SANS Software Security Institute
SANS Software Security Institute
Home > Resources > Top-20

The Top 20 Most Critical Internet Security Vulnerabilities - (2004-2005 Archive)

Introduction

The SANS Top 20 Internet Security Vulnerabilities

The vast majority of worms and other successful cyber attacks are made possible by vulnerabilities in a small number of common operating system services. Attackers are opportunistic. They take the easiest and most convenient route and exploit the best-known flaws with the most effective and widely available attack tools. They count on organizations not fixing the problems, and they often attack indiscriminately, scanning the Internet for any vulnerable systems. The easy and destructive spread of worms, such as Blaster, Slammer, and Code Red, can be traced directly to exploitation of unpatched vulnerabilities.

Four years ago, the SANS Institute and the National Infrastructure Protection Center (NIPC) at the FBI released a document summarizing the Ten Most Critical Internet Security Vulnerabilities. Thousands of organizations used that list, and the expanded Top-20 lists that followed one, two, and three years later, to prioritize their efforts so they could close the most dangerous holes first. The vulnerable services that led to worms like Blaster, Slammer, and Code Red, as well as NIMDA worms - are on that list.

This SANS Top-20 2004 is actually two Top Ten lists: the ten most commonly exploited vulnerable services in Windows and the ten most commonly exploited elements in UNIX and Linux environments. Although there are thousands of security incidents each year affecting these operating systems, the overwhelming majority of successful attacks target one or more of these twenty vulnerable services.

The Top-20 is a consensus list of vulnerabilities that require immediate remediation. It is the result of a process that brought together dozens of leading security experts. They come from the most security-conscious government agencies in the UK, US, and Singapore; the leading security software vendors and consulting firms; the top university-based security programs; many other user organizations; and the SANS Institute. A list of participants may be found at the end of this document.

The SANS Top-20 is a living document. It includes step-by-step instructions and pointers to additional information useful for correcting the security flaws. We will update the list and the instructions as more critical threats and more current or convenient methods of protection are identified, and we welcome your input along the way. This is a community consensus document -- your experience in fighting attackers and in eliminating the vulnerabilities can help others who come after you. Please send suggestions via e-mail to top20@sans.org

Version 5.0 October 8, 2004
Copyright © 2005, SANS Institute
Questions / comments may be directed to top20@sans.org.
To link to the Top 20 List, use the "SANS Top 20 List" logo

PDF | Printer Friendly Version

Top Vulnerabilities in Windows Systems

W1. Web Servers & Services

W1.1 Description
Default installations of various HTTP servers and additional components for serving HTTP requests as well as streaming media to the internet from Windows platforms have proven vulnerable to a number of serious attacks over time. The impact of these vulnerabilities can include:
  • Denial of service
  • Exposure or compromise of sensitive files or data
  • Execution of arbitrary commands on the server
  • Complete compromise of the server
HTTP servers including IIS, Apache, and iPlanet (now SunOne) have had numerous issues that have been patched as they have been discovered. Ensure that all patches are up to date for the server and that a current version is running. In most HTTP server software the default configuration is rather open leaving large avenues for exploit. Whilst this has been changed to a 'secure by default' posture for IIS 6.0, it is crucial that administrators take the time to fully understand their web server and adjust the configuration to allow only those features and services required.

IIS uses a programming hook known as ISAPI to associate files having certain extensions with DLLs (known as ISAPI extensions). Preprocessors such as ColdFusion and PHP use ISAPI, and IIS includes many ISAPI extensions to handle functions such as Active Server Pages (ASP), .Net web services, and web-based printer sharing. Many ISAPI extensions installed by default with version 5.0 and earlier of IIS are not required in most installations, and many of those filters are exploitable. Examples of malicious programs that use this type of propagation mechanism include the well-known Code Red and Code Red 2 worms. Enable only the ISAPI extensions that the web server will need to recognize and restrict the HTTP options that can be used with each allowed ISAPI extension. This tightening of security is best achieved via the IIS LockDown tool available freely from Microsoft.

Most web servers include sample applications or web sites that were designed to demonstrate the functionality of the web server. These applications were not designed to operate securely in a production environment. Versions of IIS before 6.0 include sample applications that can be exploited to allow remote viewing or overwriting of arbitrary files as well as remote access to other sensitive server information, such as system configuration settings and paths to binaries. Remove these applications prior to placing the server into production.

Any webserver installation that is not regularly maintained is also subject to vulnerabilities discovered since the software release date. Examples include the PCT and SSL vulnerabilities that are addressed by the Microsoft patch MS04-011, which could allow a Denial of Service condition or allow the attacker to take control of the server. Timely patching of publicly accessible web servers is critical.

Third-party web add-ons such as ColdFusion and php can introduce further vulnerabilities in a webserver installation, either through misconfiguration or through vulnerabilities inherent in the product.

W1.2 Operating Systems Affected
Any Microsoft Windows system with a web server installed could be affected. This includes, but is not limited to:
  • Microsoft IIS: Windows NT4.0 and above, including XP Professional
  • Apache HTTP server: Windows NT4.0 SP3 and above are supported, though it is believed to run on Win95 and Win98
  • Sun Java System/Sun One/iPlanet Web Server: Windows NT4.0 SP6 and above
Please note: Windows 2000 Server ships with IIS installed by default, as many administrators discovered during the infamous Nimda and Code Red outbreaks. As part of the Trustworthy Computing initiative, Windows Server 2003 does not enable the IIS server in a standard installation, and the default settings are configured for security. Furthermore, some third-party applications require functionality provided by IIS, possibly resulting in administrators unknowingly installing this software. Never assume a network to be immune to web server attacks simply because no such server was intentionally installed; regularly audit networks for the presence of any "rogue" web servers. See "How to Determine if you are at risk" below for more information.

W1.3 Related CVE Entries
a. IIS
CAN-2003-0225, CAN-2003-0377, CAN-2003-0227, CAN-2003-0349, CERT-VU-288308, Secunia-12647, Secunia-12638, Secunia-11563

Searchable CVE entries for IIS 2.0, CVE entries for IIS 3.0, CVE entries for IIS 4.0, CVE entries for IIS 5.0. To date no security exposures have been identified in IIS 6.0

b. Apache
CAN-2003-0987, CAN-2003-0842, CVE-2004-0009, CVE-2004-0113, CVE-2003-0993, CAN-2004-0174, CAN-2004-0492, CAN-2004-0488, CAN-2004-0748, CAN-2004-0700, CAN-2004-0751, CAN-2004-0809, CAN-2004-0786, CAN-2004-0811

CVE-2003-0016, CVE-2003-0017, CAN-2003-0460, CAN-2003-0844, CAN-2004-0493

Apache modules: CAN-2003-0844, CAN-2004-0492

c. iPlanet/Sun
CAN-2003-0411, CAN-2003-0412, CAN-2003-0414, CAN-2003-0676

CAN-2002-1315, CAN-2002-1042, CVE-2002-0845, CAN-2003-0676

d. Add-ons
CAN-1999-0455, CAN-1999-0477, CAN-1999-1124, CAN-2001-0535, CAN-2001-1120, CAN-2002-1309, CAN-2003-0172

CVE-1999-0756, CVE-1999-0922, CVE-1999-0924, CVE-2000-0410, CVE-2000-0538

ColdFusion: CAN-2002-1309, CAN-2004-0407, CVE-2000-0189, CVE-2000-0382, CVE-2000-0410, CVE-2000-0538, CVE-2002-0576

PHP: CAN-2002-0249, CAN-2003-0172

e. Other Services
CAN-1999-1369, CAN-2003-0227, CAN-2003-0349, CAN-2003-0725, CVE-2003-0905

CVE-1999-0896, CVE-1999-1045, CVE-2000-0211, CVE-2000-0272, CVE-2000-0474, CVE-2000-1181, CVE-2001-0083, CAN-2001-0524

W1.4 How to Determine if you are at risk
Any default or unpatched web server installations should be presumed vulnerable.

Most web server and service vendors provide a wealth of information regarding current security issues. Examples include: Also check any web server and associated service's patch and software revision levels, including configurations, against the vendor-supplied security information and the CVE database on a regular basis to assess potential vulnerability. It is important to realize that new issues are discovered regularly and it is best practice to consult to make good use of the Windows Update website, Microsoft Security Baseline Analyzer and Automatic Updates feature to properly assess and eliminate potential vulnerabilities.

Some remote and local vulnerability assessment tools exist to aid web server administrators in auditing their networks, including: It is recommended that remote vulnerability assessment tools be run on a network-wide basis, rather than just against known servers, to assess potential vulnerability of "rogue" web server installations.

W1.5 How to protect against these vulnerabilities

For most systems
  1. Apply the latest service packs and security updates or the HTTP service as well as for the Operating System and any applications loaded on this same host. Once the patches are up-to-date, consider using the automatic update feature to enable a higher level of security.
  2. Install host-based anti-virus and Intrusion Detection software. Be sure to keep both current on patches and review the log files frequently.
  3. Disable unused script interpreters and remove their binaries. For example; perl, perlscript, vbscript, jscript, javascript, and php.
  4. Enable logging if it is an option and review the logs frequently, preferably through an automated process that summarizes the events and reports exceptions and suspicious events.
  5. Use a syslog-like system to store Operating System and HTTPd logs safely on another system.
  6. Remove or restrict the system tools that are commonly used by attackers to assist with both the initial compromise and expansion beyond the initial victim host. For example; tftp(.exe), ftp(.exe), cmd.exe, bash, net.exe, remote.exe, and telnet(.exe).
  7. Limit the applications running on the host to the HTTP service/daemon and its supporting services.
  8. Be aware of and minimize any vectors into the inner network that enter through public web server(s). For example, NetBIOS shares or trust relationships.
  9. Use different account naming conventions and unique passwords on public facing systems than on internal systems. Any information leakage from a public facing system should not aid an attack on the internal systems.
a. IIS
Consider upgrading your IIS installation to IIS 6.0, which offers dramatically increased security. Patching a server on installation is necessary but not sufficient. As new IIS weaknesses are uncovered, patch accordingly. Windows Update and Automatic Updates are options for single-server installations. Systems Management Server (SMS) and Software Update Services (SUS) are also very good options for managed environments or administrators that have responsibilities for multiple disparate systems. MBSA, the network security Hotfix checker, assists the system administrator in scanning local or remote systems for current patches. The tool works on Windows NT 4, Windows 2000, Windows XP and Windows 2003. The current version can be downloaded from Microsoft at http://www.microsoft.com/technet/security/tools/mbsahome.mspx.

Use IIS Lockdown Wizard to harden the installation
Microsoft has released a simple tool to aid in securing IIS installations known as the IIS Lockdown Wizard. The current version can be downloaded from Microsoft at http://www.microsoft.com/technet/security/tools/locktool.asp

Running the IIS Lockdown Wizard in "custom" or "expert" mode will allow the following recommended changes to be made to an IIS installation:
  • Ensure the latest version of WebDAV is employed on the server. IIS 6.0 allows administrators to select whether or not to enable WebDAV.
  • Unmap all unnecessary ISAPI extensions (including .htr, .idq, .ism, and .printer in particular).
  • Eliminate sample applications.
  • Restrict permissions and the availability of binaries commonly found on a webserver and often used as part of an attack and compromise (e.g. cmd.exe and tftp.exe).
The SANS Reading Room contains the papers Understanding and installing the IISlockdown tool and Securing a Windows 2000 IIS Web Server - Lessons Learned. The Microsoft Security Centre also contains a wealth of prescriptive guidance for protecting and managing IIS.

Use URLScan to filter HTTP requests
Many IIS exploits, including Code Blue and the Code Red family use maliciously formed HTTP requests in directory traversal or buffer overflow attacks. The URLScan filter can be configured to reject such requests before the server attempts to process them. The current version has been integrated into the IIS Lockdown Wizard, but can be downloaded separately from Microsoft at
http://www.microsoft.com/technet/security/tools/urlscan.mspx.

b. Apache

The issues of access control, restriction by IP and the Apache security modules, along with many other topics, are discusses on the Apache Tutorials page.

In addition, Securing Apache: Step-by-Step by Artur Maj is a very helpful paper found in the SANS Reading Room that covers in detail the tasks of securing an Apache server.

c. iPlanet/Sun One

Edmundo Farinas addresses securing iPlanet in his paper Security Considerations for the iPlanet Enterprise Web Server on Solaris which is located in the SANS Reading Room.

In addition, Sun provides the Sun ONE Application Server Security Goals paper which details the recommended steps for securing an iPlanet/Sun One server.

d. Add-ons

If third-party add-ons such as ColdFusion, PerlIIS, or PHP are used check the third-party vendors' web sites for patches and configuration tips as well. For obvious reasons, Microsoft does not include third-party patches in Windows Update and related update services.

For information on securing ColdFusion, see the SANS Reading Room paper Web Application Security, with a Focus on ColdFusion by Joseph Higgins

Located in the SANS Reading Room, Securing PHP: Step-by-step by Artur Maj illustrates the process of securing PHP applications.

In addition, a helpful resource is the PHP Manual, Chapter 16. Security, which addresses PHP security in detail.

e. Other services

While there are general steps listed above that can be taken to secure most web services, each usually has its own unique set of vendor supplied updates and patches, recommended configurations, and logging features.

Review the documentation including any information posted at the vendor's web site and make sure to sign-up for each vendor's notification service and newsletter. This will help to stay informed of relevant security issues and to address them quickly and effectively.

W2 Workstation Service

W2.1 Description
Windows Workstation service is responsible for processing user requests to access resources such as files and printers. The service determines if the resource resides on the local system or on a network share, and routes the user requests appropriately.

The network management functions provided by the service can be invoked via any of the following mechanisms.
  • DCE/RPC calls over SMB protocol after connecting to the service using \\pipe\wkssvc named pipe.
  • DCE/RPC calls directly over a UDP port (> 1024)
  • DCE/RPC calls directly over a TCP port (> 1024)
     Note that the service binds to the first available TCP and UDP port over 1024.

The Workstation service contains a stack-based buffer overflow that can be triggered by a specially crafted DCE/RPC call. The problem arises because parameters are passed to the logging function without any bounds checking. This overflow can be exploited by an unauthenticated remote attacker to execute arbitrary code on the vulnerable Windows machine with "SYSTEM" privileges. The attacker can obtain complete control of the compromised machine. The exploit code for leveraging the vulnerability has been posted to the Internet and was re-used in some variants of Phatbot/Gaobot worm that infected millions of systems world-wide.

W2.2 Operating Systems Affected
Windows 2000 SP2, SP3 and SP4
Windows XP, Windows XP SP1
Windows XP 64 Bit Edition


W2.3 CVE/CAN Entries
CAN-2003-0812, CAN-2003-0813, CAN-2003-0352

W2.4 How to Determine if you are Vulnerable
Systems running Windows 2000 without the MS03-049 patch and Windows XP without the MS03-043 patch are vulnerable. Windows XP users that have installed Service Pack 2 are protected.

Check for the following registry-entries:
KB828035: Under HKLM\Software\Microsoft\Updates\Windows XP (Windows XP)
KB828749: Under HKLM\Software\Microsoft\Updates\Windows 2000 (Windows 2000)
If these registry entries are not found the Windows system may be vulnerable. For greater certainty and support in mitigating this risk, use a security scanner such as Microsoft Baseline Security Analyzer (MBSA) to check if the appropriate update has been installed. MBSA can be downloaded from http://www.microsoft.com/technet/security/tools/mbsahome.mspx

W2.5 How to Protect Against It
  1. Windows XP Service Pack 2 offers many security enhancements that protect against these and other security risks. This should be a priority for Windows XP installations and is recommended by the US-CERT.
  2. Ensure that Windows systems have all the latest security patches and service packs are installed. The configuration of Automatic Updates should be viewed as a necessity, tailored to fit individual or corporate requirements. Specifically ensure that Windows 2000 systems have MS03-049 and Windows XP systems have MS03-043 patch installed. As per the previous point, Service Pack 2 should be considered vital.
  3. Block the ports 139/tcp and 445/tcp at the network perimeter. This prevents a remote attacker from exploiting the overflow via SMB.
  4. Open only the necessary TCP ports over 1024 at the network perimeter. This prevents a remote attacker from exploiting the overflow via DCE/RPC calls. Note that it is difficult to block UDP ports above 1024 at the firewall as the ports in this range are used as ephemeral ports.
  5. Use TCP/IP Filtering available in both Windows 2000 and XP, or Windows Firewall in Windows XP systems to block inbound access to the affected ports. The Windows Firewall also offers network administrators the ability to centrally enforce policies and settings across end-user systems that can help heighten security.
  6. For Third-party applications running on customized Windows 2000/XP platforms ensure that an appropriate patch from the vendor has been applied. For example, Cisco has released an advisory stating that a number of Cisco products are vulnerable to this overflow. Cisco has also provided the patches.
  7. If the system is stand-alone (i.e. does not belong to a Windows network environment), the Workstation service can be disabled, however, this must be done with caution as it can affect applications and system functionality.
Additional information:

Microsoft Advisory
http://www.microsoft.com/technet/security/bulletin/MS03-049.mspx

eEye Advisory
http://www.eeye.com/html/Research/Advisories/AD20031111.html

CERT Advisories
http://www.cert.org/advisories/CA-2003-28.html
http://www.kb.cert.org/vuls/id/567620

CORE Security Advisory
http://archives.neohapsis.com/archives/vulnwatch/2003-q4/0066.html

Cisco Advisory
http://www.cisco.com/warp/public/707/cisco-sa-20040129-ms03-049.shtml

Gaobot Worm
http://securityresponse.symantec.com/avcenter/venc/data/w32.hllw.gaobot.gen.html

W3 Windows Remote Access Services

W3.1 Description
The family of Windows Operating systems supports a variety of different networking methods and technologies. There is native support for most industry standard networking protocols and built-in functionality for many Microsoft specific networking methods and techniques. Common avenues for exploitation include Network Shares, Anonymous Logon, remote registry access, and remote procedure calls.

NETBIOS - A set of API's that can allow the sharing files or folders across a network with other hosts through Windows network shares. The underlying mechanism of this feature is the Server Message Block (SMB) protocol, or the Common Internet File System (CIFS). These protocols permit a host to manipulate remote files just as if they were local.

Although this is a powerful and useful feature of Windows, improper configuration of network shares may expose critical system files or may provide a mechanism for a nefarious user or program to take full control of the host. One of the ways in which I-Worm.Klez.a-h (Klez Family) worm, Sircam virus (see CERT Advisory 2001-22) and Nimda worm (see CERT Advisory 2001-26) spread so rapidly in 2001 was by discovering unprotected network shares and placing copies of themselves in them. Many computer owners unknowingly open their systems to hackers when they try to improve convenience for co-workers and outside researchers by making their drives readable and writeable by network users. But when care is taken to ensure proper configuration of network shares, the risks of compromise can be adequately mitigated.

Anonymous Logon - An anonymous session is a communication link established without correct credentials (i.e. blank username and password). Null sessions can be used to display information about users, groups, shares and password policies. Microsoft Windows NT services running as the Local System account on the local computer communicate with other services over the network by establishing null sessions. Windows 2000 and later services running as the Local System account on the local computer use the local computer account to authenticate to other servers.

Remote Registry Access - Microsoft Windows 9x, Windows CE, Windows NT, Windows 2000, Windows 2003, Windows ME and Windows XP employ a central hierarchical database, known as the Registry, to manage software, device configurations, and user settings. Improper permissions or security settings can permit remote registry access or execution of code or applications that should not be allowed to run.

Remote Procedure Calls - All versions of Microsoft operating systems (Windows NT 4.0, 2000, XP, and 2003) provide an inter-process communication mechanism that allows programs running on one host to execute code on remote hosts. Three vulnerabilities have been published that would allow an attacker to run arbitrary code on susceptible hosts with Local System privileges. One of these vulnerabilities was exploited by Blaster/MSblast/LovSAN and Nachi/Welchia worms. There are also other vulnerabilities that would allow attackers to mount Denial of Service attacks against RPC components.

W3.2 Operating Systems Affected
Windows 95, Windows 98, Windows NT Workstation and Server, Windows Me, Windows 2000 Workstation and Server, Windows XP Home and Professional, and Windows 2003 are all potentially vulnerable.

Windows XP Service Pack 2 changed the behaviour of RPC. A new RPC Interface Restriction was implemented to make it more secure by default. Particularly noteworthy are the addition of a new registry key - RestrictRemoteClients. This key modifies the behaviour of all RPC interfaces on the system and will, by default, eliminate remote anonymous access to RPC interfaces on the system, effectively removing this risk.

W3.3 CVE/CAN Entries


NETBIOS
CVE-2000-0979 (Windows 95, 98, and Windows Me - ONLY), CAN-2003-0661

CAN-1999-0518, CAN-1999-0519, CAN-1999-0621, CAN-2000-1079

Anonymous Logon
CVE-2000-1200

Remote Registry Access
CAN-2000-0377, CVE-2002-0049

CAN-1999-0562, CAN-2001-0045, CAN-2001-0046, CAN-2001-0047, CVE-2002-0642, CVE-2002-1117

Remote Procedure Calls
CAN-2002-1561, CVE-2003-0003, CAN-2003-0352, CAN-2003-0528, CAN-2003-0605, CAN-2003-0715, CAN-2001-0509, CAN-2003-0813

W3.4 How to Determine if you are Vulnerable


How to determine if you are vulnerable to NETBIOS related issues.

A number of tools are available that can help to determine if there are NETBIOS related vulnerabilities on a system.

NbtScan - NetBIOS Name Network explores the NETBIOS file-sharing services available on target systems NbtScan is available at: http://www.inetcat.org/software/nbtscan.html.

NLtest - extremely powerful tool, included in Windows 2000 and 2003 Support Tools (can be found on product CD) and Windows NT4 Resource Kit. NLtest can obtain a wealth of information about potential configuration vulnerabilities.

For Windows NT (SP4), Windows 2000, Windows XP, and Windows 2003, the Microsoft Baseline Security Analyser will report hosts that are vulnerable to SMB exploits and may be used to fix the problem. The tests can be run locally or on remote hosts.

Windows NT, Windows 2000, Windows XP, and Windows 2003 users can simply type net share from the command prompt to see what resources are being shared. For more information about the net share command, type net share /?.

IMPORTANT Note: This article contains information about modifying shared resources. Before modifying any shared resource, make that it is understood how to restore the resource if a problem occurs. It is recommended that any modifications are thoroughly tested before implementation in a production environment. For information about shared resources, click the following article numbers to view the article in the Microsoft Knowledge Base:

125996 - Saving and Restoring Existing Windows Shares

308419 - HOW TO Set, View, Change, or Remove Special Permissions for Files and Folders in Windows XP

307874 - HOW TO Disable Simplified Sharing and Password-Protect a Shared Folder in Windows XP

174273 - How to Copy Files and Maintain NTFS and Share Permissions

Although File System permissions and settings will take priority, the default permissions on new shares are detailed below:

Windows NT, Windows 2000, and Windows XP (Pre Service Pack 1)
  • Everyone - Full Control
Windows XP SP1
  • Everyone - Read
Windows XP by default has one shared directory called "SharedDocs." The physical location of this share is:

C:\Documents and Settings\All Users.WINDOWS
  • The owner of the file or folder and local Computer Administrators have read and write permission to the file or folder. Nobody else may read or write to the folder or the files in it. This is the default setting for all the folders and files in each user's My Documents folder.
Most commercially-available network-based scanners will detect open shares. A quick and effective test for SMB exposures can be found at the Gibson Research Corporation web site, although the accuracy of the results is dependant upon the host system not being located behind a firewall or screening device.

Automated Scanning tools to detect share vulnerabilities: How to determine if you are vulnerable to Anonymous Logon related issues.
Try to establish a null session to the computer by issuing the following command from the command prompt (Start --> Run --> type cmd):

C:\>net use \\ipaddress\ipc$ "" /user:""

The preceding syntax connects to the hidden interprocess communications "share (IPC$) at ipaddress (/user:"") with a null () password.

If "The command completed successfully" is received, then the system is potentially vulnerable to remote interrogation and account enumeration.

The list of tools above, including Nessus and Winfingerprint, can also be used to detect null session vulnerabilities.

How to determine if you are vulnerable to Remote Registry Access related issues.
NT Resource Kit (NTRK) formerly available from Microsoft contains an executable file entitled Regdump.exe that will passively test remote registry access permissions from a Windows NT host against other Windows NT/Windows 2000 or Windows XP hosts on the Internet or internal network.

How to determine if you are vulnerable to Remote Procedure Call related issues.
Microsoft has made a hotfix, configuration, and patch-checking tool freely available for download; this is probably the best way to determine if Windows hosts are susceptible to any of these vulnerabilities. It is called the Microsoft Baseline Security Analyzer (MBSA) and is available from http://www.microsoft.com/technet/security/tools/mbsahome.mspx

There is also a standalone scanning tool that will check for missing security patches for CAN-2003-0352, CAN-2003-0528, CAN-2003- 0605 and CAN-2003-0715 only; it is available from http://support.microsoft.com/?kbid=827363. However, it is encouraged to use the MBSA, which has a wider coverage. Home or small-scale users with only a few computers to take care of will probably find it easier to visit the Windows Update site at http://windowsupdate.microsoft.com/ and check individual machines for outdated software.

W3.5 How to Protect Against It


Microsoft addresses security vulnerabilities in Service Packs and security hotfixes for Operating systems and applications. It is extremely important to have the most current Service Pack installed on a system. For example, the Sasser worm and its clones (exploiting vulnerability of LSASS system) infected a lot of unpatched systems worldwide, while systems that had hotfix MS04-011 installed were immune to this extremely dangerous vulnerability. Microsoft had hotfix MS04-011 released a few weeks prior to appearance of the Sasser worm.

NOTE: Windows 95 and Windows NT4 Workstation are no longer supported by Microsoft. Support for Windows NT4 Server expires on December 31, 2004.

For details of lifecycle for supported operating systems and products see Microsoft article Product Lifecycle Dates - Windows Product Family.

For finding relevant security hotfixes for a system, use:
  • Windows Update service. It automatically detects all required security hotfixes on the system and installs them after the user selects (approves) the hotfixes that need to be installed
  • Enable the Automatic Updates feature to provide enhancements to the operating system and applications as they are released by Microsoft.
  • Windows Security Bulletin Search online service located at: http://www.microsoft.com/technet/security/current.aspx
While having current service packs and security hotfixes addresses many software design-related problems (such as buffer overflows, code design errors etc), there are a number of dangerous features in Windows OS that have legitimate and documented functionality, but can be safely disabled or secured in many cases in order to harden system security. To better understand and highlight potential security exposures or risks, employ the Microsoft Baseline Security Analyzer (MBSA).

How to protect against NETBIOS related attacks.
Several actions can be taken to mitigate the risk of exploitation of vulnerability through Windows Networking. NOTE: Extra care must be taken before disabling sharing or netbios facilities as these can have adverse effects on enterprise applications and services. In all circumstances, ensure that the changes are effectively tested before being implemented into a production environment.

If the system does not need to provide file/print services and does not need to be remotely administered (most home workstations fit into this category), the Server service can be disabled.

On Windows NT4/2000/2003/XP systems disable service Server by selecting Start - Programs - Administrative Tools - Services - select service Server - double-click it - set Startup type to value Disabled - press button Apply - press button Stop - press button OK.

If the system does require service Server running, it is recommended that systems are configured in line with current Best Practice outlined at the Microsoft Security Guidance Center. In addition, the following steps can be made to secure Windows NT4/2000/2003/XP systems:
  1. Enumerate all default hidden shares ( C$, D$, E$ etc) by typing command:

    Net share

    From system command prompt. Make note of existing shares.
  2. Delete default hidden shares. Note that removing hidden shares will often break enterprise applications such as backup and management applications. To ensure the shares remain deleted following a reboot, the adjustments to the registry (outlined in following steps) must be also undertaken. To delete the hidden shares, issue the following command:

    Net share C$ /delete

    from system command prompt. In most cases all alphabet shares (C$, D$, E$ etc) and share ADMIN$ can be safely deleted. It is not recommended to delete default share IPC$ on any system.
  3. In order to make deletion of default shares permanent (they would be restored automatically on system restart or restart of service Server), it is necessary to make following Registry modifications:
    • Open Registry editor;
    • Navigate to Registry key:
      HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\
      lanmanserver\parameters
    • Create new Registry value under this key:
    • Value name: AutoShareWks
    • Value type: DWord
    • Value: 00000000
    • Create new Registry value under this key:
    • Value name: AutoShareServer
    • Value type: DWord
    • Value: 00000000
Review existing non-default (custom-created) shares on system. That can be done through:
  • Graphical interface (My Computer - right-click - Manage - Shared Folders - Shares). Select shares that need to be disabled - right-click - select Stop Sharing.
  • Command line (from system prompt or as part of any script):
    • Enumerate all shares by typing command:

      Net share

      From system command prompt. Make note of existing shares.
  • Delete unnecessary shares by typing command:

    Net share ShareName /delete

    from system command prompt.
That will permanently delete non-default (custom-created) shares only. For permanent deletion of default hidden shares C$, D$, ADMIN$ see procedures in previous paragraph.
  • Windows 95/98/Me clients that are a part of a Windows NT domain are recommended to be setup with user-level file share access controls.
  • Do not permit sharing with hosts on the Internet. Ensure all Internet-facing hosts have Windows network shares disabled in the Windows network control panel. File sharing with Internet hosts should be achieved using SCP, FTP ,or HTTP.
  • Do not permit unauthenticated shares. If file sharing is required, then do not permit unauthenticated access to a share. Configure the share so a password is required to connect to the share.
  • Restrict shares to only the minimum folders required. Shares should be generally only one folder and possibly sub-folders of that folder.
  • Restrict permissions on shared folders to the minimum required. Be especially careful to only permit write access when it is absolutely required.
  • For added security, allow sharing only to specific IP addresses as DNS names can be spoofed.
How to protect against Anonymous Logon problems on your systems. IMPORTANT Note: This article contains information about modifying the registry. Before modifying the registry, make sure to back it up and make sure that it is understood how to restore the registry if a problem occurs. It is recommended to thoroughly test any modifications before implementation in a production environment. For information about how to back up, restore, and edit the registry, click the following article numbers to view the article in the Microsoft Knowledge Base:

256986 - Description of the Microsoft Windows Registry
 323170 - HOW TO Backup, Edit, and Restore the Registry in Windows NT 4.0
 322755 - HOW TO Backup, Edit, and Restore the Registry in Windows 2000
 322756 - HOW TO Backup, Edit, and Restore the Registry in Windows XP Windows Server 2003

Windows NT Domain controllers require null sessions to communicate. Therefore, if working in a Windows NT domain or Windows 2000/2003 Active Directory running in mixed mode, which allows Pre-Windows 2000 compatible access, it is possible to minimize the information that attackers can obtain, but not stop all leakage by setting the RestrictAnonymous registry value to 1. For example; GetAcct from Security Friday sidesteps RestrictAnonymous=1 and will enumerate the SID and UserID. The ideal solution with a native Windows 2000/2003 Active Directory is to set the RestrictAnonymous registry value to 2.

To restrict information available via null sessions, click the following article numbers to view the articles in the Microsoft Knowledge Base:

143474 - Restricting Information Available to Anonymous Logon Users in Windows NT
 246261 - How to Use the RestrictAnonymous Registry Value in Windows 2000

To troubleshoot the RestrictAnonymous registry value, click the following article number to view the article in the Microsoft Knowledge Base:

296405 - The RestrictAnonymous Registry Value May Break the Trust to a Windows 2000 Domain

Windows NT:
  1. Start Registry Editor "regedit.exe" and go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Contro\Lsa
  2. Set the following registry value:
    Name: RestrictAnonymous
    Type: REG_DWORD Value: 1
  3. Restart your computer.
Windows 2000:
  1. Start "Control Panel-->Administrative Tools-->Local Security Policy".
  2. Open "Local Policies-->Security Options".
  3. Make sure "Additional restrictions of anonymous connections" is set to "No access without explicit anonymous permissions".
  4. Restart your computer.
Windows XP:
  1. Start "Control Panel-->Administrative Tools-->Local Security Policy".
  2. Open "Local Policies-->Security Options".
  3. Make sure the following two policies are enabled:
    • Network Access: Do not allow anonymous enumeration of SAM accounts
    • Network Access: Do not allow anonymous enumeration of SAM accounts and shares
  4. Restart your computer.
How to protect against Remote Registry Access on your systems.
To address this threat, access to the system registry must be restricted and the permissions set for critical registry keys reviewed. Users of Microsoft Windows NT 4.0 should also ensure that Service Pack 4 (SP4) or later has been installed before adjusting the registry.

Important Note: This article contains information about modifying the registry. Before modifying the registry, make sure to back it up and make sure that it is understood how to restore the registry if a problem occurs. It is recommended to thoroughly test any modifications before implementation in a production environment. For information about how to back up, restore, and edit the registry, click the following article numbers to view the article in the Microsoft Knowledge Base:

256986 - Description of the Microsoft Windows Registry
 323170 - HOW TO Backup, Edit, and Restore the Registry in Windows NT 4.0
 322755 - HOW TO Backup, Edit, and Restore the Registry in Windows 2000
 322756 - HOW TO Backup, Edit, and Restore the Registry in Windows XP Windows Server 2003

Restrict Network Access. To restrict network access to the registry, follow the steps listed below to create the following Registry key:
  • HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\ Control\ SecurePipeServers\winreg
  • Description: REG_SZ
  • Value: Registry Server
Security permissions set on this key define the Users or Groups that are permitted remote Registry access. Default Windows installations define this key and set the Access Control List to provide full privileges to the system Administrator and Administrators Group (and Backup Operators in Windows 2000).

Changes to the system registry will require a reboot to take effect. To create the registry key to restrict access to the registry:

For Windows 2000 and NT:
  1. Start Registry Editor "regedit.exe" and go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control
  2. On the "Edit" menu, click "Add Key."
  3. Enter the following values: Key Name: SecurePipeServers Class: REG_SZ
  4. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\SecurePipeServers
  5. On the "Edit" menu, click "Add Key."
  6. Enter the following values: Key Name: winreg Class: REG_SZ
  7. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ SecurePipeServers\winreg
  8. On the "Edit" menu, click "Add Value."
  9. Enter the following values: Value Name: Description Data Type: REG_SZ String: Registry Server
  10. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ SecurePipeServers\winreg
  11. Select "winreg." Click "Security" and then click "Permissions." Add Users or Groups to which to grant access.
  12. Exit Registry Editor and restart Microsoft Windows.
  13. If at a later stage it is required to change the list of users that can access the registry, repeat steps 10-12.
For Windows XP and 2003:
  1. Start Registry Editor "regedt32.exe" and go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control
  2. On the "Edit" menu, click "Add Key."
  3. Enter the following values: Key Name: SecurePipeServers Class: REG_SZ
  4. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\SecurePipeServers
  5. On the "Edit" menu, click "Add Key."
  6. Enter the following values: Key Name: winreg Class: REG_SZ
  7. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ SecurePipeServers\winreg
  8. On the "Edit" menu, click "Add Value."
  9. Enter the following values: Value Name: Description Data Type: REG_SZ String: Registry Server
  10. Go to the following subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ SecurePipeServers\winreg
  11. Select "winreg." Click "Edit" and then click "Permissions." Add Users or Groups to which to grant access.
  12. Exit Registry Editor and restart Microsoft Windows.
  13. If at a later stage it is required to change the list of users that can access the registry, repeat steps 10-12.
Limit Authorized Remote Access. Enforcing strict restrictions upon the registry can have adverse side effects upon dependent services, such as the Directory Replicator and the network printer Spooler service.

It is therefore possible to add a degree of granularity to the permissions by adding either the account name that the service is running under to the access list of the "winreg" key or by configuring Windows to bypass the access restriction to certain keys by listing them in the Machine or Users value under the AllowedPaths key:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ SecurePipeServers\winreg\AllowedPaths
Value: Machine
Value Type: REG_MULTI_SZ - Multi string
Default Data: System\CurrentControlSet\Control\ProductOptionsSystem\ CurrentControlSet\Control\Print\PrintersSystem\CurrentControlSet\ Services\EventlogSoftware\Microsoft\WindowsNT\CurrentVersionSystem\ CurrentControlSet\Services\Replicator
Valid Range: (A valid path to a location in the registry)
Description: Allow machines access to listed locations in the registry provided that no explicit access restrictions exist for that location.
Value: Users
Value Type: REG_MULTI_SZ - Multi string
Default Data: (none)
Valid Range: (A valid path to a location in the registry)
Description: Allow users access to listed locations in the registry provided that no explicit access restrictions exist for that location.

In the Microsoft Windows 2000 and Windows XP Registry:
Value: Machine
Value Type: REG_MULTI_SZ - Multi string
Default Data: System\CurrentControlSet\Control\ProductOptionsSystem\ CurrentControlSet\Control\Print\PrintersSystem\CurrentControlSet\ control\Server ApplicationSystem\CurrentControlSet\Services\Eventlog\ Software\Microsoft\Windows NT\CurrentVersion
Value: Users (does not exist by default)


It is common for there to be a lag between the time a vulnerability becomes public and the time a patch is made available. Or for other policy reasons, the vulnerability must be allowed. To mitigate the risk an organization can limit access through firewalls or routers. An additional measure would be to write new rules for an IDS (Intrusion Detection System) like Snort which would alert or trigger a response by the organization. Examples of documented rules for Snort are located here.

How to protect against Remote Procedure Call-related issues on your systems.

The best way by far is to apply the relevant patches as identified by the MBSA, Automatic Update features or Windows Update: see "How to determine if you are vulnerable to Remote Procedure Call related issues" above. Failing that there are a number of ways to disable or restrict some Remote Procedure Call functionality, and some can be found in the excellent synopsis at http://www.ntbugtraq.com/dcomrpc.asp.

BE WARNED: disabling or restricting Remote Procedure Call functionality may break Windows services that are relied on, so always test any modifications in a non-production environment first.

If systems cannot be patched, then certainly block ports associated with Windows remote procedure calls (TCP ports 135, 139, 445 and 593; UDP ports 135, 137, 138 and 445) at the network perimeters. Of course, it is always best practice to block *all* unnecessary services at the network perimeter by default.

For more information:

Microsoft Knowledge Base Article 153183. How to Restrict Access to NT Registry from a Remote Computer.

Another source is Microsoft Security Bulletin Search.

MSDN Library (Search for Securing Registry)

Microsoft Knowledge Base Article 310426 : HOW TO: Use the Windows XP and Windows Server 2003 Registry Editor

Network access: Remotely accessible registry paths and subpaths

Windows Server 2003 Security Guide

W4 Microsoft SQL Server (MSSQL)

W4.1 Description
The Microsoft SQL Server (MSSQL) contains several serious vulnerabilities that allow remote attackers to obtain sensitive information, alter database content, compromise SQL servers, and, in some configurations, compromise server hosts.

MSSQL vulnerabilities are well-publicized and actively under attack. Two recent MSSQL worms in May 2002 and January 2003 exploited several known MSSQL flaws. Hosts compromised by these worms generate a damaging level of network traffic when they scan for other vulnerable hosts. Additional information on these worms can be found at

SQLSnake/Spida Worm (May 2002) SQL-Slammer/SQL-Hell/Sapphire Worm (January 2003) Port 1433 and 1434 (MSSQL server and monitor default ports) have also been regularly registered as two of the most frequently scanned ports by the Internet Storm Centre.

SQLSnake's exploit routine depends on the default administrative account, or "sa" account, having a null password. It is essential to the proper configuration and defence of any system to ensure that all system accounts are password protected, or completely disabled if not in use. You can find more information regarding setting and managing sa account passwords in the Microsoft Developer Network documentation under Changing the SQL Server Administrator Login, as well as Verify and Change the System Administrator Password by Using MSDE. The sa account should have a complex, hard-to-guess password even if it is not used to run your SQL/MSDE implementation.

SQL Slammer's exploit routine is based upon a buffer overflow in the SQL Server Resolution Service. This buffer overflow is brought to bear and host security is thus compromised when the worm sends crafted attack packets to UDP port 1434 of vulnerable target systems. If a machine runs SQL services that are subject to this stack buffer overflow and it receives packets of this nature, it will usually result in total server and system security compromise. The most effective means of defence against this worm is diligent patching, proactive system configuration practices, and ingress/egress UDP port 1434 filtering at network gateways.

The Microsoft Server 2000 Desktop Engine (MSDE 2000) can be thought of as "SQL Server Lite". Many system owners don't even realize that their systems are running MSDE and that they have a version of SQL Server installed. MSDE 2000 is installed as a part of the following Microsoft products:
  • SQL/MSDE Server 2000 (Developer, Standard and Enterprise Editions)
  • Visual Studio .NET (Architect, Developer and Professional Editions)
  • ASP.NET Web Matrix Tool
  • Office XP
  • Access 2002
  • Visual Fox Pro 7.0/8.0
In addition there are many other software packages that make use of the MSDE 2000 software. For an up-to-date list please check http://www.SQLsecurity.com/applicationslistgridall.aspx. Since this software uses MSDE as its core database engine, it has the same vulnerabilities as SQL/MSDE Server. MSDE 2000 can be configured to listen for incoming client connections in a multitude of different ways. It can be configured such that clients can use named pipes over a NetBIOS session (TCP port 139/445) or sockets with clients connecting to TCP port 1433, or both. Whichever method is used, SQL Server and MSDE will always listen on UDP port 1434. This port is designated as a monitor port. Clients will send a message to this port to dynamically discover how the client should connect to the server.

The MSDE 2000 engine returns information about itself whenever presented with the single byte packet 0x02 on UDP port 1434. Other single byte packets cause a buffer overflow without ever having to authenticate to the server itself. What further exacerbates these issues is that the attack is channelled over UDP. Whether the MSDE 2000 process runs in the security context of a domain user or the local SYSTEM account, successful exploitation of these security holes may mean a total compromise of the target system.

Since SQL Slammer exploits a buffer overflow on the target system, following best practices of timely patching and conscientious system configuration helps to mitigate this threat. By downloading and using defensive tools such as the Microsoft SQL Critical Update Kit, one can check local systems for vulnerability to this exploit, scan entire domains or networks for the existence of vulnerable systems, and automatically update affected files with SQL Critical Update.

Please see the report and analysis on incidents.org for more details on the SQL/MSDE Slammer worm. This particular attack affected the Internet Backbone for a few hours on the morning of January 25, 2003.

W4.2 Operating Systems Affected
Any Microsoft Windows system with Microsoft SQL/MSDE Server 7.0, Microsoft SQL/MSDE Server 2000 or Microsoft SQL/MSDE Server Desktop Engine 2000 installed, as well as any system which uses the MSDE engine separately.

W4.3 CVE/CAN Entries
CVE-2000-0202, CVE-2000-0402, CVE-2000-0485, CVE-2000-0603,CVE-2001-0344, CVE-2001-0879, CAN-2002-0649

CVE-2002-0186, CVE-2002-0187, CAN-2002-0224, CAN-2002-0624, CAN-2002-0641, CVE-2002-0642,CAN-2002-0643, CAN-2002-0644, CAN-2002-0645, CAN-2002-0649, CVE-2002-0650,CAN-2002-0695, CAN-2002-0721, CVE-2002-0729, CVE-2002-0859, CAN-2002-0982,CVE-2002-1123, CVE-2002-1137, CVE-2002-1138, CAN-2002-1145, CAN-2003-0118, Secunia-12680

W4.4 How to Determine if you are Vulnerable
Microsoft has published a set of security tools at http://www.microsoft.com/sql/downloads/securitytools.asp. The toolkit named the SQL Critical Update Kit contains valuable tools such as SQL Scan, SQL Check, and SQL Critical Update.

Chip Andrews of sqlsecurity.com released a tool called SQLPingv2.2. This tool sends a single byte UDP packet (byte value of 0x02) to port 1434 of either a single host or an entire subnet. SQL Servers listening on UDP 1434 will respond by divulging system details such as version number, instances, etc. SQLPingv2.2 is considered a scanning and discovery tool much like Microsoft's SQL Scan, and will not further compromise your system and network security. Additional SQL security tools can be found at Chip Andrew's SQL/MSDE Security Web site.

W4.5 How to Protect Against It
Summary:
  1. Disable SQL/MSDE Monitor Service on UDP Port 1434 (appreciate that this might interfere with remote administration or backup services).
  2. Apply the latest service pack for Microsoft SQL/MSDE server and/or MSDE 2000.
  3. Apply the latest cumulative patch that is released after the latest service pack.
  4. Apply any individual patches that are released after the latest cumulative patch.
  5. Enable SQL Server Authentication Logging.
  6. Secure the server at system and network level.
  7. Minimize privileges of the MSSQL/MSDEServer service and SQL/MSDE Server Agent.
  8. Take Best Practice guidance on securing this infrastructure at Microsoft.
Detail:
  1. Disable the SQL/MSDE Server Monitor on UDP Port 1434.

    This can be easily accomplished by installing and using the functionality within SQL Server 2000 Service Pack 3a. Microsoft's database engine MSDE 2000 exhibits two buffer overflow vulnerabilities that can be exploited by a remote attacker without ever having to authenticate to the server. What further exacerbates these issues is that the attack is channeled over UDP. Whether the MSDE 2000 process runs in the security context of a domain user or the local SYSTEM account, successful exploitation of these security holes may mean a total compromise of the target system. MS-SQL/MSDE Slammer sends a 376 byte long UDP packet to port 1434 using random targets at a very high rate. Compromised systems will immediately start sending identical 376 byte packets once they are infected. The worm sends traffic to random IP addresses, including multicast IP addresses, causing a Denial of Service on the target network. Single infected machines have reported traffic in excess of 50 Mb/sec after being infected.
  2. Apply the latest service pack for Microsoft SQL/MSDE server and MSDE 2000.

    The current Microsoft SQL/MSDE Server service pack version is:
    • SQL/MSDE Server 7.0 Service Pack 4
    • MSDE/SQL Server 2000 Service Pack 3a
    To ensure that you are current with any future upgrades, monitor Make Your SQL/MSDE Servers Less Vulnerable from Microsoft TechNet.
  3. Apply the latest cumulative patch that is released after the latest service pack.

    The current cumulative patch for all versions of SQL/MSDE/MSDE Server is available at MS02-061 Elevation of Privilege in SQL/MSDE Server Web Tasks (Q316333/Q327068).

    To ensure that you are current with any future upgrades, you can check for the latest cumulative patch for Microsoft SQL/MSDE Server at: Apply any individual patches that are released after the latest cumulative patch and enable the Automatic Update feature and consider subscribing the Microsoft Security Notification Service. Currently, there is no individual patch after the release of the MS02-061 Elevation of Privilege in SQL/MSDE Server Web Tasks (Q316333/Q327068).
  4. Enable SQL Server Authentication Logging.

    SQL Server Authentication Logging is commonly not enabled. This can be done through Enterprise Manager (Server properties; tab Security).
  5. Secure the server at system and network level.

    One of the most commonly attacked MSSQL/MSDE exposures is that the default administrative account (known as "sa") is installed with a blank password. If your SQL/MSDE "sa" account is not password-protected, you effectively have no security and can be affected by worms and other exploits. Therefore, you should follow the recommendation from the "System Administrator (SA) Login" topic in SQL/MSDE Server Books Online to make sure that the built-in "sa" account has a strong password, even if your SQL/MSDE server does not run using this account. Microsoft Developer's Network has documentation on Changing the SQL Server Administrator Login and how to Verify and Change the System Administrator Password by Using MSDE.
  6. Minimize privileges of the MSSQL/MSDEServer service and SQL/MSDE Server Agent.

    Run the MSSQL/MSDEServer service and SQL/MSDE Server Agent under a valid domain account with minimal privileges, not as a domain administrator or the SYSTEM (on NT) or LocalSystem (on 2000 or XP) account. A compromised service running with local or domain privileges would give an attacker complete control of your machine and/or your network.
    1. Enable Windows NT Authentication, enable auditing for successful and failed logins, and then stop and restart the MSSQL/MSDEServer service. If possible, configure your clients to use NT Authentication.
    2. Packet filtering should be performed at network borders to prohibit specifically non-authorized inbound or outbound connections to MSSQL specific services. Ingress and egress filtering of TCP/UDP ports 1433 and 1434 could prevent internal or external attackers from scanning and or infecting vulnerable Microsoft SQL/MSDE servers on your network or the networks of others that are not explicitly authorized to provide public SQL/MSDE services.
    3. If TCP/UDP ports 1433 and 1434 need to be available on your Internet gateways, enable and customize egress/ingress filtering to prevent misuse of this port.
Additional information on securing Microsoft SQL/MSDE Server can be found at

W5 Windows Authentication

W5.1 Description
Passwords, pass-phrases and security codes are used in virtually every interaction between users and information systems. Most forms of user authentication, as well as file and data protection, rely on user-supplied passwords. Since properly authenticated access is often not logged, or even if logged not likely to arouse suspicion, a compromised password is an opportunity to explore a system from the inside virtually undetected. An attacker would have complete access to any resources available to that user, and would be significantly closer to being able to access other accounts, nearby machines, and perhaps even administrative privileges. Despite this threat, accounts with bad or empty passwords remain extremely common, and organizations with good password policy are far too rare.

The most common password vulnerabilities are:
  • User accounts have weak or nonexistent passwords.
  • Regardless of the strength of their password, users fail to protect it.
  • The operating system or third-party applications create accounts with weak or nonexistent passwords.
  • In many commercial and Open Source applications, the hashing algorithms is known and often the hashes are stored where they can be accessed by standard users. Whilst system policies cannot help protect against hashing implementations or short-comings, the use of strong passwords can help thwart attacks against the hashes to recover the pass-phrase.
Microsoft Windows does not store or transmit passwords in clear text - it uses a hash of password for authentication. A Hash is a fixed-size result obtained by applying a mathematical function (the hashing algorithm) to an arbitrary amount of data (also known as the message digest). There are three Windows authentication algorithms: LM (least secure, most compatible); NTLM and NTLMv2 (most secure and least compatible). Although most current Windows environments have no need for LAN Manager (LM) support, Microsoft Windows locally stores legacy LM password hashes (also known as LANMAN hashes) by default on Windows NT, 2000 and XP systems (but not in Windows 2003). Since LM uses a much weaker encryption scheme than more current Microsoft approaches (NTLM and NTLMv2), LM passwords can be broken in a relatively short period of time by a determined attacker. Even passwords that otherwise would be considered "strong" can be cracked by brute-force in under a week on current hardware.

http://msdn.miscrosoft.com/library/default.asp?url=/library/en-us/security/Security/h_gly.asp

The weakness of LM hashes derives from the following:
  • Passwords are truncated to 14 characters.
  • Passwords are padded with spaces to become 14 characters.
  • Passwords are converted to all upper case characters.
  • Passwords are split into two seven character pieces.
This hashing implementation means that an attacker needs only to complete the trivial task of cracking two seven-character, upper-case passwords to gain authenticated access to your system. Since the complexity of cracking hashes increases geometrically with the length of the hash, each seven-character string is at least an order of magnitude simpler to attack by brute-force than would a combined fourteen-character string. Since all strings are exactly seven characters (including spaces) and entirely upper-case, a dictionary-style attack is also simplified. If the pass-phrase consists of 15 characters or more, the LM hash is sets a null string which effectively thwarts brute forcing against the hash. It is important to note that this insecurity is not just specific to Windows environments - whenever password hashes can be accessed without due authorisation, there is the potential for an attacker to brute-force the correct pass-phrase.

In addition to the risk posed by having legacy LM hashes stored in the SAM, the LAN Manager authentication process is often by default enabled on clients and accepted by servers. As a result, Windows machines capable of utilizing stronger hash algorithms instead send weak LM hashes across the network, making Windows authentication vulnerable to eavesdropping by packet sniffing, and therefore easing the efforts of an attacker to obtain and crack user passwords.

W5.2 Operating Systems Affected
All Microsoft Windows operating systems.

W5.3 CVE/CAN Entries
CVE-2000-0222

CAN-1999-0504, CAN-1999-0505, CAN-1999-0506

W5.4 How to Determine if you are Vulnerable
Although there are observable symptoms of general password weakness, such as the existence of active accounts for users who have departed the organization or services which are not running, the only way to know for certain that each individual password is strong is to test all of them against the same password cracking tools used by attackers.

Please Note: Never run a password scanner, even on systems for which you have administrative access, without explicit and preferably written permission from your employer. Administrators with the most benevolent of intentions have been fired for running password cracking tools without authority to do so.

A few of the best cracking tools available are: LC6 (L0phtcrack version 5) and John the Ripper

Regarding the issue of a locally stored LAN Manager hash:
  • If you are running a default installation of NT, 2000 or XP, you are vulnerable since LAN Manager hashes are stored locally, although by default only the system administrator has access privileges.
  • If you have legacy operating systems in your environment that require LM authentication in order to communicate to servers, then you are vulnerable because those machines send NTLM hashes which can be sniffed off the network.
W5.5 How to Protect Against It
The best and most appropriate defence against password weaknesses is a strong policy which includes thorough instructions to engender good password habits and proactive checking of password integrity.
  • Ensure that passwords are consistently strong. Given enough hardware and enough time, any password can be cracked by brute force. But there are simpler and very successful ways to learn passwords without such expense. Password crackers employ what are known as dictionary-style attacks. Since hashing methods are known, cracking utilities simply compare the hashed form of a password against the hashed forms of dictionary words (in many languages), proper names, and permutations of both. Therefore a password whose root in any way resembles a known word is highly susceptible to a password-cracking attack. Many organizations instruct users to generate passwords by including combinations of alphanumeric and special characters, and users more often than not adhere by taking a word ("password") and converting letters to numbers or special characters ("pa$$w0rd"). Such permutations can protect against a dictionary attack, but are likely to fall if subject to a brute-force attempt on every printable character.

    A good password therefore cannot have a word or proper name as its root. A strong password policy should direct users to generate passwords from something more random, like a phrase or the title of a book or song. By concatenating a longer string (taking the first letter of each word, or substituting a special character for a word, removing all the vowels, etc.), users can generate sufficiently long strings which combine alphanumeric and special characters in a way which dictionary attacks will have great difficulty cracking. And if the string is easy to remember, then the password should be as well.

    Once users are given the proper instructions for generating good passwords, procedures should be put in place to assure that these instructions are followed. The best way to do this is by validating the password whenever the user changes it by employing Passfilt (NT4).

    Windows 2000, XP, 2003 have powerful tools for enforcing password policy. To view your current password policy on most Windows systems, follow these steps (Start - Programs - Administrative Tools - Local Security Policy - select Account Policies - Password Policy). The Local Security Policy has following settings:
    • Password must meet complexity requirements. Determines whether passwords must meet complexity requirements. Complexity requirements are enforced when passwords are changed or created. If this policy is enabled, passwords must meet the following minimum requirements:
      • Not contain all or part of the user's account name
      • Be at least six characters in length
      • Contain characters from three of the following four categories:
        • English uppercase characters (A through Z)
        • English lowercase characters (a through z)
        • Base 10 digits (0 through 9)
        • Nonalphanumeric characters (e.g., !, $, #, %)
    • Enforce password history (range: 0-24): Determines the number of unique new passwords that have to be associated with a user account before an old password can be reused. The value must be between 0 and 24 passwords. Setting this parameter to 0 passwords remembered enables password recycling; setting it to 24 passwords remembered requires 24 changes of password before initial password can be recycled. This policy enables administrators to enhance security by ensuring that old passwords are not reused continually. To maintain the effectiveness of the password history, do not allow passwords to be changed immediately when you configure the minimum password age.
    • Maximum password age (range: 0-999 days): Determines the period of time (in days) that a password can be used before the system requires the user to change it. You can set passwords to expire after a number of days between 1 and 999, or you can specify that passwords never expire by setting the number of days to 0.
    • Minimum password age (range: 0-999 days): Determines the period of time (in days) that a password must be used before the user can change it. You can set a value between 1 and 999 days, or you can allow changes immediately by setting the number of days to 0. The minimum password age must be less than the maximum password age. Configure the minimum password age to be more than 0 if you want Enforce password history to be effective. Without a minimum password age, users can cycle through passwords repeatedly until they get to an old favorite. The default setting does not follow this recommendation, so that an administrator can specify a password for a user and then require the user to change the administrator-defined password when the user logs on. If the password history is set to 0, the user does not have to choose a new password. For this reason, password history is set to 1 by default.
    • Minimum password length (range: 0-14 characters): Determines the least number of characters that a password for a user account may contain. You can set a value of between 1 and 14 characters, or you can establish that no password is required by setting the number of characters to 0. Minimum password length should conform to corporate security policy (otherwise it is recommended that it be set to 8 or more characters; National Security Agency (NSA) recommends 12 characters).
    • Store password using reversible encryption for all users in the domain: Determines whether Windows 2000, 2003 and XP Professional store passwords using reversible encryption. This policy provides support for applications using protocols that require knowledge of the user's password for authentication purposes. Storing passwords using reversible encryption is essentially the same as storing plaintext versions of the passwords. For this reason, this policy should never be enabled unless application requirements outweigh the need to protect password information.
An approach that can be used to automatically generate and assign complex passwords to user accounts is as follows - run the following command (from Command line prompt of Windows NT4, 2000, XP, 2003):

Net user username /random

Execution of this command will assign random complex (but always 8-characters long) passwords to an account and will print this password on the console screen. This method is usually more appropriate for assigning passwords for service accounts, rather than for actual users.

The best way to audit the quality of passwords is to run password cracking utilities in a stand-alone mode as part of routine scanning.

Important Note: Never run a password scanner, even on systems for which you have administrative access, without explicit and preferably written permission from your employer. Administrators with the most benevolent of intentions have been fired for running password cracking tools without authority to do so.

Once you have acquired authority to run cracking utilities on your system, do so regularly on a protected machine. Users whose passwords are cracked should be notified confidentially and given instructions on how to choose a good password. Administrators and management should develop these procedures together so that management can provide assistance when users do not respond to these notifications.

Another way to protect against nonexistent or weak passwords is to use an alternative form of authentication such as password-generating tokens or biometrics.
  1. Protect strong passwords. Even if passwords themselves are strong, accounts can be compromised if users do not protect their passwords. Good policy should include instructions that a user never tell his or her password to anyone else, never write a password down where it could be read by others, and properly secure any files in which a password is stored to automate authentication (passwords are easier to protect when this practice is only used if absolutely necessary). Password aging should be enforced so that any passwords which slip through these rules are only vulnerable for a short window of time, and old passwords should not be reused. Make sure that the users are given warning and chances to change their password before it expires. When faced with the message "Your password has expired and must be changed," users will tend to pick a bad password.
  2. Prevent password hashes and SAM database from being copied. Password cracking tools, mentioned in this section, obtain password hashes by:
    • Sniffing passwords from the network. Countermeasures: 1. Use of switched networks; 2. Detection and removal of network cards in promiscuous mode (they can be detected by most commercial security assessment tools, such as free tools like ethereal).
    • Copying SAM file (located in %SystemRoot%\System32\Config\ folder usually C:\Winnt\System32\Config\ - on Windows NT4 and 2000 or C:\Windows\System32\Config\ - on Windows XP and 2003). This file is normally locked by Windows OS and can be copied only when system booted to alternative OS. SAM file also can be obtained by restoring backup of SAM file or System State (Windows 2000, 2003, XP). SAM file also located on NT4 Repair Disk.
    Countermeasures: Limit and monitor physical access to computer systems (especially domain controllers), backup media and Repair Disk.

    The following Microsoft articles provide useful references:
  3. Tightly control accounts.
    • Any service-based or administrative accounts not in use should be disabled or removed. Any service-based or administrative accounts which are used should be given new and strong passwords.
    • Audit the accounts on your systems and create a master list. Do not forget to check passwords on systems like routers and Internet-connected digital printers, copiers and printer controllers.
    • Develop procedures for adding authorized accounts to the list, and for removing accounts when they are no longer in use.
    • Validate the list on a regular basis to make sure no new accounts have been added and that unused accounts have been removed.
    • Have rigid procedures for removing accounts when employees or contractors leave or when the accounts are no longer required.
  4. Maintain strong password policy for the enterprise. In addition to operating system or network service-level controls, there are many comprehensive tools available to help manage good password policy. Many sample policies templates, policy development guidelines, password security fundamentals, and links to numerous security policy web sites (which include password policy information) can be found at the SANS Security Policy Project site.
  5. Disable LM authentication across the network. The best replacement in Windows for LAN Manager authentication is NT LAN Manager version 2 (NTLMv2). NTLMv2 challenge/response methods overcome many weaknesses in LM by using stronger encryption and improved authentication and session security mechanisms. The registry key that controls this capability in both Windows NT and 2000 is:

    Hive: HKEY_LOCAL_MACHINE
    Key: System\CurrentControlSet\Control\LSA
    Value: LMCompatibilityLevel
    Value Type: REG_DWORD - Number
    Valid Range: 0-5
    Default: 0

    Description: This parameter specifies the type of authentication to be used.

    0 - Send LM response and NTLM response; never use NTLMv2 session security
    1 - Use NTLMv2 session security if negotiated
    2 - Send NTLM authentication only
    3 - Send NTLMv2 authentication only
    4 - DC refuses LM authentication
    5 - DC refuses LM and NTLM authentication (accepts only NTLMv2)

    On Windows 2000, 2003, and XP the same functionality can be implemented by configuring the setting LAN Manager authentication level (Windows 2000) or Network security: LAN Manager authentication level (Windows XP, 2003) (Start - Programs - Administrative Tools - Local Security Policy - Local Policies - Security Options).

    If all of your systems are Windows NT SP4 or later, you can set this to 3 on all clients and 5 on all domain controllers to prevent any transmission of LM hashes on the network. However, legacy systems (such as Windows 95/98) will not use NTLMv2 with the default Microsoft Network Client. To get NTLMv2 capability, install the Directory Services Client. Once installed, the registry value name is "LMCompatibility," and the allowed values are 0 or 3.

    If you cannot force your legacy clients to use NTLMv2, you can gain a slight improvement over LM hashing by forcing NTLM (NT Lan Manager, version 1) at the domain controller (set LMCompatibilityLevel to 4 or if you use tool Local Security Policy set LAN Manager authentication level to value: Send NTLMv2 Response only\Refuse LM). But the most secure option with regard to legacy systems is to migrate them to newer operating platforms, since the older operating systems do not allow this minimum security level to be supported.
  6. Prevent the LM hash from being stored. One major problem with simply removing the LM hashes being passed over the network is that the hashes are still created and stored in the SAM or Active Directory. Microsoft has a mechanism available for turning off the creation of the LM hashes altogether, but only in Windows 2000, 2003 and XP. On Windows 2000 systems (SP2 or later), the following registry key controls this function:

    Hive: HKEY_LOCAL_MACHINE
    Key: System\CurrentControlSet\Control\LSA\NoLMHash

    If this key is created on a Windows 2000 Domain Controller, the LanMan hashes will no longer be created and stored in Active Directory.

    On Windows XP and 2003, the same functionality can be implemented by enabling setting Network security: Do not store LAN Manager hash value on next password change (Start - Programs - Administrative Tools - Local Security Policy - Local Policies - Security Options).

    After making these modifications to a Windows 2000 system a restart is necessary for the changes to take effect.

    Important Note: This only prevents new LM hashes from being generated. Existing LM hashes are removed individually the next time each user changes his or her password.


W6 Web Browsers

W6.1 Description
The Browser is the means by which computer users access the web on Microsoft Windows systems. The dominant web browser is Microsoft Internet Explorer (IE), which is the default web browser installed on Microsoft Windows platforms. Other Web browsers include Mozilla, Firefox, Netscape and Opera. The latest version of IE is 6, and the US-CERT has issued an advisory outlining the security enhancements and features available in IE 6. The vulnerabilities discussed here are also applicable to Mozilla versions 1.4 - 1.7.1, Firefox version 0.9.x, Netscape version 7.x, and Opera version 7.x.

The problems are six-fold:
  1. Large Number of vulnerabilities over the last few years in comparison to other browsers - 153 IE vulnerabilities since April 2001, according to the Security Focus Archive.
  2. Longer Time to patch known IE vulnerabilities - Users have had to wait in excess of six months from the time the vulnerability is disclosed before Microsoft issues a patch.
  3. Active X and Active Scripting controls themselves have not been found to be open to particular exploitation, but can be used to bypass the security constructs of the browser and potentially impact upon the host system.
  4. Large number of unpatched vulnerabilities - 34, according to http://umbrella.name/
  5. Spyware/Adware vulnerabilities - This affects all browsers and systems that facilitate access and use of web resources.
  6. Integration of IE browser into the Operating System, which makes the OS more vulnerable to exploitation.
All web browsing applications have had their share of vulnerabilities and bugs that have created security exposures. A malicious web designer can create web pages to exploit such failings while simply browsing the web pages. A prime example of this is the "Download.Ject" vulnerability, present 'in the wild' for many months, and utilized Active X vulnerabilities. Even after an exploit was published on June 8, 2004, a patch for IE was not released until July, 2004. Due to its widespread use through industry and home-user environment, Microsoft Internet Explorer is by far the most attractive target for attackers. However, the majority of risks that tend to come to light, such as cross-site-scripting attacks, affect the majority of browsers in quite the same fashion. Attacks against browsers can include disclosure of cookies, local files or data, execution of local programs, download and execution of arbitrary code, or complete takeover of the vulnerable system.

W6.2 Operating Systems Affected
These vulnerabilities exist on Microsoft Windows systems running any version of these browsers. It is important to note that IE is installed with a wide variety of Microsoft software and is therefore typically present on all Windows systems, even though the user may not wish it installed or utilized. All other browsers are installed at the user's discretion and the user decides whether the browser is used by other applications.

W6.3 Browser Vulnerabilities, courtesy of Secunia overlap as some vulns expressed on all browsers
  1. Internet Explorer:
    2004 - 15 Security Advisories (As Of July 30, 2004)
    1. Microsoft Internet Explorer Multiple Vulnerabilities
    2. Internet Explorer Frame Injection Vulnerability
    3. Internet Explorer File Download Error Message Denial of Service Weakness
    4. Internet Explorer Security Zone Bypass and Address Bar Spoofing Vulnerability
    5. Internet Explorer Local Resource Access and Cross-Zone Scripting Vulnerabilities
    6. Microsoft Internet Explorer and Outlook URL Obfuscation Issue
    7. Windows Explorer / Internet Explorer Long Share Name Buffer Overflow
    8. Microsoft Outlook Express MHTML URL Processing Vulnerability
    9. Internet Explorer/Outlook Express Restricted Zone Status Bar Spoofing
    10. Multiple Browser Cookie Path Directory Traversal Vulnerability
    11. Internet Explorer Cross Frame Scripting Restriction Bypass
    12. Internet Explorer File Identification Variant
    13. Internet Explorer Travel Log Arbitrary Script Execution Vulnerability
    14. Internet Explorer File Download Extension Spoofing
    15. Internet Explorer showHelp() Restriction Bypass Vulnerability
  2. Mozilla Vulnerabilities
    2004 - 7 Secunia Advisories
    1. Mozilla Fails to Restrict Access to "shell:"
    2. Mozilla XPInstall Dialog Box Security Issue
    3. Multiple Browsers Frame Injection Vulnerability
    4. Mozilla Browser Address Bar Spoofing Weakness
    5. Mozilla / NSS S/MIME Implementation Vulnerability
    6. Multiple Browser Cookie Path Directory Traversal Vulnerability
    7. Mozilla Cross-Site Scripting Vulnerability
  3. Netscape Vulnerabilities
    2004 - 2 Secunia Security Advisories
    1. Mozilla Fails to Restrict Access to "shell:"
    2. Multiple Browsers Frame Injection Vulnerability
  4. Opera Vulnerabilities
    2004 - 8 Secunia Security Advisories
    1. Opera Browser Address Bar Spoofing Vulnerability
    2. Multiple Browsers Frame Injection Vulnerability
    3. Opera Address Bar Spoofing Security Issue
    4. Opera Browser Favicon Displaying Address Bar Spoofing Vulnerability
    5. Multiple Browsers Telnet URI Handler File Manipulation Vulnerability
    6. Opera Browser Address Bar Spoofing Vulnerability
    7. Multiple Browser Cookie Path Directory Traversal Vulnerability
    8. Opera Browser File Download Extension Spoofing
W6.4 Identification and Protection from Browser Vulnerabilities
If you are using Internet Explorer on your system, there is no current way to know if you are vulnerable, due to the large number of unpatched vulnerabilities which exist. The Windows Update Site should be visited regularly and where possible the Automatic Updates feature enabled, ensuring that IE is protected from vulnerabilities that patches are available for. Users interested in further protection from browser vulnerabilities should consider employing a blend of the following recommendations:
  1. By far the most effective step towards a safe and secure browsing experience is to ensure the latest version of your web client is installed, which will feature new controls for heightened security and eliminate concerns identified in older versions of the application. For instance, the latest version of IE can be downloaded for free at: http://www.microsoft.com/windows/ie/
  2. Most Internet sites do not make use of ActiveX, but disabling this feature could have adverse effects on other aspects of the system. Since the