This really awesome set of real life examples. I developed multiple of highload apps using Vert.x and results pretty good. Perfomance of apps are just impossible high. Thanks for shadding light on vert.x
While I understand the model, when it came to the Port authority with say 200 sensors, the data will not be coming in that quickly. This is human level effort and you can imagine about 100 crates moving around per hour. Each sensor point will be hit by a crate probably once every 15 seconds. A single simple server could easily manage all incoming IoT sensor data without blinking an eye. Lambda in AWS easily could handle this. So you are looking at about 5000-8000 sensor hits per hour-that is nothing in terms of computing power. A raspberry pi could handle that load, much less a real server. He mentions 7 million transactions, again, that's not a lot of data. If anyone has ever worked in Telcom you get 100-300 million per hour with data. This is managed by on-premise servers using traditional servers with HTTP or sockets. No new frameworks, just good solid code.
This guy simply deserves more time to be on stage. Amazing talk.
This really awesome set of real life examples.
I developed multiple of highload apps using Vert.x and results pretty good.
Perfomance of apps are just impossible high.
Thanks for shadding light on vert.x
this is the best presentation ever! 🎯
Love the fun way of delivery. Amazing real life examples and full of new tech stuff to learn. Thank you for uploading :)
Hardly any vertx... until 39:30.
what hardware you use?
I have learned a lot.
Great talk! Hmm, I'm wondering how good is the Java support on that VSC.
are the slides online somewhere?
Wow!
Very interesting.
Where is the code btw?
awesome
github repo??
check out link to the presentation, it has all the github links inside
While I understand the model, when it came to the Port authority with say 200 sensors, the data will not be coming in that quickly. This is human level effort and you can imagine about 100 crates moving around per hour. Each sensor point will be hit by a crate probably once every 15 seconds. A single simple server could easily manage all incoming IoT sensor data without blinking an eye. Lambda in AWS easily could handle this. So you are looking at about 5000-8000 sensor hits per hour-that is nothing in terms of computing power. A raspberry pi could handle that load, much less a real server. He mentions 7 million transactions, again, that's not a lot of data. If anyone has ever worked in Telcom you get 100-300 million per hour with data. This is managed by on-premise servers using traditional servers with HTTP or sockets. No new frameworks, just good solid code.
54:50 :p :p
Lol, could write C code in 8K of code to do the same thing, and it would outperform what you have showing.
Take a look at techempower's benchmarks. JVM langs are on part of the C, C++ & Rust web framework performance. It's scary fast!