Formula Used
Non persistent HTTP opens a separate connection for each requested item.
A common timing model uses two RTTs per item.
One RTT starts the connection.
Another RTT sends the request and receives the first response byte.
Base page time = DNS + connection RTTs + server time + HTML transfer time.
Embedded object time = request waves × connection RTTs + wave server delay + total object transfer time.
Total time = base page time + embedded object time.
Transfer time is calculated from kilobytes, overhead, and bandwidth.
Understanding Non Persistent HTTP Time
Non persistent HTTP is a simple request model. Each file uses a new connection. The browser gets the main HTML file first. It then discovers extra objects. These objects may be images, scripts, style sheets, fonts, or small media files. Every object adds delay because the connection does not stay open.
Why RTT Matters
RTT means round trip time. It is the time needed for a signal to travel to the server and return. In this model, every request needs connection setup and request response travel. That is why two RTTs are often used for each object. High RTT can make a small page feel slow, even when bandwidth is strong.
Bandwidth and Transfer Time
Bandwidth controls how fast data moves. Large files need more transfer time. Header data also adds small weight. Protocol overhead can increase the transferred amount. The calculator includes these parts. This helps compare a light page with a media heavy page. It also shows when file size matters more than latency.
Parallel Requests
Browsers may open several connections at once. This reduces waiting time for embedded files. The calculator groups embedded objects into waves. More parallel connections create fewer waves. Fewer waves reduce RTT overhead. The bandwidth is still shared by all transferred data. This keeps the estimate practical.
Planning Use
Use this tool for teaching, testing, page planning, and network review. Try different RTT values for local, regional, and distant servers. Raise object count to see the cost of many small files. Lower bandwidth to test slow links. Add extra RTTs when security handshakes or proxy delays apply. The result is an estimate, not a live speed test. Still, it gives a clear view of the main timing drivers.
FAQs
What is non persistent HTTP?
It is a request method where each object uses a separate connection. The connection closes after the response. This increases timing cost when a page has many objects.
Why does the calculator use two RTTs?
One RTT commonly represents connection setup. Another RTT represents the request and first response byte. Extra RTTs can be added for secure or delayed handshakes.
What is an embedded object?
An embedded object is any extra file loaded by the main page. Examples include images, scripts, style files, fonts, icons, and media assets.
Should DNS time be included?
Include DNS time when the lookup is not cached. Use zero when DNS is already resolved or when you only want request transfer timing.
What does parallel connections mean?
It means several objects are requested at the same time. The calculator groups objects into waves and reduces repeated RTT waiting.
Does bandwidth affect every request?
Yes. Bandwidth affects HTML, object bodies, and header transfer. Larger files or lower bandwidth increase the transfer portion of total time.
Why add protocol overhead?
Overhead accounts for extra bytes from transport, framing, and network layers. It gives a more realistic estimate than file size alone.
Is this a real page speed test?
No. It is a planning estimate. Real page speed also depends on caching, server behavior, compression, browser limits, and client processing.