Traditional voice services that are delivered using PSTN provide excellent voice quality, very high reliability, carry critical services such as E911, enable federal agencies with ability for lawful intercept, all while offering an extremely high level of security. For VoIP to become a the mainstream communications medium, users of VoIP will demand the exact if not better metrics that PSTN exhibits. Thus, those providing VoIP must be able to ensure that voice networks are able to deliver the same quality, reliability, and security to that of PSTN. With various providers at different strata's publicly committing to VoIP, security is one of the biggest barriers to the successful deployment of VoIP. To securely implement VoIP networks, understanding that data and voice network security must be addresses differently is critical.

VoIP is a complex service. Over the years, service providers and PBX vendors have established their respective brands with traditional telephony technologies as being synonymous with high levels of reliability, quality and security and must preserve these parameters with VoIP, meanwhile dealing with the unique requirements of VoIP; it's high sensitivity to Quality of Service parameters, its real-time nature, a heterogeneous infrastructure, protocols and applications, and interaction with legacy PSTN networks. Special techniques and methodologies must be in place to address each facet of the aforementioned requirements. For example, on the data side, Denial of Service (DoS) attacks often result a system being unusable for extended periods of time. On the VoIP side, to come anywhere near the 99.999% of reliability of PSTN networks, a DoS attack would be disastrous. In reality, 99.999% translates into zero DoS attacks all the time. Is that possible? Can a system be built to ensure this against current and future unknown attacks? Yes; these systems are inevitable. Furthermore, VoIP's sensitivity to packet delay, packet loss, and packet jitter does not align with current data security technologies implementing encryption on a per-hop-basis. There is simply no effective way to secure voice traffic via encryption like an email without interference, interference that is more costly in the voice plane than that of data.

Long story short, VoIP security requires more than a single white knight, more than a single technology, it requires new approaches to established paradigms. In the VoIP plane, maximum packet delay is considered to be 150 ms after which "voice" service is considered affected. However, the multi-layer nature of security infrastructure could add significant delays and jitter that would make the VoIP services unusable. There is also the issue of balance between encryption and QoS. Existing encryption engines will introduce additional jitter and delay that would be cumulative due to hop by-hop encryption schemas foreseen to be used by VoIP calls. As PSTN and VoIP networks coexist media gateways that provide internetworking between carrier’s IP network and TDM based PSTN networks will be required. This could enable crossnetwork security attacks which impact existing PSTN networks. VoIP is a real-time service. All communications are happening in real-time and no information is stored anywhere on the network. As result, any loss of information cannot be recovered or retransmitted. This makes VoIP services very susceptible to worms and DoS attacks that could very easily disrupt voice communication. Finally, the complex nature of VoIP infrastructure demands a different approach to security. A VoIP network consists of a wide range of components and applications such as telephone handsets, conferencing units, mobile units, call processors/call managers, gateways, routers, firewall's and specialized protocols. As a result, a system level approach where security is built into all the infrastructure layers and coordinated via a centralized control center is required.

Enterprise Deconstructed

This month we'll introduce you to what we hope will be a reoccurring meeting point for reading some really compelling and interesting literature related to VoIP security. The literature will try to be kept at a level we like to call "technical for the masses", we don't wan tot drown we with technical jargon but rather get the point across. So let let's take a look at a typical 2 branch enterprise:

Notice that the majority of contexts above are labeled as having some inherent security risk. If this worries you, do not be afraid, with the proper planning and implementations, most of not all of these risks can be mitigated. However, the first step is knowledge and understanding of the details of each risk. Let's take look. Check out our BleedingVoIP Security Series section.

---