Ok, I wrote almost everything in the title, but let me expand on that a bit :)
As you might guess, the “alien engine” is Fusion — I still struggle to find a perfect one-line description for it :)
Thanks to Dmitry Filippov (̶g̶i̶r̶l̶s̶, just look at this huge pull request!), Fusion got .NET Standard 2.0 / .NET Framework 4.7+ support. Crazy, but almost every test runs on .NET Framework as well now — even the ones that use Fusion’s distributed features! To make it happen, Dmitry added .NET Standard 2.0 support to every Fusion’s assembly except
Stl.Fusion.Server, which had to be re-implemented as
Stl.Fusion.Server.NetFx due to quite significant API difference between ASP.NET Core and regular ASP.NET.
An interesting observation I’ve made recently is that nearly all UI architectures are quite similar.
Let’s start from Flux:
A brief description of what happens here:
Now, let’s look at Redux:
If you believe Virtual DOM is a “lightweight copy” of a real DOM helping to speed up the updates by batching them together— continue reading. Albeit being correct, it’s a highly misleading answer. Surprisingly, it’s also the most popular one: googling for “react virtual DOM” leads you to very similar descriptions: 1, 2, and you get ~ the same results even if you stick to the top answers on StackOverflow: 1, 2, 3;
The content below equally applicable to React and Blazor.
#1. It doesn’t highlight the key problem Virtual DOM solves — namely, the elimination of unnecessary re-renders. And…
The same day .NET 5 was released I shared a single screenshot showing how much faster .NET 5 is relatively to .NET Core 3.1. I promised to share more data later — and here it is.
Code: https://github.com/servicetitan/Stl.Fusion.Samples , see “Caching Sample” section there.
Overall, these tests stress an API endpoint fetching a small object from SQL database, the only difference is how it is exposed (locally or via HTTP) and if Fusion is used there to provide transparent caching.
My initial test was performed on Window & AMD CPU, but almost every production service runs on Linux nowadays, so…
I migrated Fusion to .NET 5 today — and honestly, I was absolutely astonished by the performance boost it brings:
The output is produced by Fusion’s “Caching” sample, which uses EF Core 5 and ASP.NET Core. The speed on tests producing 20M+ operations/s (#1, #3) is mainly constrained by Fusion’s logic and Castle.DynamicProxy. And tests producing around 100K operations/s are constrained by either EF & SQL Server (#2, #5) or ASP.NET Core (#4 — it’s ~ the same as #5, but relying on Fusion’s caching features). All of this means that:
Fusion — a small library created as an attempt to challenge the existing architecture of high-load and real-time apps — is going to be 4 months old soon. Here is what happened over these months:
Even though Amazon’s example isn’t unique in the context of this post, I’ll use it as a first example of efficiency of a modern web service:
Disclaimer: I am the author of Stl.Fusion — an open-source library helping to address the long-standing problem described in this post. But the problem is real, and the post is relevant even if this library won’t exist.
To answer this question, let’s ask the opposite one first: what’s the most rudimentary UX feature used by almost any web app today?
I nominate this button: