Looking For Direction
in [Java Programming]
|
From: Martin Gregorie on 12 Jun 2010 17:17 On Sat, 12 Jun 2010 14:32:34 -0400, Arne Vajhøj wrote: > On 12-06-2010 14:11, Martin Gregorie wrote: >> On Sat, 12 Jun 2010 12:07:40 -0400, Arne Vajhøj wrote: >>> But I don't understand why you think web server platform should relate >>> to either Java EE platform or overall server platform. >> >> LAMP >> >> The acronym wouldn't exist if it didn't describe a reasonably common >> software stack. Nothing to do with Java, of course, but this thread had >> diverged into the demographics of servers. > > I know. > > But we have: > - numbers for Java EE servers > - numbers for servers total > - numbers for web servers > > And I read your comments as if you think the last numbers indicate that > one of the first sets of number are not valid. > > Which I don't understand. > I don't think *any* of the numbers are valid, partly because none of the sources of these metrics seem to have published their collection methods and partly because many of the sources would seem to have agendas. I'd believe them if they came from the webcrawlers of large search engines, there was general agreement among these statistics and there was an agreed disambiguation method to avoid the bias due to large webhosting farms, i.e. a single machine supporting 'n' websites with a common software stack should not be scored as 'n' instances of each element of the software stack. -- martin@ | Martin Gregorie gregorie. | Essex, UK org |
From: Arne Vajhøj on 12 Jun 2010 19:03 On 12-06-2010 14:53, ilan wrote: > Arne Vajhøj<arne(a)vajhoej.dk> writes: > >> On 11-06-2010 06:57, ilan(a)whatever.com wrote: > >>> Here is something factual. >>> >>> http://en.wikipedia.org/wiki/Web_server >>> http://news.netcraft.com/archives/2010/01/ >>> >>> Vendor Product Web Sites Hosted (millions) Percent >>> Apache Apache 111 54% >>> Microsoft IIS 50 24% >>> Igor Sysoev nginx 16 8% >>> Google GWS 15 7% >>> lighttpd lighttpd 1 0.46% >> >> Well - 24% only runs on Windows and 62% can run on Windows, so >> you have proved that Windows is used on 24-86% of web servers. > > Hmm.. I think we can explore this a little bit. I was only looking at > web servers in this instance. I doubt for example the majority of Oracle > databases run on Windows.. even if Oracle _can_ run on Windows. > > I also doubt the majority of enterprises put their ERP systems onto > Windows boxes. > > But lets stay with web servers; for now. > > So is there any of determining of the 54% running Apache; how many > are using Windows? I would suspect that the vast majority are not - but > I don't like guesses. I would love to know the statistics on that. > > I have googled a little bit; but can not find anything. Probably got my > search criteria wrong. It is many years since NetCraft publicised OS info. I don't think it is available. Arne
From: Arne Vajhøj on 12 Jun 2010 19:05 On 12-06-2010 17:17, Martin Gregorie wrote: > On Sat, 12 Jun 2010 14:32:34 -0400, Arne Vajhøj wrote: > >> On 12-06-2010 14:11, Martin Gregorie wrote: >>> On Sat, 12 Jun 2010 12:07:40 -0400, Arne Vajhøj wrote: >>>> But I don't understand why you think web server platform should relate >>>> to either Java EE platform or overall server platform. >>> >>> LAMP >>> >>> The acronym wouldn't exist if it didn't describe a reasonably common >>> software stack. Nothing to do with Java, of course, but this thread had >>> diverged into the demographics of servers. >> >> I know. >> >> But we have: >> - numbers for Java EE servers >> - numbers for servers total >> - numbers for web servers >> >> And I read your comments as if you think the last numbers indicate that >> one of the first sets of number are not valid. >> >> Which I don't understand. >> > I don't think *any* of the numbers are valid, partly because none of the > sources of these metrics seem to have published their collection methods > and partly because many of the sources would seem to have agendas. The background for the Java survey was most certainly published. I don't think the analysis companies reveal too much about how they get the numbers. Neither respondents to the Java survey not the analysis companies seems to have good reason to bias their results. Arne
From: Arne Vajhøj on 12 Jun 2010 19:17 On 12-06-2010 14:53, ilan wrote: > Hmm.. I think we can explore this a little bit. I was only looking at > web servers in this instance. I doubt for example the majority of Oracle > databases run on Windows.. even if Oracle _can_ run on Windows. > > I also doubt the majority of enterprises put their ERP systems onto > Windows boxes. The OS's obviously have their niches. Linux is very strong in web servers (Apache, PHP etc.). The commercial Unixes are very strong in databases. Windows is very strong in file/print/mail servers. Java is probably one of the more diversified areas. And it is a myth that business critical stuff is not put on Windows. There are a lot of ERP systems running on Windows. Not so much Oracle's, but a lot of SAP systems are running on Windows. Arne
From: David Lamb on 13 Jun 2010 13:36
On 12/06/2010 5:17 PM, Martin Gregorie wrote: > i.e. a single machine supporting 'n' websites with a common > software stack should not be scored as 'n' instances of each element of > the software stack. I'm not sure a single machine running one stack should count as much as a farm of 1000, either. I suppose one hopes for a surveyor to provide the raw data so different people can push their own interpretations. |