting scenarios comes down to instant dedicated servers, multiple locations, and an easy way to verify real network performance with the exact steps we just walked through.