Server Capability Question

I am trying to chart out my learning path to learn my way
directly to
my goal....and would appreciate your insight. Right now, I
am learning
Python, (I have commercially programmed in BASIC, 20 years
ago...but
everything today, is a new experience to me as I have been
out of the
computer world), and am looking hard at using Turbo-Gears,
which uses
Cherrypy, to build my web business with....when I become
capable of
doing so. Of course, I am looking for the best and shortest
path to
create my business on the web. Hopefully, those two things
requirements
can be met in the Python world.

I noticed that Turbo-gears leaves room to use either Apache
or
Cherrypy's server. Does this mean there is a point at which
the
Cherrypy server becomes too loaded down to handle the volume
of traffic
hitting a website that is getting more and more popular?
Assume my
business, if sucessful, (one hopes!) would be similar in
nature and
volume to ebay or Amazon.com, and database driven. (I have
been
thinking Postgre...) Which of these two servers should I
plan on
learning and implementing? And why?

I really appreciate any insight you can give me on this
issue.
Have a great day!


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

On 3/8/06, wildcard <terry.kemmerergmail.com> wrote:
> I noticed that Turbo-gears leaves room to use either
Apache or
> Cherrypy's server. Does this mean there is a point at
which the
> Cherrypy server becomes too loaded down to handle the
volume of traffic
> hitting a website that is getting more and more
popular? Assume my
> business, if sucessful, (one hopes!) would be similar
in nature and
> volume to ebay or Amazon.com, and database driven. (I
have been
> thinking Postgre...) Which of these two servers should
I plan on
> learning and implementing? And why?

"similar in nature and volume to ebay or
Amazon.com". The #7 and #12
sites on the net. Nothing like thinking big 

Being able to scale like that is less a function of server
choice and
more a function of how you design your app. If you design
your app to
linearly scale upward by adding new boxes, you're in good
shape.

Kevin

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

Kevin. Would you mind explaining just a bit more about what
you meant
by "adding new boxes" so that when I run into
that area in my studies I
can pay particular attention as to how to set up my app to
linearly
scale upward? Thanks!

P.S. I always think big......it makes a much bigger splat on
my feet! 


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

--- wildcard <terry.kemmerergmail.com> wrote:

> 
> Kevin. Would you mind explaining just a bit more about
what you meant
> by "adding new boxes" so that when I run
into that area in my studies I
> can pay particular attention as to how to set up my app
to linearly
> scale upward? Thanks!

You mentioned Amazon/ebay in your previous post.  The
throughput experienced by
these websites cannot be handled by one computer.  Instead,
they will likely
have a number of computers performing specific tasks.  There
might be 10
database servers for example.  Each time you request a
resource from amazon,
all the servers will "team up" to handle your
reqeust.  How they team up
depends on how the system is designed.

Large sites like these have designed their systems so that
if they experience
poor performance in some part of their system, they simply
have to buy another
machine, plug it into the network, and hey presto,
performance is improved. 
For your studies, look out for terms like "Distributed
Systems", "Enterprise
Applications".  You would find more by googling these
terms.

Hope you remember your fellow pythonistas when your business
takes off 

Regards,
Andy


	
	
		
___________________________________________________________ 
Yahoo! Messenger - NEW crystal clear PC to PC calling
worldwide with voicemail http://uk.messenger.yah
oo.com

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

On 3/9/06, wildcard <terry.kemmerergmail.com> wrote:
>
> Kevin. Would you mind explaining just a bit more about
what you meant
> by "adding new boxes" so that when I run
into that area in my studies I
> can pay particular attention as to how to set up my app
to linearly
> scale upward? Thanks!
>
> P.S. I always think big......it makes a much bigger
splat on my feet! 

In addition to Andy's message, I always like to highlight
the "shared
nothing" (google it) approach to webapp development as
being key.
Anything you can do to minimize server-side state (by
stuffing more in
the browser, for example) makes scaling easier. Anything you
can do to
allow splitting your database up into chunks makes scaling
easier. If
you have something that requires a big calculation, being
able to
split that up across machines makes scaling easier
(http://l
abs.google.com/papers/mapreduce.html).

I'd also caution against going for an overly complex
architecture
before you need it. (You can think about how you would
partition your
database beforehand without actually doing it and cluttering
your code
before you need it.)

Kevin

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

On 3/9/06, wildcard <terry.kemmerergmail.com> wrote:
>
> Kevin. Would you mind explaining just a bit more about
what you meant
> by "adding new boxes" so that when I run
into that area in my studies I
> can pay particular attention as to how to set up my app
to linearly
> scale upward? Thanks!

To add to what has already been said, but on a different
level, take a
look at Pound:
http://www.apsis.ch/pound/

htt
p://www.debian-administration.org/articles/249
There are also commercial programs that load balance web
servers in
the same way.  All major web sites use something like this,
regardless
of what web server they use.
Of course, that is just the front end.  At the back end
there will be
multiple database servers with replication.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

> if sucessful would be similar in nature and
> volume to ebay or Amazon.com, and database driven.

At Google all searches originating from the United States
(excluding
Puerto Rico) are served by python CherryPy server configured
with 25
threads and a static filter enabled for the Google logo,


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---

Istvan Albert wrote:
>> if sucessful would be similar in nature and
>> volume to ebay or Amazon.com, and database driven.
> 
> At Google all searches originating from the United
States (excluding
> Puerto Rico) are served by python CherryPy server
configured with 25
> threads and a static filter enabled for the Google
logo,

And what server handles Puerto Rico?  Perhaps CGIHTTPServer?



--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at http://
groups.google.com/group/cherrypy-users
-~----------~----~----~----~------~----~------~--~---