tag:blogger.com,1999:blog-37798047.post6540543009655029495..comments2024-01-16T05:48:33.523-05:00Comments on Errata Security: Scalability: it's the question that drives usDavid Maynorhttp://www.blogger.com/profile/09921229607193067441noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-37798047.post-80032771302135892562020-11-27T05:50:53.372-05:002020-11-27T05:50:53.372-05:00This comment has been removed by a blog administrator.Hugohttps://www.blogger.com/profile/12896452500086339918noreply@blogger.comtag:blogger.com,1999:blog-37798047.post-88856934276122718302014-06-21T02:12:53.585-04:002014-06-21T02:12:53.585-04:00It would be interesting to see how the libdispatch...It would be interesting to see how the libdispatch version of Apache fares in similar tests<a href="http://goo.gl/mtxlE6" rel="nofollow">me</a> ...!!! jambuhttps://www.blogger.com/profile/04435394996693207796noreply@blogger.comtag:blogger.com,1999:blog-37798047.post-5598940680859199652013-11-01T13:30:22.475-04:002013-11-01T13:30:22.475-04:00The concept is correct, but in some places you sho...The concept is correct, but in some places you should probably use the word capacity. For example: "Even using a server 16 times as fast as the original, we still haven't even doubled scalability". It should end with "haven't even doubled capacity". This is a scalability problem. Scalability isn't a quantity you can double. It's a property (i.e., a quality). The definition I use is: The ability to add resources to a system in order to attain or sustain a desired performance level. I'm sure that requires modification in some other contexts. But your point is still correct.Tomhttps://www.blogger.com/profile/06093731178507441020noreply@blogger.comtag:blogger.com,1999:blog-37798047.post-73033719748005786622013-02-26T00:51:15.328-05:002013-02-26T00:51:15.328-05:00It would be interesting to see how the libdispatch...It would be interesting to see how the libdispatch version of Apache fares in similar tests.<br />It currently only runs on OS X or FreeBSD and scoreboard doesn't work but it might make a huge difference to the overhead.<br /><br />http://libdispatch.macosforge.org/trac/wiki/apachemikehttps://www.blogger.com/profile/16218389948653669064noreply@blogger.comtag:blogger.com,1999:blog-37798047.post-57697333695478022892013-02-19T17:55:41.329-05:002013-02-19T17:55:41.329-05:00"It'll happily run out of memory and then..."It'll happily run out of memory and then fall over."<br />Seriously? Then it does not scale.<br /><br />The technology behind NodeJS is not new, it exists for tons of other languages. Why not talk about them, instead?<br /><br />NodeJS is reinventing the wheel, except badly.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-37798047.post-14776748840140059632013-02-19T16:17:27.507-05:002013-02-19T16:17:27.507-05:00In response to the post above:
Have I tried Node...In response to the post above: <br /><br />Have I tried NodeJS? Yes. I've written an app on NodeJS and tested its scalability vs. Apache. It scales vastly better.<br /><br />The only people I've found that dislike NodeJS are those who don't have scalability problems.<br /><br />Of course, NodeJS has lots of other problems. It'll happily run out of memory and then fall over. I'm not sure that I would use it for most problems. I just use it as an example of something that, unlike Apache, does scale.<br />Robert Grahamhttps://www.blogger.com/profile/09879238874208877740noreply@blogger.comtag:blogger.com,1999:blog-37798047.post-73805913264900289362013-02-19T12:14:39.727-05:002013-02-19T12:14:39.727-05:00Did you actually try NodeJS in a real environment,...Did you actually try NodeJS in a real environment, or do you believe their marketing crap?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-37798047.post-31988278124235218332013-02-18T22:37:17.659-05:002013-02-18T22:37:17.659-05:00It seems wrong to be talking about scalability abs...It seems wrong to be talking about scalability absent from looking at the work actually being done for each request. Node can be scalable up the wazoo, but if there's 5 seconds of CPU time per request, the solution is not scalable. The only time a solution like that makes sense is if your clients are mostly idle.EJ campbellhttp://sports.yahoo.comnoreply@blogger.comtag:blogger.com,1999:blog-37798047.post-53332389412692311372013-02-18T15:03:25.699-05:002013-02-18T15:03:25.699-05:00I find it easier to read with requests/s on the x ...I find it easier to read with requests/s on the x axis and non-error responses/s on the y axis. 99th percentile latency vs. requests/s is another good metric. I don't care about concurrency per se, just supportable load.Wes Felterhttps://www.blogger.com/profile/01395217775195260835noreply@blogger.com