magicianspell
asked on
Laptop Security over VPN
Hi..
Pleas tell me ways to implement "two factor security" for my remote mobile users who access my corporate network via Nortell contivity VPN over internet.
Please tell me what are the options available for security like ...
Digital certificates or
tokens
smart cards etc...
what is the best way of securing a laptop in case it gets stolen.... ?
can anybody tell me the price of having a "digital certificate" per user from verisign etc...
Pleas tell me ways to implement "two factor security" for my remote mobile users who access my corporate network via Nortell contivity VPN over internet.
Please tell me what are the options available for security like ...
Digital certificates or
tokens
smart cards etc...
what is the best way of securing a laptop in case it gets stolen.... ?
can anybody tell me the price of having a "digital certificate" per user from verisign etc...
If you're using a Nortel Contivity box, then I'd beg to differ with Chris. RSA SecurID tokens offer far better security than simply using a username and password.
Thinking what would need to be compromised if you used this style of authentication on the Nortel, your hacker would need to know the following:
A valid username
A valid groupname
The password associated with the groupname above
The PIN number associated with your username's token
The current number on the securID token associated with username above
The minimum version of software he could connect to your Nortel server with (it can be locked down).
Even if you knew the seeded pseudo number generation algorithm that is used on the token (the seed is individual to each token and the number displayed is nothing to do with the time of day, your token needs syncronising before first use) then you'd still have to get all the information above to get on. You have to admit it's a lot of trouble to go to, and potentially as a hacker you then have to authenticate yourself on the network to Windows Domains and so on.
Any vulnerabilities in the implementation of RSA SecurID are not going to be cracked by some kiddie with his scripting tools, that you can be certain.
Take a read of the following commentary if you're not convinced:
http://seclists.org/lists/bugtraq/1996/Sep/0014.html
And I'm sure you know where to go to find info on RSA SecuriD/ACE Servers :-)
Have fun,
R.
Thinking what would need to be compromised if you used this style of authentication on the Nortel, your hacker would need to know the following:
A valid username
A valid groupname
The password associated with the groupname above
The PIN number associated with your username's token
The current number on the securID token associated with username above
The minimum version of software he could connect to your Nortel server with (it can be locked down).
Even if you knew the seeded pseudo number generation algorithm that is used on the token (the seed is individual to each token and the number displayed is nothing to do with the time of day, your token needs syncronising before first use) then you'd still have to get all the information above to get on. You have to admit it's a lot of trouble to go to, and potentially as a hacker you then have to authenticate yourself on the network to Windows Domains and so on.
Any vulnerabilities in the implementation of RSA SecurID are not going to be cracked by some kiddie with his scripting tools, that you can be certain.
Take a read of the following commentary if you're not convinced:
http://seclists.org/lists/bugtraq/1996/Sep/0014.html
And I'm sure you know where to go to find info on RSA SecuriD/ACE Servers :-)
Have fun,
R.
You are wrong, o' obi-wan.
The SecurID algorithm has been cracked, cracked, cracked. See the @stake research report on the same, along with the dozens of script-kiddie scripts to exploit it. You merely need to observe/sniff a single instance of the number that's generated by the thing to be able to generate all future numbers.
This means it is not a "what you have" but a "what you know" authenticator, and therefore not two-factor.
Better than a regular password? Only if your system allows poorly chosen passwords that can be cracked by a dictionary attack.
Otherwise, about the same as a password with good password selection/update hygene.
The SecurID algorithm has been cracked, cracked, cracked. See the @stake research report on the same, along with the dozens of script-kiddie scripts to exploit it. You merely need to observe/sniff a single instance of the number that's generated by the thing to be able to generate all future numbers.
This means it is not a "what you have" but a "what you know" authenticator, and therefore not two-factor.
Better than a regular password? Only if your system allows poorly chosen passwords that can be cracked by a dictionary attack.
Otherwise, about the same as a password with good password selection/update hygene.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
WRT the @stake conclusions: Agreed. SecurID is not useless, and encrypting the communications can greatly reduce the risk. On the other hand, the same is true of passwords. IMO, a password-based system with strong password controls is on the same order of security as SecurID. If you want strong two-factor, I'd go for X.509-cert unlocked by PIN (good), SmartCard/dongle holding X.509-cert unlocked by PIN (better), or SmartCard/dongle holding X.509-cert unlocked by biometric (best).
As for the scripting exploits, you do no have to compromse the 64-bit seed. You merely need a copy of the token output plus the timestamp when it was generated. You can derive the 64-bit seed from that (that's why the algorithm is broken). The @stake paper talks about doing this, and there was source-doce posted to Bugtraq a couple of years ago.
As for the scripting exploits, you do no have to compromse the 64-bit seed. You merely need a copy of the token output plus the timestamp when it was generated. You can derive the 64-bit seed from that (that's why the algorithm is broken). The @stake paper talks about doing this, and there was source-doce posted to Bugtraq a couple of years ago.
Depends on what your VPN solution support. Most support certificates. Some support tokens or smart cards. Contact the vendor(s).
Note that not all supposed two-factor solutions really are two-factor. For example, RSA SecurID tokens have a reversable algorithm based on the time of day and the pin, so they're really no different than a password.
Securing a laptop in case its stolen:
Your best bet is filesystem encryption. The encrypting filesystem support in Windows XP (assuming Windows) is reasonably decent. There are also commercial solutions from places like PGP and Norton, as well as solutions for non-Windows systems like Mac and Linux.
Price of a verisign cert:
Personal cert is something like $25/year/person. Server cert is more like a few hundred $/year/server.