This section is the first of a three-part series on the topic of "VoIP islands." This series will focus on what the term means, how VoIP islands came into being and their implications for enterprise networks. Moving beyond this focus, the second segment will address the topic of how these islands can be connected and what this means for enterprises. Finally, the third article will examine the business case for bridging these islands, along with ROI scenarios that enterprises are achieving today.
To answer the question "how did we get here?" we must start with the initial transition from legacy telephony to IP telephony that began early in this decade. The concept of VoIP islands arose from the adoption of IP PBXs -- in either hybrid or all-IP form -- over TDM-based PBXs. Within the enterprise environment, typically the LAN, IP PBXs provided comparable functionality to legacy systems but with a lower price tag. This proposition was particularly attractive to midsized businesses or branch offices, which could not previously justify the capital expense of a legacy PBX but wanted more features than their existing key systems could provide.
This may have been a good idea within the enterprise, but all IP PBX deployments were essentially islands, in that there was no practical way for them to connect directly to one another. The reason for this is quite simple -- at the time, IP PBXs were voice-centric systems, and vendors were concerned only with connecting them to the PSTN and not to other IP PBXs. Telecom vendors were still using proprietary technologies and deployed various protocols to enable this connectivity -- primarily H.323, but also MGCP and H.248.
This was sufficient for enterprises because the vast majority of telecom networks were PSTN, so voice traffic needed to be converted to/from VoIP and TDM via a media gateway at the network edge. As a result, voice traffic may have been packetized inside the enterprise, but outside it was TDM, and the by-product was VoIP islands. It should be made clear that this outcome was for the convenience of the telecom vendor and not by the enterprise's design.
At the time, VoIP was not mature enough to displace TDM, so the path of least resistance was to use VoIP only inside the LAN and use the tried-and-true PSTN for as much of the transport as possible. Unfortunately, this kept the cost of telecom service -- especially trunking -- fairly high, but this was not the vendors' problem, and as long as the PBX was a voice-only system, they could still provide an attractive entree for enterprises to adopt IP telephony. In this scenario, enterprises could get PBX-caliber functionality at lower cost and maintain the reliability of PSTN service. The trade-off was having these islands of VoIP.
All of this changed with the advent of SIP (Session Initiation Protocol). While SIP has become the de facto standard for real-time IP communications, it was slow to gain acceptance and was initially deployed mainly for telephony. SIP is fundamentally different from the protocols cited earlier, as they were developed to facilitate the connection of IP telephony to the PSTN. After all, the PSTN will be with us for years to come, so the need for this is very real. In most cases, this conversion takes place in the enterprise's media gateway using these protocols, and everyone is happy. The enterprise gets more features for less money, the vendor sells more IP PBXs and media gateways, and the service provider maintains the status quo with high-margin POTS.
This model works well for everyday telephony, but with SIP, so much more is possible, and that is where VoIP islands become problematic. VoIP islands may advance the IP revolution by a few steps, but they offer only a Telecom 1.0-type solution, in that all enterprises are getting is a replication of PBX service at a lower cost. The real value that comes from SIP is to enable a broader vision of IP communications within which telephony is just one mode. IP PBXs can touch on this in a limited way, but with SIP, enterprise networks can deliver a much richer environment to end users, allowing us to seamlessly mix voice, data, chat, IM, video and so on.
Convergence is the term most often used here, whereby voice and data are combined onto a single IP network. This lays the foundation for a Telecom 2.0 evolution, where the role of protocols is not to connect IP PBXs to the PSTN, but rather to other IP networks. Clearly, this is not in the interest of incumbent telecoms, nor is it ideal for telecom vendors. On the other hand, this is good news for telecoms with IP-based networks, vendors building products around SIP and, of course, enterprises that see how SIP can help them leverage far more ROI from their IT budgets.
In the next part of this series, I'll explore these ideas further and explain why linking these islands will be better, not worse, for everyone here.