SIP Trunking IS More Expensive. Until It’s Not.

They say you learn everything you need to know in kindergarten.  I have a confession to make: I never attended. As a result, I never learned the secret handshake or whatever else it is you kindergarten grads pick-up at those hallowed halls.  Word has it that amongst these valuable life-skills, kindergarteners learn that sharing is good.  While this behavior is noble, altruistic and all that jive, it is also economical and in our own self-interest.  Sharing your basketball at the playground is a sure way to get picked for one of the teams (not to mention it beats shooting hoops all by yourself).  Sharing a cab from the airport to the hotel can greatly diminish the impact on your wallet (and make the conversation on the way there more pleasant).

Sharing even plays an important role in telecommunications.  Enter SIP Trunking.  All the hype (particularly from hardware manufacturers pitching their wares) says it’s supposed to save you a bundle. But do the math and something’s amiss.

  • SIP uses more bandwidth.  Depending on the network provider’s setting, an uncompressed SIP call (using G.711 as the codec) uses 96 Kbps –assuming 64 Kbps of payload and roughly 32 Kbps of packet overhead. This means a DS-1, which in the stodgy TDM world could handle 24 voice channels (or 23 voice plus one signaling), is now limited to 16 voice channels (2/3 the capacity).
  • SIP has higher port charges.  Furthermore, carrier SIP trunks (on a per-simultaneous-call basis) can be twice as expensive as TDM.

Paying more to get less does not sound like a recipe for savings. So, what’s all the hype about?  True, you can significantly decrease the bandwidth by using a compressed codec, but even if you sacrifice on quality to triple the SIP calls you can put on that pipe, with the premium you’re paying on the ports, you’re barely breaking-even.

The true secret to savings is sharing.

  • Your SIP trunks share the same access facilities as your data network, which means you can eliminate duplicative infrastructure. And voice is not a bandwidth-intensive application, so while you should make sure your data network is tuned to play well with voice, in many cases, this will not translate to the need for bigger pipes.
  • Using a shared session model (which, by the way, not all carriers offer), multiple sites can share the same CCPs, which means not having to buy blocks of 23 calls which sit idle most of the day (some of this capacity may never get used). It also means when one location is idle (say, because it’s lunchtime in the east coast) another location can pick-up that calling capacity slack (say, the west-coast folks starting-off their days on the right foot).

Sharing is good.  Check!  There may be hope yet for a kindergarten reprobate like myself.  Now if I could only get someone to teach me the secret handshake, I’d be set.

The following two tabs change content below.
The Level 3 Editor manages content for Level 3's Beyond Bandwidth blog.

3 thoughts on “SIP Trunking IS More Expensive. Until It’s Not.

  1. Pingback: The Top 5 Blog Posts of 2012 - Beyond Bandwidth

  2. What you should also mention is that SIP usually goes hand in hand with centralisation, VoIP and data redundancy plans.
    i.e. an office used to have a data circuit and a voice circuit with probably no redundancy, and if redundancy was required then 4 circuits were needed (2 data, 2 voice). Now those remote offices only need 2 data circuits for full redundancy, data and voice – and don’t forget the reduction in hardware required for fewer circuits and less expensive interface cards in the routers.
    Plus as you’ve centralised your calls to the central location the cost of the central locations data services is much less as the larger the data connection the lower the cost per Mbit.

    • Thank you Kean! There’s obviously a disaster recovery angle which I do not explore in this post – and it is true that you can have a more resilient voice service with a cheaper TCO using SIP. While I agree with your observation on premise hardware costs becoming smaller, I’m not sure I would argue the additional voice payload would get customers up to cheaper cost per Megabit (because voice constitutes a relatively small load on the overall network).

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload the CAPTCHA.