Voice over IP – VoIP – is inevitable in government. It is the hottest area, and many government departments are exploring how they can take advantage of its benefits – like greater flexibility in service and features and possibly lower costs. This rating is reserved for the most critical problems.
On the dark side, however, is a concern with security. As usual, new security vulnerabilities go hand in hand with new IT services. A lot of work often needs to be done just to achieve the same baseline of security that was available with the previous generation of equipment – if in fact the old level is even achievable any more.
The PAL initiative
In the federal government, Public Safety and Emergency Preparedness Canada (PSEPC), Communications Security Establishment (CSE), Industry Canada and other government departments all have roles to play in VoIP. Industry Canada, for example, established the Protocol Analysis Lab (PAL) within Michael Binder’s organization in 2002 to assist in providing engineering support for telecommunications policy development.
Among other activities, PAL analyses newly discovered vulnerabilities in IP and telecommunications systems and helps formulate a government response in collaboration with industry and other government partners. The work at this lab has recently taken them to the cutting edge in VoIP security, especially in crucial areas such as understanding how to shield vulnerable systems from hacker attacks.
The origins of the PAL team go back to an infamous security issue in communications systems, when the University of Oulu in Finland published the results of a research project in protocol vulnerability testing. The work uncovered problems in the commonly used network management protocol based on ASN.1. The result rendered many IT and telecom systems vulnerable. A later generation of this tool, known as PROTOS, was released in 2004; it was designed explicitly to find vulnerabilities in SIP implementations, a commonly used protocol in VoIP systems.
This was the kick-start the hacker community needed, and a day after the PROTOS release commercial SIP telephones were knocked out. This led to a new focus on VoIP for the PAL team. A good understanding of what this entails can be seen by looking at a day in the life of the lab when a security vulnerability hits the Internet. However, to put this in context, a brief tour of the major VoIP security concerns is required.
VoIP = new technology and new security concerns
The technology newness and flexibility of VoIP is at the heart of the security issue. The technology incorporates new IP-based protocols that are responsible for call-processing signalling, and these run on the same network as the voice communication payload. These open protocols provide for much greater flexibility and application development, which is the key for many higher valued returns promised for the future. However, the openness and greater flexibility represents opportunities for attackers as well.
On top of this, these new call-processing protocols such as Session Initiation Protocol (SIP) were designed with minimal security built in. The standards committee developing the specification at the time believed that security would be “added on later” by running SIP over a secure link. This violates the fundamental security tenant of “build it in” – as opposed to trying to patch later; telecom equipment that faithfully implements the SIP specification could be in for a rough ride from attackers, and special attention must be paid to detect and stop attacks that can compromise service.
So far, VoIP has not experienced many big security issues, but don’t take this as a forecast of the future. VoIP is only starting to reach critical mass as a “target rich environment.” As soon as this happens, the hacking community will turn to it with an eye to making money. The recent VoIP related fraud and arrest last June of Edwin Pena, 23, the owner of two small Miami VoIP companies, is a sign of the times ahead. In this case Pena and his 22-year-old hacker collaborator Robert Moore hacked into third party VoIP services and routed their customer’s calls over those networks. Pena collected the fees from customers, to the tune of $1 million, but did not pay any service provider. The publicity on this case is sure to inspire similar attacks.
There are many threats to VoIP, but attacks that exploit the open nature of the protocol have high potential for damage. There is no difference here when compared to life on the Internet for any other application, such as e-mail. As soon as basic application operation is locked down, the most damaging attacks are those that exploit vulnerabilities in software. However, here is where greater danger lurks for VoIP users. The traditional defence for generic software vulnerabilities is to keep up with patches. The problem is that applications such as VoIP will have vulnerable software embedded into many hardware devices, such as PBXs and telephone handsets, which are more difficult to update. This gives rise to a newer defence strategy of shield first, patch second. It is the shielding aspect, among others, of VoIP vulnerabilities that Industry Canada’s PAL team is exploring.
Top VoIP security Issues – what is a real issue, what is just noise?
It is becoming common to see technical presentations or articles on VoIP security providing a long list of security threats, along with matching security tips. The top three are denial of service, fraud and abuse, and privacy and confidentiality. It is not unusual to see conference demonstrations of interceptions and replay of voice calls. However, many of these threats such as interception are much more theoretical than practical. Think of the attackers; they are motivated by money, and attacks such as the Edwin Pena fraud scheme come quickly to mind. Certainly unauthorized interception is a problem, but it is not the first thing to worry about. However, if you encrypt every single e-mail in your environment, you would be concerned with encrypting VoIP payloads. But for most environments, you want to keep a reliable service operating and not let it become a platform for hackers to launch other attacks. The reason hackers can launch malware is because software has vulnerabilities. VoIP software is no different, and vulnerabilities will be the number one security issue.
In general, VoIP security means dealing with the same security issues as you would around any other application. If you follow best practices, and do well as part of routine IT security audits, you are well positioned to deploy VoIP with confidence. Particular attention must be paid to the continuous process of staying secure by addressing software vulnerabilities and the appropriate shielding/patching/response mechanisms.
As with any other application exposed to the internet, the root cause of malware is a known vulnerability. Hackers then launch viruses or worms that exploit the vulnerability. Typical mitigation practice has been to update virus scanner signatures and patch vulnerable software. These techniques are simply not enough any more, as zero-day exploits and targeted attacks are becoming common. A zero-day exploit means you are hit before you knew there was a problem. In this case hackers keep information about a vulnerability to themselves until they are ready to launch their attack. As these attacks are motivated by financial gain, they are often designed to take over machines to harvest them as part of a bot-net. When someone controls tens of thousands of machines they are used to launch denial-of-service attacks on, for example, an Internet gambling site in order to extort “protection” money. So rather than create a nuisance by writing a virus, hackers are now individually targeting vulnerable systems to take them over for use as a platform for other attacks. In a VoIP environment, it is easy to see the appeal in compromising a server that will result in infecting all connected soft phones, for example.
Read about the state of security and emergency services in Canada