Yikes, that's a long one - not really what I was after but interesting
anyway.

On another tiny side-note:
-----------------
> Small side note - I've heard people say that CFID/CFTOKEN is insecure
because people sitting in the middle 
> (like at an ISP) can sniff the URL & hijack the session. Well cookies are
no different, 
> the information is still going across the wire in plain text & anyone who
can look at the 
> HTTP headers can hijack the session. The only way to secure sessions is
the use HTTPS.
-----------------

I reckon it has nothing to do with CFID/CFTOKEN any session identifier from
any language is interceptable to this kind of hijacking. I reckon you can
make it a bit more difficult by looking at the remote IP or
xhttp_forwarded_for (I believe that's what it is) if behind a proxy.
I know, still not 100% secure, but securer...

-----------------
> The only way I can see this happening is that you might have been passing
CFID/CFTOKEN values 
> through the URL. If I have a URL that looks like
http://www.domain.com/index.cfm?CFID=XXXX&CFTOKEN=XXXXXXXX 
> and I posted it to this mailing list - everyone who clicked the link would
be the same client as 
> far as CF server was concerned & pick up the same session variables
because they are using the same PIN number.
>
> So if you are testing in two browsers in this way you are going to get the
same problem. However if you 
> clear the CFID/CFTOKEN off the link there is no way  that your session in
Netscape is going to be shared with your session in IE.
-----------------

I should have posted the whole thread I was talking about, someone said that
if you open a browser and start a session, then open a new browser, not by
clicking file -> new, but by actually clicking the browser icon again and
opening a new browser and start another session - you can maintain two
sessions, I thought it was a bit weird myself, but other people confirmed if
I am not mistaken..

Taco Fleur
Blog http://www.tacofleur.com/index/blog/
Methodology http://www.tacofleur.com/index/methodology/
0421 851 786
Tell me and I will forget
Show me and I will remember
Teach me and I will learn 


> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf 
> Of Mark Stanton
> Sent: Tuesday, 20 January 2004 6:33 PM
> To: CFAussie Mailing List
> Subject: [cfaussie] RE: Old Post RE: Browser Sessions - elaborate
> 
> 
> This is (not so) briefly how I understand the session & 
> client scopes in CF:
> 
> Session variables:
> Session variables are stored in memory on the SERVER. Each 
> session is referenced by a unique session identifier 
> (JSESSIONID or CFID/CFTOKEN pair).
> 
> HTTP is stateless, there is no inherent way to say "this 
> request is part of that session". So in for the server to 
> work this out some information has to be passed back & forth 
> between the server & the client on each request/response. 
> This information is the session identifier (JSESSIONID or 
> CFID/CFTOKEN pair).
> 
> The first time a browser hits a CF page the CF server makes a 
> new session, creates the session identifier & gives this back 
> to the client, this is kind of like the session PIN number. 
> The client then uses from them on to access its session on the server.
> 
> On each subsequent request the client sends this session 
> identifier along with its request (POST or GET) and the 
> server can then match the request up to the session. There 
> are two ways that this session identifier can be sent back; 
> either in the URL or in the HTTP headers as cookie data. This 
> might look something like this when it goes over the wire:
> 
> Cookie: JSESSIONID=96301828141074585835953;
> CFTOKEN=9c76f3e1c540c048-324E99B1-E168-32E6-4F58135858045667; 
> CFID=211501
> 
> 
> Regardless of which method is used the server looks at the 
> PIN, works out if the session is valid and if it is it will 
> use the relevant session variables.
> 
> Ok so what is the difference between using cookies & or the 
> URL. Well some fools don't allow cookies in their browsers 
> and/or some developers can't work out how to support IE's 
> privacy policies correctly (its not exactly easy but you can 
> usually get around everything but the highest privacy setting 
> if you implement P3P - www.w3.org/P3P/). When cookies fail 
> you can revert to passing CFID/CFTOKEN through the URL.
> 
> At the end of the day there are really no major differences 
> in terms of security or anything else.
> 
> Small side note - I've heard people say that CFID/CFTOKEN is 
> insecure because people sitting in the middle (like at an 
> ISP) can sniff the URL & hijack the session. Well cookies are 
> no different, the information is still going across the wire 
> in plain text & anyone who can look at the HTTP headers can 
> hijack the session. The only way to secure sessions is the use HTTPS.
> 
> Client Variables:
> Client variables pretty similar to session variables in most 
> ways. The main difference is that they persist (where as 
> session variables exist in memory & expire after a certain 
> period). Storing client variables in the registry or a 
> database is pretty much the same thing in terms of what we 
> are discussing now - the key point is the values get stored 
> on the server. Storing client variables in a cookie means 
> that you actually send them all back & forth between the 
> browser & server on each request.
> 
> I think that is pretty much session & client scopes in a nutshell.
> 
> 
> > <quote>Because the session is being tied to the "Client" 
> itself - i.e. 
> > the actual browser. If you try the same thing but with a different 
> > browser on the same machine it should work correctly.</quote>
> >
> > I tried it myself and did not get this to work, did anybody else?
> 
> The only way I can see this happening is that you might have 
> been passing CFID/CFTOKEN values through the URL. If I have a 
> URL that looks like 
> http://www.domain.com/index.cfm?CFID=XXXX&CFTOKEN=XXXXXXXX 
> and I posted it to this mailing list - everyone who clicked 
> the link would be the same client as far as CF server was 
> concerned & pick up the same session variables because they 
> are using the same PIN number.
> 
> So if you are testing in two browsers in this way you are 
> going to get the same problem. However if you clear the 
> CFID/CFTOKEN off the link there is no way  that your session 
> in Netscape is going to be shared with your session in IE.
> 
> 
> > Can anyone elaborate on the following "particular client and 
> > application and that must be saved between user sessions"
> 
> Client variables persist (as I mentioned before) so if I log 
> into a site tha t supports them, then go away & come back a 
> week later my session will be long gone, but my client 
> variables should still exist. This makes client variables 
> good for things like "Remember me" functionality or tracking 
> when I last visited the site & how many times I've been to 
> the site. In fact coldfusion even does some of the work for 
> you by automatically providing client.hitCount, 
> client.lastVisit & client.timeCreated.
> 
> Client variables are application specific, so if I have one 
> CF server with 3 apps on it I'll get unique values in my 
> client scope for each application.
> 
> hth
> 
> 
> Cheers
> 
> Mark
> 
> 
> ------------------
> Mark Stanton
> Technical Director
> Gruden Pty Ltd
> Tel: 9956 6388
> Mob: 0410 458 201
> Fax: 9956 8433
> http://www.gruden.com
> 
> 
> ---
> You are currently subscribed to cfaussie as: 
> [EMAIL PROTECTED] To unsubscribe send a blank email to 
> [EMAIL PROTECTED]
> 
> MXDU2004 + Macromedia DevCon AsiaPac + Sydney, Australia 
http://www.mxdu.com/ + 24-25 February, 2004


---
You are currently subscribed to cfaussie as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]

MXDU2004 + Macromedia DevCon AsiaPac + Sydney, Australia
http://www.mxdu.com/ + 24-25 February, 2004

Reply via email to