tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From André Warnier ...@ice-sa.com>
Subject Re: [OT] RE: Max Number of users
Date Mon, 09 Mar 2009 15:55:04 GMT
Ken Bowen wrote:
>>
> All,
> 
> Being bored and trying to avoid starting any real work on a slushy snowy 
> morning,  I organized the discussion for inclusion on Andre's list 
> (below). BTW, GIF/WIF:  
> http://en.wikipedia.org/wiki/Lies,_damned_lies,_and_statistics ::
> "Lies, damned lies, and statistics" is part of a phrase attributed to 
> Benjamin Disraeli and popularised in the United States by Mark Twain
> 
> Cheers...Ken
> 
>     From:     rbs100@gmx.com
>     Subject:     Max Number of users
>     Date:     March 9, 2009 2:53:26 AM EDT
>     To:     users@tomcat.apache.org
>     Reply-To:     users@tomcat.apache.org
> 
> Hi,
> Can any tell the maximum number of users can login into tomcat at the same
> time?
> Can I use tomcat to run in an online internet application used by n number
> of users and the application runs for 24*7?
> Is it reliable?
> Thanks
> RBS
> --------------
>     From:     rc46fi@googlemail.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 5:24:08 AM EDT
> 
> All depends on your configuration, directly porpotional to "n".
> 
> And no, Tomcat is not reliable at all, it's more kinda toy for bored
> developers such as me.
> 
> That's why we are using Tomcat in a 24/7-environment with a couple of
> 1000 users, never had any unscheduled dropouts.
> 
> Rgds
> 
> Gregor
> 
> PS.:http://wiki.apache.org/tomcat/FAQ/Performance_and_Monitoring
> --------------
>     From:     rainer.jung@kippdata.de
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 6:37:57 AM EDT
> 
> On 09.03.2009 10:24, Gregor Schneider wrote:
> And no, Tomcat is not reliable at all, it's more kinda toy for bored
> developers such as me.
> 
> :)
> --------------
>     From:     aw@ice-sa.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 8:40:23 AM EDT
> 
> Rainer Jung wrote:
> On 09.03.2009 10:24, Gregor Schneider wrote:
> And no, Tomcat is not reliable at all, it's more kinda toy for bored
> developers such as me.
> :)
> A statistical answer :
>  - searching Google for "tomcat problems" gives approximately 2,730,000 
> links.
>  - searching Google for "Tomcat works" gives 5,070,000 links.
>  - 2,730,000 + 5,070,000 = 7,800,000 links
>  - 5,070,000 / 7,800,000 = 0.65
> 65% reliability is not bad for a free product.
> --------------
>     From:     rosenberg.leon@googlemail.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 8:52:54 AM EDT
> 
> what a glorious example of self cozenage.
> Why do you assume that "tomcat works" is not included in the "tomcat
> problems" result?
> +tomcat +problems -works -> 1,830,000 results.
> +tomcat -problems +works -> 227,000
> 
> 227/(227+1830) -> 11% Reliability ... :-)
> 
> of course this statement is as un-serious as yours :-)
> --------------
>     From:     aw@ice-sa.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 9:20:39 AM EDT
> 
> I must admit that my first statistical analysis was flawed, and 
> furthermore lacked an element of comparison.
> So I refined it and created this table, applying the same objective 
> meta-analysis criteria to a range of products :
> 
> Product            problems   works     total      reliability %
> 
> Apache (httpd)     647,000    857,000   1,504,000    81
> IIS              6,530,000  5,900,000  12,430,000    47
> Tomcat           2,730,000  5,070,000   7,800,000    65
> Jboss              967,000  1,050,000   2,017,000    52
> Glassfish          125,000    155,000     280,000    55
> Windows         37,500,000 44,300,000  81,800,000    54
> 
>  From this, we can draw a number of conclusions :
> a) there are a lot more Windows users than anything else
> b) there are about 5 times more Tomcat users than Apache httpd users
> c) Apache httpd is at least 15% more reliable than any servlet server
> d) Tomcat is the most reliable servlet server
> e) Most IIS users have problems
> f) on a price/reliability ratio, Apache httpd and Tomcat win hands down
> --------------
>     From:     rc46fi@googlemail.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 9:36:00 AM EDT
> 
> Well, maybe this list helps to answer the question about Tomcat's 
> reliablity:
> http://wiki.apache.org/tomcat/PoweredBy
> 
> Besides, I guess a lot of companies (like us) don't advertise using
> Tomcat due to security-concerns.
> 
> I'm just wondering if the OP could digest any information from those
> answers now ;)
> 
> Rgds
> 
> Gregor
> --------------
>     From:     Chuck.Caldarale@unisys.com
>     Subject:     RE: Max Number of users
>     Date:     March 9, 2009 9:40:33 AM EDT
> 
> From: André Warnier [mailto:aw@ice-sa.com]
> Subject: Re: Max Number of users
> 
>  From this, we can draw a number of conclusions :
> 
> "There are three kinds of lies: lies, damned lies, and statistics." - 
> Benjamin Disraeli (maybe)
> 
> - Chuck
> --------------
>     From:     aw@ice-sa.com
>     Subject:     Re: Max Number of users
>     Date:     March 9, 2009 9:50:14 AM EDT
> 
> Caldarale, Charles R wrote:
> From: André Warnier [mailto:aw@ice-sa.com] Subject: Re: Max Number of 
> users
> 
>  From this, we can draw a number of conclusions :
> "There are three kinds of lies: lies, damned lies, and statistics." - 
> Benjamin Disraeli (maybe)
> A statistician is a person who can affirm that another person having one 
> foot caught in ice and the other foot in a bucket of boiling water, is 
> on average in a comfortable position.
> --------------
>     From:     Peter.Crowther@melandra.com
>     Subject:     [OT] RE: Max Number of users
>     Date:     March 9, 2009 10:02:56 AM EDT
> From: André Warnier [mailto:aw@ice-sa.com]
> A statistician is a person who can affirm that another person
> having one
> foot caught in ice and the other foot in a bucket of boiling water, is
> on average in a comfortable position.
> 
> ... but only if they're feeling mean.
> 
>                - Peter
> --------------
> 
> http://en.wikipedia.org/wiki/Lies,_damned_lies,_and_statistics ::
> "Lies, damned lies, and statistics" is part of a phrase attributed to 
> Benjamin Disraeli and popularised in the United States by Mark Twain
> 
Thanks to Ken for bringing a bit of order into this statistical chaos.

With the idea of refining the very rough analysis and table above, I 
tried to check if there would be any cultural bias in the results as 
provided by www.google.com, by crosschecking them by language.
I thus called up www.google.es (spanish), and searched for respectively 
"Tomcat problemas" (2,850,000 hits) and "Tomcat funciona" (only 88,000 
hits), for a total thus of 2,938,000 hits, which by the same calculation 
would give us a Spanish reliability ratio for Tomcat of only a dismal 
2.9 %. The reality is in fact even worse, because the search on "Tomcat 
funciona" generates a substantial number of hits on "Tomcat no 
funciona", so the 88,000 hits do not even represent the real total of 
satisfied users.
We would thus have to assume either that Spanish-speaking users, 
culturally, are less grateful and much more likely to post messages when 
things do not work than when they do work, or else have to admit that 
Tomcat is somehow incompatible with the Spanish language.
Would any of the committers know if there is substance to either one of 
the previous suspicions, or dispose of a statistical technique allowing 
to overcome this possibly cultural bias ?
As a control, I performed the same analysis on IIS, with the following 
results :

IIS problemas : 6,730,000
IIS funciona : 165,000
total : 6,895,000
reliability ratio : 2.3 %

which would tend to indicate that the bias, if any, also applies to 
other software products, even non-free.
To remove any high-technology related effect, I also tried to apply the 
same analysis to themes unrelated to Tomcat, such as "coche" (car) 
(4,850,000/2,200,000), "mujer" (woman) (15,900,000/10,200,000), and 
"hombre" (man) (17,200,000/11,300,000) and so far it seems to lead to 
the conclusion that Spanish-speaking people, independently of sex, are 
rather grouchy, although less so on these other kinds of products than 
on web software.



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Mime
View raw message