• You are not logged in. | Login

Post a reply

December 26, 2006 9:23 am

admin
Administrator
Administrator

Discuss article: "Password for a page. Part I."

You can discuss that article here

Discussed article: "Password for a page. Part I."


 

 

April 3, 2007 2:20 pm

senjor_itc
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

Hi! Sure, it is interesting but there is nothing new  sad  I’d like to ask maybe you know if it is possible to authorize a user through your form if the directory is closed with .htaccess i.e. to place data into the variables $PHP_AUTH_... within script. I have seen similar questions on many forums but I haven’t found answers to them.

In most articles devoted to protection is written that if there is a possibility for authorization by means of server, you shouldn’t invent anything new. As to the “Ability to view contents of files .htaccess and .htpasswd”, directive forbidding giving away files beginning with .ht

Last edited by senjor_itc (April 3, 2007 2:21 pm)


 

 

April 3, 2007 2:36 pm

Conker
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

I've heard the question if it's possible to authorize this way but I haven't heard any answer to it. And I know about ".ht"


 

 

April 3, 2007 2:46 pm

reetesh
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

I’m impatiently waiting for it to be continued!... And in what way letter ‘Z’ makes hash selection more difficult?


 

 

April 3, 2007 2:55 pm

Conker
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

It can be replaced every moment. And selector won’t be aware of what is there. By the way, instead of using one letter only you may simply write a non-sense combination of 10-20 letters/numbers. Such one is impossible to select. However it is not a universal protection.


 

 

April 3, 2007 3:08 pm

steven9x
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

The matter is that .htaccess files and form are incompatible due to two simple reasons

1. variables $PHP_AUTH_USER,$PHP_AUTH_PW cannot be changed from script.

2. to install them you are to send to user BasicRealm which is not a form, of course.

Two variants are possible…

Good luck.


 

 

April 3, 2007 3:34 pm

bandlist12
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

It’s better to use guest’s IP instead of letter ‘Z’ (as a variant forvarded_for if it is present). Selection becomes complicated and at the same time no one will succeed when entering from another address of someone else with this hash.
(IE has got such a bug – you may get cookie with known name even on (!!!) other server (which is different from the one that has given it.))


 

 

April 3, 2007 3:45 pm

bobbee
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

Gentlemen, how can I organize authorization if the server is not Apache but IIS.  Authorization request via header function doesn’t work and correspondingly $PHP_AUTH_USER and PHP_AUTH_PW don’t work either. cookies won’t help as far as they are not accepted by users.


 

 

April 4, 2007 6:58 am

Keeper
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

Of course, this is a method but what can we do with corporate users?
They all have the same IP even if they are 100 and if there is a hacker among them who is willing to select a password on your server, after some efforts he is simply going to disconnect the entire his net from access to your server…
I think there is no universal solution here.
It’s nearly impossible to stop the person who is going to break down your service.
There are definite control methods for the non-web interfaces’ users which are quite different from the ones used for users of web-interfaces. I think it would be more rational to control a web user by means of sessions. In such a case we deal with session ID which is really unique. Limiting of system access efforts number within one session is a rather simple task but it’s funny to rely upon the probability that a hacker will be sitting at the browser window and post forms with password selection to you is really funny. It is more likely to be a special soft which starts a new session when you have closed the previous one… And you won’t be able to do anything in this case. There is one way out – traffic analysis. What I mean is following: it’s necessary to control and analyze the number of sessions opened within system from definite IP. At this you are not to control all of them but only those that have been closed due to the password errors. You are either to guess or find out by an experiment critical number for this case and close access to the system for this IP when this number is achieved. Analyzing traffic you are not to forget about distribution of access rejection in some period of time in order not to mix usual users’ activity splash with hacker’s attack…


 

 

April 4, 2007 7:30 am

Conker
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

No one prevents user (selecting password) from choosing optional proxy each time which means that IPs may be different all the time.
My opinion is following: if someone is eager to select a password, you cannot do anything with it (neither with IP verification, neither with session nor with cookies).
Correct me if I’m mistaken


 

 

April 4, 2007 7:43 am

mambojim
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

Tell me, please, what is HTTP_X_FORWARDED_FOR? I have nothing like this in my manual.


 

 

April 4, 2007 7:50 am

Conker
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

$HTTP_X_FORWARDED_FOR. If proxy-server or firewall aren’t anonymous, it informs to which IP it is going to be connected to (usually there are some IPs separated with comma).
Eveyone has forgotten about flock() sad

Last edited by Conker (April 4, 2007 7:52 am)


 

 

April 4, 2007 8:00 am

IBdaMac
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

Sure, flock should be mentioned here as an alternative method...
But the system is rather complicated.


 

 

April 4, 2007 8:27 am

monkeydude
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

I have following remark concerning password selection.
The author gives such example: someone attaches to banner rollers applying to our site involving values substitution. It looks like a ‘selector’ is rejected by the system after 5 attempts. But banner will keep trying to get to the site. In an hour (it doesn’t matter for what period of time user is blocked by us) banners will try again. It doesn’t matter to them because it is a program. If it was a human, he would give up this in a couple of hours trying. So the main idea is fighting against the program for password selection.
But the standard solution comes at once: how to distinguish a user from program? We are to place a picture containing symbols’ set created randomly. If login and password are entered along with this code, we can be sure that we deal with a human and act as usually; if the code isn’t entered – we reject it and even don’t try to connect with the base or do anything else.
Why is this technology used by the registration of a new user only? Why isn’t it used for the authentication process?


 

 

April 4, 2007 8:42 am

jjjlc1983
Member
Ranks

Re: Discuss article: "Password for a page. Part I."

And what if we block name in the base? It’s possible to select passwords to names. It doesn’t take an hour, just a minute. You are to add ‘frequently forgotten logins’ into variables and not to involve the base…


 

 
  • Actions
  • Top
ITCrimea. Ukraine Web Development Company. Professional Developers and Web Designers Team
Custom Web Designs, Internet Applications, E-Commerce Websites, Interactive Sites, Database-Driven Sites and Services