Timeline for Re-architecting CPU intensive Node application to handle multiple users
Current License: CC BY-SA 4.0
27 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 11, 2021 at 21:08 | history | protected | gnat | ||
| Jan 10, 2021 at 19:14 | vote | accept | Mark | ||
| S Jan 10, 2021 at 19:14 | history | bounty ended | Mark | ||
| S Jan 10, 2021 at 19:14 | history | notice removed | Mark | ||
| Jan 10, 2021 at 15:05 | answer | added | lennon310 | timeline score: 4 | |
| Jan 9, 2021 at 20:20 | comment | added | Mark | @oshaiken Why do you say that? | |
| Jan 8, 2021 at 23:31 | comment | added | oshaiken | In this case you can only scale horizontally, because adding more CPU will not yeld expected results. | |
| Jan 8, 2021 at 22:50 | answer | added | Stack Exchange Broke The Law | timeline score: -1 | |
| Jan 8, 2021 at 21:22 | comment | added | RaspberryK | You should be able to handle this with a worker thread. Using Async Local Storage might provide some benefit as well (depending on how things are setup). Alternatively, consider the possibility of doing it all client side in a web worker. | |
| Jan 8, 2021 at 19:43 | comment | added | Mark | @oshaiken Too late for that! | |
| Jan 8, 2021 at 16:12 | comment | added | oshaiken | I think the issue is that you do not use NodeJS for CPU intensive processes. It is single threaded! | |
| Jan 4, 2021 at 21:26 | answer | added | imel96 | timeline score: 2 | |
| S Jan 3, 2021 at 21:39 | history | bounty started | Mark | ||
| S Jan 3, 2021 at 21:39 | history | notice added | Mark | Canonical answer required | |
| Jan 1, 2021 at 22:37 | audit | Close votes | |||
| Jan 1, 2021 at 22:38 | |||||
| S Dec 29, 2020 at 13:48 | history | suggested | cngzz1 | CC BY-SA 4.0 |
fixed grammar.
|
| Dec 29, 2020 at 9:36 | review | Suggested edits | |||
| S Dec 29, 2020 at 13:48 | |||||
| Dec 27, 2020 at 9:51 | comment | added | Doc Brown | @Mark: if you need to have this work with very different clients, including "small" devices like smart phones or thin clients, using a server-based approach is probably the only option which avoids the necessity of creating individual applications for each device. But then your users will have to live with some latency, that's unavoidable. | |
| Dec 27, 2020 at 9:43 | comment | added | Mark | @DocBrown The are medical researchers and doctors. Problem is that these files are growing bigger quickly as the technology improves - quicker than the improvement in browsers. Also the software works on every device at the moment - something I couldnt guarantee with a frontend approach. | |
| Dec 27, 2020 at 8:48 | comment | added | Doc Brown | @Mark: who are the users of this system / what kind of assumptions can you make for their clients (hardware & browser)? Maybe you simply have to optimize that approach? | |
| Dec 27, 2020 at 7:55 | comment | added | Mark | @DocBrown That was actually how I initially did it and it worked for small files but many users have much larger ones, in those cases, there is simply too much data for the browser to handle, and it tended to crash | |
| Dec 26, 2020 at 23:26 | comment | added | Doc Brown | What about doing the rendering fully locally at the client side? With HTML5 and Javascript how it looks today in all relevant browsers, wouldn't it be possible to produce the images without any upload and processing on the server? | |
| Dec 26, 2020 at 22:26 | answer | added | B. Ithica | timeline score: 2 | |
| Dec 26, 2020 at 21:00 | history | tweeted | twitter.com/StackSoftEng/status/1342938235140501504 | ||
| Dec 26, 2020 at 19:20 | answer | added | Greg Burghardt | timeline score: 2 | |
| Dec 26, 2020 at 17:09 | comment | added | W.K.S | Off the top of my head, I would look into using database streaming and sockets | |
| Dec 26, 2020 at 8:31 | history | asked | Mark | CC BY-SA 4.0 |