Remote Control software and VPN technologies have been around for a really long time. Norton PC Anywhere was one of the first widespread remote control utilities. In fact it still has an install base today, and was in one case recently brought to my attention in a security audit; it was the only way to reach an embedded Windows 2000 HVAC control system.
Unsurprisingly, Remote Control software has also been a favorite target of hackers. In fact it was reported in February 2012 that a group of hackers, known as Yamatough but operating under the umbrella of Anonymous, released the source code for Symantec's PCAnywhere product in a failed attempt to extort a $50,000 ransom. But I digress.
System admins have utilized VPN for remote administration since Remote Access Service was first included on the Microsoft Windows Server Operating Systems, way back on Windows NT 3.51 (NT RAS). This technology largely supported “dial-up connectivity via modem”, but as Trumpet Winsock and a TCP/IP networking stack became available in later versions of Windows, admins everywhere jumped at the chance to work remotely.
Today, however, all is not well in Windows Remote Access Services utilizing Point-to-Point Tunneling Protocol (PPTP), unless specific measures are taken to safe guard it. It’s not that there is a glairing security hole here; it’s just that PPTP was developed in the late 90s. It was the very first VPN protocol that was supported by Microsoft’s servers and even though it is no longer recommended by Microsoft, it is still widely deployed.
PPTP is everywhere
Since that time, PPTP capability has been embedded in all sorts of devices – you find it in many routers, firewalls and embedded devices such as Network Attached Storage. Can it be broken? Yes; and it was successfully broken and never fixed in 1999, by Bruce Schneier and Mudge. Its vulnerability lies in the Microsoft Challenge Handshake Authentication Protocol (MS-CHAP v2): A challenge-response procedure that provides mutual authenticity – a brute force dictionary attack can figure out the password.
From a simplistic view, PPTP requires a “hole” (an open port) in your firewall. The Administrator account, in many versions of Windows Server cannot be locked out by bad password attempts, so this account is frequently used as a user ID to attack. As a result port 1723 comes under frequent attack – brute forcing “Administrator” as the user ID and many password attempts per second.
Just as PPTP is continually probed, Microsoft’s Remote Desktop Protocol (RDP) or terminal services come under frequent attack as well, and port 3389 is assaulted by many brute force password attempts just as PPTP is. Again, this remote admin feature requires a hole in the firewall. Hackers are taking a keen interest in the Remote Desktop Protocol and there are many patches and updates that address exploits that target this Windows Service. So, if you have RDP exposed to the Internet you better be up-to-date. If you’re running an older version of Windows Server, check to see if you’re set up to log failed login attempts – you may be surprised at how much activity you see in those logs.
VNC is another remote console-style application that again is targeted by hackers armed with brute force attacks against Port 5900. Finding exploits for this software is trivial and many exploits exist for the free and out-of-date versions; use this only at extreme peril.
Any port that lets traffic in is a liability
Unless you’re prepared to invest a lot of time constructing firewall Access Control Lists, in this day and age any open port on your network that lets traffic in is a vulnerability and a liability. If you have ports open to the Internet it’s frightening to think you’re not limiting traffic from certain IP address ranges to ports that should only be used by a limited number of users or administrators.
Enter the next generation of remote control utilities such as TeamViewer and Log-Me-In and any other “out bound, https-based” remote control utilities. Why are they safer than RDP, PPTP or anything else – even IPSEC – running on Port 500?
In short these remote utilities don’t need an open port on your Firewall. Instead they reach out from the computer inside your network, through the Internet over https (encrypted with SSL), to a central server that in turn connects a client requesting that machine. Think of it as a “server” that sits in-between the two connections, brokering the connections.
I know for a fact that these companies spend a lot on security and are working to make both the encryption of the connection and the “brokering” of connections more robust, stable and secure.
The harsh reality is that any remote control connection is going to be vulnerable to some sort of attack. The analogy is that anything on the Internet can be attacked, the only (now arguable) way is to air-gap the secure system from the Internet. This pretty much stops the idea of remote administration dead in it’s tracks.
Although systems like RDP and PPTP are still popular, a solution that does not open a hole in your firewall, to the ravages of the Internet is surely a better choice in the increasingly hostile threat landscape of the Internet.