Blazor Web vs WASM standalone startup time difference is huge.

[zubairkhakwani]

I have built a website using Blazor WEB. i have now converted the whole client side project in to separate Blazor standalone and the startup time difference is huge.

on mobile my blazor web app (InteractiveAuto) is taking 10 to 11 seconds to start and blazor standalone is taking 5 to 6 seconds.

on desktop my blazor web app (InteractiveAuto) is taking 3.45 seconds and blazor standalone is taking 2.5 sec to start.

and with InteractiveWebAssembly blazor web app is still taking around 7.40 to 8+ seconds.

the project and all dependencies are 100% same.

is this expected behaviour.

[oroztocil]

Hi @zubairkhakwani, thank you for contacting us.

Without more details about your project, hosting setup and measurements (e.g. first load vs repeated loads) we cannot say much. If you are using .NET 10, please make sure to upgrade to the latest patch (10.0.2) as we have fixed a Wasm asset caching bug present in the initial 10.0.0 release.

[zubairkhakwani]

Hi @zubairkhakwani, thank you for contacting us.

Without more details about your project, hosting setup and measurements (e.g. first load vs repeated loads) we cannot say much. If you are using .NET 10, please make sure to upgrade to the latest patch (10.0.2) as we have fixed a Wasm asset caching bug present in the initial 10.0.0 release.

Thanks for the response.

To clarify a few points:

• I am using self-contained deployment on a server.
• The runtime is .NET 10.0.2 (latest patch).
• Wasm asset caching is working correctly.
• The load times I mentioned are for repeated loads after everything is fully cached, not first load.
• Both the Blazor Web (InteractiveAuto / InteractiveWebAssembly) app and the standalone Blazor WebAssembly app are: Hosted on the same server.
• Tested on the same internet connection
• Built from the same project and dependencies

Despite this, there is still a consistent and noticeable startup time difference, where the standalone Blazor WebAssembly app loads faster than the Blazor Web app after caching.

You can also observe this behavior in a freshly created project:
If you create a new standalone Blazor WebAssembly project and a Blazor Web project (InteractiveAuto or InteractiveWebAssembly), publish and host both, the standalone version consistently starts faster after assets are cached.

If you need any additional information or measurements, please let me know and I’ll be happy to provide them.

[oroztocil]

@zubairkhakwani Thank you for the information. How are you hosting the static files for the standalone app? (nginx, IIS, etc)

In general, the Blazor Web pipeline is more complex (including antiforgery, auth, and other middleware) than the standalone hosting model. However, we have not seen reports of significant differences in load time (or time-to-interactivity) such as yours.

[javiercn]

@zubairkhakwani can you check for differences on TTFB?

At least the initial request will have to be processed on the server. Are you using prerendering?

[zubairkhakwani]

@zubairkhakwani Thank you for the information. How are you hosting the static files for the standalone app? (nginx, IIS, etc)

In general, the Blazor Web pipeline is more complex (including antiforgery, auth, and other middleware) than the standalone hosting model. However, we have not seen reports of significant differences in load time (or time-to-interactivity) such as yours.

I'm using plesk panel on windows for both projects.

I will record and attach a small video showing the difference tomorrow. I will try to keep it as simple as possible.

[zubairkhakwani]

@zubairkhakwani can you check for differences on TTFB?

At least the initial request will have to be processed on the server. Are you using prerendering?

I'm not using pre-rendering. And i will provide the request details tomorrow.