Re: Isolate systems

From: Steven L Umbach (
Date: 02/08/05

Date: Tue, 8 Feb 2005 11:20:11 -0600

Some ports obviously need to be available for legitimate users. Beyond using
some sort of port/protocol/Ip/mac"filtering" via switches, ipsec filtering,
firewall, routers, personal firewalls, you will need to rely on enforcing
strong passwords, hardening, and patching to ward off the evil. I assume
those ports are being isolated from users outside of the subnets that
contain the domain such as the internet? You should try to scan your
firewall yourself from outside the network, even if you use a self scan site
such as to see if it is blocking the traffic you

If legitimate users are trying to attack your computers you may have to see
if there is a user policy in force [or can be created] that you can user for
some type of disciplinary action after gathering info from audit logs and
such. I understand due to politics your hands may be tied to that approach
as many seem to be forced to tolerate more and more bad behavior these days.
Or you may have infected/compromised computers on the network and the user
does not even know they are infected in which case you should notify them
and block access from them until the problem is resolved. Firewalls are best
configured with a block all default rule and then you add the allowed
exceptions. If you are already using Languard you probably already have a
good idea of network risk assessment. Your thought sounds right on and in
step with the least privilege principle. --- Steve

"Bob Smith" <> wrote in message
> Steve,
> Thanks for the great info, I do have access to the firewall and I have
> used
> ipsec policies previously, I also run Languard to check against
> vulnerabilities, my major attacks are coming from the basic MS ports and
> how
> to isolate these ports without removing basic services, also we need to
> maintin management, my thought here is to allow management access to two
> subnets (server room and vpn subnets), however systems like domain
> controllers I will have to leave open (of course I have these locked down)
> to
> the community, I guess the answer here is to evaluate each system for the
> specific needs and isolate based on that info.
> Regards,
> Bob Smith
> "Steven L Umbach" wrote:
>> If you have access to the firewall, you might be able to configure what
>> IP
>> addresses can and can not access your network/servers and on what ports
>> using what protocols. If you can not access the firewall you can use
>> ipsec
>> filtering policy on your computers which is a policy that uses rules with
>> permit and block filter actions to act as a built in packet filtering
>> firewall. Ipsec policies are best when trying to configure for a subnet
>> range or a small range of IP addresses as you can not specify IP
>> addresses
>> "ranges" in an ipsec policy. You can also create an ipsec rule
>> "blacklist"
>> to add the IP address of attackers to block their access. Software
>> firewalls
>> such as the ones from Sygate could be another option. Depending on your
>> network layout [operating system, domain, etc] you may be able to
>> implement
>> ipsec negotiation security to block access from non domain computers or
>> domain computers that are not configured with at least a matching ipsec
>> client/respond policy. Ipsec can also use certificates for computer
>> authentication. Only Windows 2000/XP Pro/W2003 MS computers are ipsec
>> aware.
>> Ipsec negotiation polices also need to exempt domain controllers for
>> traffic
>> between domain members and domain controllers. The links below are about
>> ipsec.
>> Disable file and print sharing on any computers that do not need to offer
>> shares and do not need to be managed remotely via Computer Management or
>> command line tools that rely on the ports you mentioned. You also may be
>> able to take advantage of the user rights for "logon locally and deny
>> logon
>> locally" to restrict what users can access a computer, though that will
>> not
>> stop users from trying to make attempts to guess passwords. Such user
>> rights
>> and ipsec policies can be managed via Group Policy for consistent
>> application and ease of administration to larger number of computers. A
>> managed switch may be another option as they offer options such as mac
>> filtering and port isolation [HP Procurve] to further restrict access to
>> your network. Mac filtering can be spoofed but it would be another
>> barrier
>> to access and will deter most curious attackers. 802.1X switches are a
>> better access restricting option but they are not foolproof either and
>> require compatible operating systems, a Certificate Authority to issue
>> computer certificates, and an IAS server on the network. Also run the
>> Microsoft Baseline Security Analyzer on your computers to check for basic
>> vulnerabilities such as weak passwords, missing patches, and unneeded
>> services.--- Steve
>> --- MBSA.
>> "Bob Smith" <> wrote in message
>> > Due to the large number of attacks against Windows Server we would like
>> > to
>> > block windows systems from the larger community (Large college) to
>> > prevent
>> > systems from getting attack, does anyone have any help, suggestions,
>> > info
>> > for
>> > blocking ms port (135, 137, 139, & 445) from the community.
>> >
>> > Thanks in advance,
>> > Bob Smith