Re: internet growth article

This is absolutely on-topic, and you have a good point.

My belief is that the exponential growth of the Internet (in terms of
number of users) has slowed considerably, given that in most developed
nations access is nearly ubiquitous to the home and office. Still, the
sheer number of users is enormous -- well into the hundreds of millions
-- which is more than enough to overwhelm any given website.

My own web page has a pitiful number of hits per day, but my concern is
with "mega sites" like Yahoo, Google, and CNN, which see traffic in the
millions of hits per day range. If you read Slashdot, you are already
familiar with the Slashdot effect, where links to under-provisioned
sites (i.e., a single machine running Apache) are often dead for the
first few hours after being posted. So the load spike issue effects both
large sites and small ones (in special circumstances).

Matt

On Sat, 2003-09-06 at 02:15, Leigh Wanstead wrote:
> Hello everyone,
> 
> Execuse me if I asked this sounds simple qustion, I really want to know
> the answer and I am very confused about that.
> 
> I read following article from internet.
> 
> http://www.dtc.umn.edu/~odlyzko/doc/itcom.internet.growth.pdf
> 
> I raise this question is how this effect us to try to offer quality web
> service to handle massive web request demand. As stated in the article,
> "every 100 days internet volume double" is a myth. And whole dotcom
> collapse is because over investment on the internet volume requirement
> never exist. And at that time, majority people ignore the simle
> calculation. It seems it is true every year internet volume double from
> what I read. 
> 
> Does that mean slashdot effect might not be so serious?
> 
> Maybe I am wrong to try to compare internet volume to concurrent
> connection, not apple to apple compare.
> 
> I have followed the link on Dr. Matt website.
> 
> http://www.nedstatbasic.net/s?tab=1&link=1&id=1196094
> http://www.nedstatbasic.net/s?id=710309
> 
> The point is to show these public available statics do not show big demand
> for web request. As Pageviews 2712445 on most busiest day only translate
> to 31 pageview/second. Apache can handle around one thousand requests on a
> few connections in controled lab, note not thousands concurrenct
> connection make single web request. It is not that much. Of course, it did
> not tell what is the number of concurrent connections. Where can we found
> statics website to show concurrent connections number available to public
> for free?
> 
> Does web server/web application server consider to be internet back bone.
> I guess not. But I am not exactly sure about this.
> 
> I am sorry if you consider that to be offtopic.
> 
> TIA
> 
> Regards
> Leigh
> 
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> SEDA-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/seda-users



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf