Ping and Latency Explained: Response Time, Delay and Timeouts
Latency is the delay involved in communication between systems and is commonly expressed in milliseconds. One measurement describes one path and method at one moment; it does not determine overall connection quality by itself.
How it works
This site's Ping / Latency Test measures how long its server takes to establish one TCP connection to port 443 on a selected public hostname or IP address. It is not a traditional ICMP ping from your browser or device, so its result should be interpreted as a server-to-target TCP connection observation.
Traditional ping and this TCP test are different
Traditional ping normally sends ICMP echo requests and measures the time until corresponding replies return. Operating systems can repeat those requests and report individual round trips, packet loss, and summary statistics. Networks may deprioritize or block ICMP even when an application remains reachable.
A web application cannot assume that a visitor's browser can send arbitrary ICMP packets. This site's tool instead asks the application server to open a TCP connection to the target's HTTPS port, 443. A successful TCP handshake demonstrates reachability for that port from the service environment, not ICMP availability from your device.
The resulting status—reachable, connection refused, or timed out—has the same limited interpretation as other open, closed, and timed-out port results. It does not prove that the application behind the port is healthy or available from every network.
What the current tool actually measures
The tool accepts a public hostname or public IP. For a hostname, the server performs DNS resolution and applies public-address safety checks before selecting the displayed resolved address. DNS resolution occurs before the connection timer starts, so DNS lookup time is not included in the reported TCP latency.
Timing begins immediately before the server creates a TCP socket to the resolved address on port 443 and ends when the socket reports a successful connection. The number therefore mainly reflects that server-to-target network path and TCP handshake. It does not include a TLS handshake, an HTTP request, page download, application processing, rendering, or work performed by the target after accepting the connection.
Each run makes one connection attempt. A successful connection is shown as Reachable with one TCP latency value. A refused connection is shown as Connection refused, while a socket that does not complete within four seconds is shown as Timed out. Those unsuccessful states do not receive a latency value.
Interpreting milliseconds and changing results
A smaller number generally means this one TCP connection completed sooner than a larger number measured with the same method and endpoints. There is no universal threshold that makes every result good or bad. Interactive calls, games, page loads, APIs, and background transfers have different requirements, and one sample cannot show normal variation or sustained performance.
Physical distance, the route between networks, congestion, filtering, and conditions at the destination can change connection time. Server load can affect how promptly a connection is accepted, although this measurement stops before application processing. Results can also differ when DNS selects another public address or network routes change.
Wi-Fi quality and a user's VPN can affect latency measured from that user's own device, but they are not part of this tool's outbound server-to-target TCP path. The result describes the site's server environment. Compare it with measurements from the client network that actually matters before diagnosing a local wireless or VPN problem.
Timeouts, bandwidth, and measurement limits
A timeout means this single connection did not complete within the configured four-second limit. A firewall can silently filter the attempt, a route can fail, the selected address can be unreachable from the service environment, or a transient condition can intervene. It does not establish that the destination is permanently offline.
Latency and bandwidth are different. Latency describes delay, while download or upload speed describes how much data can be transferred over time. A connection can have low latency and limited throughput, or high latency and substantial throughput. This tool does not run a download, upload, or speed test.
The check does not report ICMP round trips, packet loss, jitter, averages, percentiles, route hops, TLS time, HTTP response time, or end-user Wi-Fi performance. It also tests only port 443 and one selected public address per run, so another port, address, region, or client can produce a different result.
Practical guidance
- Repeat the check at different times and treat each result as one server-to-target TCP observation rather than a universal rating.
- Compare the resolved address and port before comparing numbers from different tools or locations.
- Use a measurement from the affected client network when investigating local Wi-Fi, VPN, or device latency.
Common question
Why does this result differ from ping in my terminal? Your terminal may use repeated ICMP requests from your device, while this tool makes one TCP connection from the site's server to port 443. The origin, protocol, target address, and timing boundaries can all differ.