Tales from ITC: SIP Trunking
Familiar with the concept of breakeven analysis? The idea is pretty simple: Save some money on one product or service, put those savings in your pocket, and eventually use the funds to pay for Item B (something else you really want). How long you have to wait for the incoming savings to accumulate before you have enough to pay for Item Ba week, a month, a yearis your breakeven. After you've reached that point, further savings are gravy.
But what if a vendor told you that there was a zero breakeven on their product or service? Would that get your attention? A special conference session titled SIP Trunking: Realizing Rapid ROI Today, presented at the Internet Telephony Conference and Expo this week in San Diego, explored this concept. The working premise was that the deployment of Session Initiation Protocol (SIP) connections as replacements for traditional T1 and ISDN trunks could save you enough money on monthly telecommunications charges to pay for the interfacing hardware that makes it all play together.
The session was sponsored by hardware developer Ingate Systems, a Swedish firm with a long history of firewall product development. Joining Ingate for the presentation were two other groups of companies: IP-PBX vendors Mitel, Objectworld Communications, Pingtel, and Sphere Communications; and SIP trunking providers BandTel, Bandwidth.com, Level 3 Communications, and VoEX.
Here's the picture that unfolded at the session:
Unless your enterprise network is completely wireless, you need some type of hard-wired connection into the Public Switched Telephone Network (PSTN) to provide connectivity between the various locations. With traditional time division multiplexing (TDM)-based architectures, high-speed circuits such as T1 lines or ISDN Primary Rate Interface (PRI) lines, (both operating at 1.544 Mbps) are typically used to connect the switching systems (PBXs), to the service provider. Each PBX, be it based on a traditional TDM or the newer Internet Protocol (IP) architecture, will require a gateway to connect to that T1 or ISDN line, providing functions such as data protocol and transmission rate conversions. But if no gateway is requiredbecause the service provider hands you a SIP-compatible trunk and takes care of any required PSTN connections on their end of the networkthen you save the cost of the gateway hardware. Moreover, a SIP trunk provides both Internet (data) and voice connectivity, further leveraging economies of scale, since, in effect, you get two network connections with one service.
Furthermore, each traditional T1 or ISDN PRI trunk provides 24 channels, 1 for signaling and 23 for end-user voice or data traffic. This results in one of biggest challenges to network growth: When you use all of the capacity of one T1 or ISDN PRI, you have to order up another entire circuiteven if all you need is one more channel. This means you may end up paying for more network bandwidth than you actually need. But SIP trunks are provisioned separately, so that if you need 5 channels you end up paying for only 5 trunksand not the extra 18 that you would have if you had to order an additional T1 or ISDN PRI circuit.
This brings us back to the concept of zero breakeven analysis that we mentioned earlier. First, deploy a SIP-compatible IP-PBX, which can connect softphones and other SIP-based devices more easilyand likely with a lower cost that adding on terminal adapters to your existing legacy PBX. Second, find a carrier that can provide you with SIP-based trunks, so that you dont need a gateway between the IP-PBX and the traditional PSTN network. Stir both ingredients for a short time and voiláyou have an end-to-end SIP network that is both more efficient (since the multiple protocol conversions are not required), less expensive (since the gateway can be eliminated and the trunks can be provisioned more effectively), and one that takes advantage of the numerous converged network applications now available.
As with many appealing stories, it's not quite that simple in practice. There are still a few challenges to be solved, which is where Ingates SIPerator device comes inoperating as a very intelligent interface between the IP-PBX and SIP trunking provider.
- Challenge 1: SIP traffic is typically blocked at the enterprise firewall, and must be allowed to pass into the enterprise network, while still adhering to the corporate security strategies.
- Challenge 2: Network Address Translation (NAT) is frequently employed as an additional security measure, converting from the set of public IP addresses on one (public) side to another set of private IP addresses on the other (private) side.
- Challenge 3: The SIP traffic must be secured, encrypting the SIP signaling using Transport Layer Security (TLS) and encrypting the voice or video signals (media) using Secure Real Time Protocol (SRTP), so that eavesdropping on private sessions cannot occur.
- Challenge 4: Conversions between SIP and PSTN addresses (phone numbers) are required for many calls.
You probably won't be astonished to learn that the SIPerator is designed to surmount all four of these obstacles.
So check out the idea of SIP trunking for your enterprise. You might disappoint your existing service providers and hardware vendors but end up saving some operating capital that can be redirected toward other projects.
Copyright Acknowledgement: © 2006 DigiNet ® Corporation, All Rights Reserved
Mark A. Miller, P.E. is President of DigiNet ® Corporation, a Denver-based consulting engineering firm. He is the author of many books on networking technologies, including Voice over IP Technologies, and Internet Technologies Handbook, both published by John Wiley & Sons.