Add hapi (Node)

[waghanza]

https://github.com/hapijs/hapi

[armand1m]

@waghanza great work mate! I see the PR is merged but no docs were added, nor a place in the benchmarks. These would be extremely helpful as well =)

[aichholzer]

Isn't Hapi like the slowest framework around?

[armand1m]

@aichholzer that is the point =)

[waghanza]

@armand1m Thanks. Good catch ;-)

due to intense work 😛 hapi has been dropped (unintentionally)

hapi will be added to the next release

@aichholzer We will see ... the next release will no be production-ready, the next after will (but I do not know when I will have the time to do so)

PS : Some previous results

Language (Runtime)	Framework (Middleware)	Requests / s	Throughput
node	fastify	61358.00	91.68 MB
node	polka	52718.00	78.61 MB
node	rayo	47574.00	71.10 MB
node	koa	36394.00	76.90 MB
node	express	33598.00	82.17 MB
node	restify	34149.00	59.80 MB
node	hapi	29352.00	54.91 MB

[aichholzer]

How did Fastify get almost 10K req/s ahead of the rest? - Does not seem to be realistic.

[aichholzer]

If you are running on a multi-core server, then you should enable @rayo/storm in the latest version of rayo.

It will increase it's performance (and throughput) by almost 200% percent, hence blowing the socks off the other contenders, although it will probably not be a fair competition... 🙈

[waghanza]

@aichholzer 😛 I'm sure you'll be impatient to compete ^^

Feel free to PR usage for @rayo/storm. I'll be glad to merge

PS : All benchmarks are using PM2, it could be changed ... I'm not so skilled in node

[aichholzer]

PM2 will run on multi-cores if enabled and if the server supports it. I will try to get a PR across. Still curious on Fastify's performance gains.

[waghanza]

@aichholzer sending those result was a (gentle) provocation 😛.

Above results are garbage / draft

[aichholzer]

🤣
Well done!