Testing your network before buying a VPS is a basic requirement for experienced users. However, many beginners, when choosing a VPS, often only focus on marketing terms like "CN2 GIA" and "optimized lines" on the package page, only to find after payment that the speed is abysmal during peak hours and the packet loss rate is ridiculously high. Looking Glass is the best way to strip away the "advertising filters" from service providers.

What exactly is Looking Glass?

Looking Glass literally translates to "peeping mirror." In the network operations and maintenance field, it's an online network probing tool proactively provided by the service provider, usually deployed within the provider's own data center.

In simpler terms: you open a webpage and remotely control a "test computer" located in the provider's data center, allowing it to execute network commands on any IP address from within the data center, and then display the results to you in real time.

You can think of it as seeing the world from the server's perspective. It allows you to directly execute network commands (such as Ping and Traceroute) on the provider's server, or download files from the server, testing the network quality from the server to the outside world, and testing the server's actual downlink bandwidth.

This is fundamentally different from users testing from their own local network "outside in"—it more accurately reflects the server's outbound performance, routing path, and interconnection quality with other networks within the provider's network.

Most reputable VPS providers publicly offer a Looking Glass entry point, usually located at the bottom of their website, in the Help Center, or on the "Network Status" page.

What can Looking Glass measure? Three core functions

A fully functional Looking Glass page typically offers the following testing options:

1. Ping Test: View latency and packet loss

Enter your local IP address or any target IP address in the input box on the LG page, select Ping, and click Test. LG will continuously send data packets from the data center node to the target IP, returning the average latency and packet loss rate.

A common misconception is that many people only ping a server from their own computer before purchasing, and if the latency is low, they feel confident enough to pay. However, local Ping measures the outbound journey (your home → data center), while LG Ping measures the inbound journey (data center → your home). Many so-called "optimized routes" only optimize the outbound journey; the return journey still follows congested routes—when you access a server from within China, the actual experience is determined by the return journey; a better outbound route is useless.

2. Traceroute/MTR: Examining the Routes Clearly

Traceroute displays every hop of the data packet's path from the LG's data center to the target IP. MTR adds continuous ping to Traceroute, simultaneously displaying latency and packet loss rate for each hop.

This step is crucial for distinguishing between "genuine and fake optimized routes." For example, if you're testing a CN2 GIA server, the return route will show numerous IP ranges starting with 59.43.xxx.xxx (AS4809), indicating the data is indeed using the CN2 GIA backbone. Conversely, if the return route consists entirely of nodes starting with 202.97 (163 backbone) or detours through Japan or Europe before returning, it means the provider has only performed "one-way optimization," and congestion will still occur during peak hours.

3. Speed ​​Test File Download: Verifying Actual Bandwidth

Many service providers' Looking Glass pages integrate a speed test file download function, offering several test files of different sizes (e.g., 1MB, 10MB, 100MB) for direct download. This step directly verifies the gap between "advertised bandwidth" and "actual available bandwidth"—if you use a browser or wget to download the test file and the speed consistently falls short of the theoretical value advertised in your package, be wary of potential issues such as inflated bandwidth claims or over-selling of shared bandwidth.

Why Use Looking Glass? —What Problem Does It Solve?

Ordinary users test their internet speed by simply pinging the server's IP address from their own computer. The problem with this method is that you only see the path "from your home to the server," not the return path "from the server to your home."

Internet routing is asymmetrical—you might take route A on the way there, but route B on the way back. If you only test the outbound journey, you won't be aware of any detours on the return journey. Looking Glass solves this problem precisely: it allows you to initiate tests from the server side, clearly showing the exact return route.

In addition to the above, Looking Glass can also help you:

Assess global access performance: Many large service providers deploy Looking Glass servers in multiple data centers worldwide. Users can utilize these distributed test nodes to simulate accessing target services from major regions around the world.

Identify network bottlenecks: Through Traceroute results, you can accurately identify which nodes are experiencing network bottlenecks—if the latency of a particular hop is abnormally high, it may indicate an unstable or congested node in the network.Verify bandwidth accuracy: By downloading test files, you can directly verify whether there is a discrepancy between the advertised bandwidth and the actual available bandwidth.

Hands-on Guide: Complete Pre-Purchase Testing Process

Before you prepare to pay, it is recommended to spend at least 10 minutes completing the following steps:

Step 1: Locate the Looking Glass entry point. It is usually located at the bottom of the merchant's official website, in the help center, or on the "Network Status" page. If you can't even find LG, you can directly request the test IP and LG address from customer service.

Step 2: Record the merchant's test IP. The LG page usually provides the IP address of this server; write it down for future reference.

Step 3: Perform basic local testing. Ping this IP address on your own computer and record the latency. But this is just a warm-up; the real core is in the next step.

Step 4: Use Looking Glass to test the return route. Enter your local broadband's public IP address (or any IP address in your region) in the input box on the Looking Glass page and execute MTR or Traceroute. Focus on whether the return route uses an optimized line (such as the CN2 GIA 59.43 node) and whether there are any detours or nodes with high packet loss.

Step 5: Download the speed test file. If the Looking Glass page provides a test file download, use a browser or wget to download it and see if the actual speed reaches the package's advertised value.

Looking Glass is the most worthwhile tool to use during the VPS selection process, bar none.

It allows you to clearly see before placing an order: what route the server's return route actually takes, whether the latency is stable, and whether the bandwidth is falsely advertised. Understanding these three questions will reduce your chances of making mistakes by at least 80%.

Before buying a server, spend 10 minutes using Looking Glass—this is the lowest-cost operation that ensures you are responsible for your business.

Jtti provides users with cloud servers in multiple regions, including Hong Kong, the United States, and Singapore. All servers come standard with premium CN2 GIA lines and dedicated bandwidth, and test IPs are publicly available so that every user can personally verify the network quality before purchasing.