Network Vulnerability Scan with OpenVAS

What is a Network Vulnerability Scanner?

The network perimeter of a company is the 'wall' which isolates the internal network from the outside world. However, because the outside world needs to access various resources of the company (ex. the website), the network perimeter exposes some network services (ex. FTP, VPN, DNS, HTTP, and others).

A Network Vulnerability Scanner is designed to map all the services exposed on the network perimeter and detect if they are affected by vulnerabilities.

Details about our scanner

The Light version of our Network Vulnerability Scanner performs a very fast security assessment with minimum interaction with the target system. It starts by first running Nmap to detect the open ports and services. Then, based on the results returned by Nmap, our network scanner interrogates a database with known vulnerabilities to see if the specific versions of the services are affected by any issues.
This detection method, while being very fast, it is prone to returning false positives because it relies only on the version reported by the services (which may be inaccurate).

The Full version of the Network Vulnerability Scanner uses OpenVAS as a scanning engine. OpenVAS is the most advanced open-source vulnerability scanner, which can actively detect thousands of vulnerabilities in network services such as SMTP, DNS, VPN, SSH, RDP, VNC, HTTP, and many more. OpenVAS does vulnerability detection by connecting to each network service and sending crafted packets to make them respond in certain ways. Depending on the response, the scanner reports the service as vulnerable or not.

We have pre-configured and fine-tuned OpenVAS on our servers and have also added a very simple interface on top of its complex functionalities. The engine is running in a distributed environment and it can perform multiple parallel scans.

How does the OpenVAS scanner work?

OpenVAS is a fork of the old Nessus scanner, performed in 2005 when Nessus became a commercial product. OpenVAS is currently developed and maintained by Greenbone Networks with support from the community.

OpenVAS implements each test in a plugin called NVT (Network Vulnerability Test) which is written in a scripting language called NASL (Nessus Attack Scripting Language). It currently has more than 57000 active plugins which can detect a huge number of vulnerabilities for numerous services and applications.

For instance, here is how one simple NVT looks like. It is called fortigate_detect.nasl and it tells if the target device is a Fortigate Firewall:

#
#  This script was written by David Maciejak 
#  This script is released under the GNU GPL v2
#

if(description)
{
 script_id(17367);
 script_name("Fortinet Fortigate console management detection");
 script_family("General");
 script_dependencies("http_version.nasl");
 script_require_ports(443);
 exit(0);
}

#
# The script code starts here
#
include("http_func.inc");

function https_get(port, request)
{
    if(get_port_state(port))
    {

         soc = open_sock_tcp(port, transport:ENCAPS_SSLv23);
         if(soc)
         {
            send(socket:soc, data:string(request,"\r\n"));
            result = http_recv(socket:soc);
            close(soc);
            return(result);
         }
    }
}

port = 443;

if(get_port_state(port))
{
  req1 = http_get(item:"/system/console?version=1.5", port:port);
  req = https_get(request:req1, port:port);
  #<title>Fortigate Console Access</title>

  if("Fortigate Console Access" >< req)
  {
    security_note(port);
  }
}
            

OpenVAS Scanning Policy

While OpenVAS has multiple predefined policies, our scanner uses the one called Full and Fast. This policy uses the majority of the NVTs and it is optimized to use the information collected by the previous plugins. For instance, if a previous plugin detects the FTP service running on port 2121, it will run all the FTP-related plugins on that port. Otherwise, it won't.

Open Ports Detection

We have configured OpenVAS to scan for a default list of ports containing the most common 6000 ports (TCP and UDP). However, please note that the scanner first attempts to detect if the host is alive or not before doing the port scan. If the host is not alive (ex. does not respond to ICMP requests) it will show zero open ports found.

Note: If the scanner does not find any open ports even though you know there are, we recommend you re-running the scan with the option "Check if host is alive" disabled. This will skip host discovery and just start the port scan.

How long does an OpenVAS scan take?

Since the OpenVAS scanner performs a considerable number of tests, the full scan can take from 30 minutes to several hours. It highly depends on the number of open ports found on the target host. As this number is larger, the scanning time increases because OpenVAS will have to run a higher number of NVTs.