Elmatdor:
You are correct, and you make a profound point. Basically, the cost of producing a switch in the last quarter of the 20th century was 80/20 sf/hw; for accounting reasons, as you note, it is presented as 20/80. (I've left out installation costs: they matter also).
Since the accounting staff, or to use the technical engineering term, 'bean counters', buy the equipment, they are the real customers.
Software is really a service, not a product. and has high maintenance expenses- ~80% or higher of total lifecycle development cost is maintenance after deployment. The problem with some VOIP assumptions is they ignore the cost, complexity and scheduling of software development and maintenance. The big switches solved this with the 80/20 swapping rule. This won't work with some of the modern VOIP models. This means the VOIP hardware will 'seem' cheaper and VOIP software will seem more expensive than the big iron switches. Another nail in the road on the voice IP information highway.
Mind you it works with browers/HTML/HTTP/IP; there software that is free. But web pages are stateless, and HTML is a very very simple protocol compared to the 20 or 30 protocols with another 100 variants you need to handle call software. An 80% of a switch's software is operations, administration, or maintenance. And then there is provisioning.
This explains a major reason the failure of anything that lets an outside entity take control of your switch: you will lost the software revenue unless you charge a LOT for the interface.
1) That is why they switch vendors don't like AIN- the interface cost is just an additional set of TCAP messages on existing SS7 interfaces
2) as you noted, that is why they don't like V5.2 or TR303,
although they can charge an incremental circuit cost for this interface
Perhaps the future of VOIP in the carrier site of the last mile will be driven by accounting practices.
justone opinion. |
