A simple question... about hyper-express HOT 3 CLOSED

I got it. Your explanation is deeper than this faq. Cache is always a smart solution.
tks again. 👍

from hyper-express.

There is no minimal infrastructure special to uWebsocket. You can use both Express/HyperExpress on any hardware and HyperExpress/uWebsocket will beat Express in every way when it comes to pure throughput performance.

What that statement is stating is this: let's say you had a route called /users/:id that made a SELECT query to your SQL database and returned results, you will actually notice that HyperExpress/Express wil have very similar performance. This is simply an illusion because since each request makes an SQL query to your database, you are essentially stressing your database not the webserver. When he says to cache requests as much as possible, he means to implement intermittent caches. So for example, let's say you had an endpoint called /currency/consume which made a third-party API request to retrieve fresh currency data. What you would do to maximize performance is use a package like https://github.com/kartikk221/cached-lookup and put the third-party API behind a 5 second cache. This way if your server got let's say 50k requests/second, your server server will continue to serve the latest cached data from memory while fetching fresh data every 5 seconds as needed. This would make full use of the available throughput without stressing the third-party API.

from hyper-express.

cached-lookup => Of course, I will include in my PoC.

from hyper-express.