RE: Fuseguard processing time

2011-06-14 Thread Brook Davies

In production our average page times are between 200-300 ms during peak
load. This for all the pages that are DB intensive ect. The 20ms page I
tested on had no DB queries or much of any code.  I guess you could be right
that the hot spot compiler might kick in and optimize those classes/code.  I
might have to just run it in production and see how it fares..

Brook

-Original Message-
From: Sean Corfield [mailto:seancorfi...@gmail.com] 
Sent: June-13-11 4:06 PM
To: cf-talk
Subject: Re: Fuseguard processing time


On Mon, Jun 13, 2011 at 11:09 AM, Brook Davies cft...@logiforms.com wrote:
 And my test was done under zero load, so under peak load this number 
 could go up.

It could also go down. Until you test, you won't know.

Under heavy load, the HotSpot compiler in the JVM may work to your benefit
and speed things up. You may also benefit from the code being shared across
many concurrent requests. Do you know what your typical response times are
for your pages today in production under peak load?
--
Sean A Corfield -- (904) 302-SEAN
An Architect's View -- http://corfield.org/ World Singles, LLC. --
http://worldsingles.com/ Railo Technologies, Inc. --
http://www.getrailo.com/

Perfection is the enemy of the good.
-- Gustave Flaubert, French realist novelist (1821-1880)



~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345274
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-13 Thread Peter Boughton

 200ms is still a good page load time.
Not when the original was 20ms!

A page that takes 0.2s to load is no longer instant, there's a detectable 
delay, which isn't good.



Does it really take 145ms to check for SQL Injection? :/

What's it doing that takes that long!? 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345228
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-13 Thread Russ Michaels

LOL, ok well perhaps I am just used to see much worse loading times on most
peoples sites.

On Mon, Jun 13, 2011 at 3:40 PM, Peter Boughton bought...@gmail.com wrote:


  200ms is still a good page load time.
 Not when the original was 20ms!

 A page that takes 0.2s to load is no longer instant, there's a detectable
 delay, which isn't good.



 Does it really take 145ms to check for SQL Injection? :/

 What's it doing that takes that long!?

 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345233
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-13 Thread Dominic Watson

Of course, a measure on single hits is not a good measure of the
performance impact. You might be happyish adding a 200ms average at
peak load to a high traffic site - but who knows what the real impact
is without a proper performance test / pushing it live and hoping...

Dominic

On 13 June 2011 16:57, Russ Michaels r...@michaels.me.uk wrote:

 LOL, ok well perhaps I am just used to see much worse loading times on most
 peoples sites.

 On Mon, Jun 13, 2011 at 3:40 PM, Peter Boughton bought...@gmail.com wrote:


  200ms is still a good page load time.
 Not when the original was 20ms!

 A page that takes 0.2s to load is no longer instant, there's a detectable
 delay, which isn't good.



 Does it really take 145ms to check for SQL Injection? :/

 What's it doing that takes that long!?



 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345240
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


RE: Fuseguard processing time

2011-06-13 Thread Brook Davies

And my test was done under zero load, so under peak load this number could
go up.  Hmmm... I don' t know if I should install fuseguard or not in
production.  Now I am undecided...

Brook

-Original Message-
From: Dominic Watson [mailto:watson.domi...@googlemail.com] 
Sent: June-13-11 10:02 AM
To: cf-talk
Subject: Re: Fuseguard processing time


Of course, a measure on single hits is not a good measure of the performance
impact. You might be happyish adding a 200ms average at peak load to a high
traffic site - but who knows what the real impact is without a proper
performance test / pushing it live and hoping...

Dominic

On 13 June 2011 16:57, Russ Michaels r...@michaels.me.uk wrote:

 LOL, ok well perhaps I am just used to see much worse loading times on 
 most peoples sites.

 On Mon, Jun 13, 2011 at 3:40 PM, Peter Boughton bought...@gmail.com
wrote:


  200ms is still a good page load time.
 Not when the original was 20ms!

 A page that takes 0.2s to load is no longer instant, there's a 
 detectable delay, which isn't good.



 Does it really take 145ms to check for SQL Injection? :/

 What's it doing that takes that long!?



 



~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345242
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-13 Thread Peter Boughton

Well ideally you have a non-development staging server, which closely mimics 
your live production server, against which you can run load testing to help 
determine this.

The other question is, how secure is your code? If it's riddled with 
vulnerabilities then it might be safer to take this hit, at least until you can 
solve the various issues, rather than take the risk of someone exploiting your 
site. 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345243
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-13 Thread Sean Corfield

On Mon, Jun 13, 2011 at 11:09 AM, Brook Davies cft...@logiforms.com wrote:
 And my test was done under zero load, so under peak load this number could
 go up.

It could also go down. Until you test, you won't know.

Under heavy load, the HotSpot compiler in the JVM may work to your
benefit and speed things up. You may also benefit from the code being
shared across many concurrent requests. Do you know what your typical
response times are for your pages today in production under peak load?
-- 
Sean A Corfield -- (904) 302-SEAN
An Architect's View -- http://corfield.org/
World Singles, LLC. -- http://worldsingles.com/
Railo Technologies, Inc. -- http://www.getrailo.com/

Perfection is the enemy of the good.
-- Gustave Flaubert, French realist novelist (1821-1880)

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345246
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm


Re: Fuseguard processing time

2011-06-10 Thread Russ Michaels

200ms is still a good page load time. I wish most of our customers had pages
that load that quick :-)
consider what it is doing, checking all scopes for injection, which is all
string comparison (regex I presume), so that is heavy work.
Pete did say it would add about 200ms, so I guess he was right.


Russ

On Fri, Jun 10, 2011 at 4:15 PM, Brook Davies br...@logiforms.com wrote:


 I finally got around to getting an evaluation version of fuseGuard up and
 running on my dev server. Install was easy, so that was good. On my dev
 server, fuseguard is adding approx. 200 ms to each page load. Super simple
 pages with no form variable input went from 18-20ms to 190-200ms..



 The values are approximate, but it looks like the SqlInjectionFIlter.cfc is
 adding the bulk of the processing time at about 145ms per request. Is this
 normal? Seems like a fair amount of overhead.  Pete? Anyone?



 Brook D.






 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:345148
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm