Five million machines potentially vulnerable to RDP exploit

Up to 5 million computers have Remote Desktop Protocol (RDP) activated meaning many were potentially exploitable via the MS12-020 bug.

The results derived from an ongoing TCP port scan to locate machines runing RDP services. It was preformed by security researcher Dan Kaminsky and was about 20 per cent complete.

It found 415,000 of 300 million machines (0.4 per cent or extrapolated to about 8.3 per cent of total internet users) had RDP activated.

The RDP vulnerability provided attackers with remote access to networks that have RDP enabled and was predicted to cause mayhem for organisations this year with small businesses most notably at risk.

Kaminsky told SC Magazine Australia he ran the scan to show the popularity of RDP and debunk sentiments of some IT security professionals that only “stupid” users would be vulnerable.

“The results have been consisent (in regular scan statistic checks) and it is fair to extrapolate that about 5 million IP addresses have RDP activated,” he said.

“It came as a result of victim-shaming from people who said 'who could be dumb enough to be vulnerable [to the RDP bug]',” he said. “It was a common question and asked out of ignorance.”

He said dispassionate and quantitative data was essential to defend against the “cyber security crisis”.

It was unknown how many systems had applied the Microsoft patch (CVE 2012-0002) released earlier this month.

Kaminsky suspected that a “large number” of systems were patched, but added that he was “not delusional because large numbers haven't been”.

Infosec professionals may be reticent to apply the patch if they could downtime as a result. Kaminsky said that was a poor excuse because networks must be built resilient to downtime.

It was the third dataset on prevalence of bugs that Kaminsky has run in recent months

The scan results were validated because Kaminsky probed systems a second time if they initially responded to ensure RDP was activated.

Vulnerabilities in RDP have been previously reported several times. In August 2009, a patch rated 'critical' was issued for RDP vulnerabilities that could allow remote code execution, provided an attacker could phish a user of Terminal Services.

Remote Desktop Protocol was “extrodinary complicated”, Kaminsky said,  as was exploiting the MS12-020  flaw, yet he noted researchers were more experienced in preauthorisation attacks.

The respected researcher said infosec professionals need to be mindful and prepared for an RDP exploit, despite that it was not the only threat on the critical attack surface.

He listed threats on the critical attack surfaces without order as:

• HTTP (Apache, Apache2, IIS, nginx)

• Web Languages (ASPX, ASP, PHP, and  parts of Perl/Python and Rails.)

• TCP/IP (Windows 2000/2003/2008 Server, Linux, FreeBSD, and iOS)

• XML (libxml, MSXML3/4/6)

• SSL (OpenSSL, CryptoAPI,  NSS)

• SSH (OpenSSH, Dropbear)

• DNS (BIND9, BIND8, MSDNS, Unbound, NSD)

• SNMP

• SMTP (Sendmail, Postfix, Exchange, Qmail, and the services of GMail/Hotmail/Yahoo)

Kaminsky was developing four mitigation strategies for the RDP bug that could be used for unpatched systems, but he did not wish to disclose them at the time of writing because he said they were not yet ready.

H recommended users may, among other strategies, consider replacing RAW RDP with RDP over virtual private network.

“But the best strategy is to apply the patch."

Copyright © SC Magazine, Australia