Ronald Guria asks the question how reliable should VoIP be? There is only one answer – completely and 100%. If any technology wants to disrupt and replace a system that has worked well for over 100 years, it has to be near darned perfect. The million odd early adopters don’t make a market – they are nothing but an extended beta test. I have argued about the ease of use and reliablity before. VoIP continues to fail the mom test. For the longest time I thought, perhaps I was alone in expecting ease of use and reliability. Glad to see that Martin Geddes is on the same page. “Vendors should think of the adoption of their product as being a courtship. Never demand too much of the other party early on. Don’t blow your nose in a mucky handkerchief while chatting a girl up,” he quips in classic British understated manner, and adds, “Can you imagine any other business where one product accounts for 80%+ of your revenue, but you never attempt to promote consumption at the point of sale?”
But tens of millions of mobile users tolerate far less than 5 9s reliability (on a relative basis), in exchange for the added convenience, and perhaps a lower price. VoIP offers price & convenience advantages, and so I’d agree with Ronald that it probably needs to be somewhere between the 5 9s offered by PTSN and 97-98% offered by mobile carriers.
Ahem… Om – my surname is GRUIA, not GURIA (which, by the way, in Portuguese means gal ;-). Looks like you propagated the typo that our mutual buddy from France (Marc Goldberg) has made.
Seriously, I will post a full response on my blog about this. Just because I asked the question and quoted Henning’s research, it does not mean I necessarily endorse less than five nines. Remember that I come from the voice world, so I do not treat it like just any other data application.