We are sure you take it as a matter of course that the whole world is going to IP and that VoIP is "the future." Of course it is! Whether you're talking about residential users migrating to Vonage and the like, large enterprises moving call centers and PBXs to IP, or carriers moving to VoIP in their core networks, the move is in progress and inevitable.Packet8’s Virtual Office solution - a hosted IP PBX service with a companion VoIP business phone that provides unlimited U.S. and Canada calling with auto attendant, conference bridging and voicemail features all included in that base price. Optional features include things like a Virtual Office Switchboard application for receptionists, local main and DID numbers, and extended auto attendant features like group routing.
Inevitable it may be, but we asked ourselves this: is VoIP ready for prime time for business use? Would we recommend that service providers offer VoIP today to their customers, and if so, what would we tell them to improve over existing offerings? Or to look at it from a different perspective, would we, as a small but highly distributed and virtualized business buy such a service? What would it offer us over more traditional alternatives?
This isn't the first time we've experimented along these lines – we've been trying out various voice services for years and years. And we've lamented in these very pages our inability to find a service that reaches all of our locations without requiring costly T-1 lines into home offices. The truly "business class" VoIP services out there just don't reach us without requiring us to get tied into data services that aren't a good fit for our needs. For example, one of our home office telecommuters is in all other respects better off with a much cheaper and considerably faster (downstream at least) cable modem connection than a T-1.
So we tried a totally different tack by skipping the pricier integrated voice and data services and going with an inexpensive "bring your own access" service. What we were looking for was a business-oriented service that could give us features like auto attendant with by-name or by-extension dialing, internal extension dialing, caller groups and call routing… basically all the features that one would expect from a PBX or CENTREX type service. The bottom line was that we were looking for a service that could replicate what we might get from the RBOC or a CLEC's small business services - plus that little "extra" in terms of management and configurability that VoIP has always promised.
What we most recently tried was
We experimented with the service for several months over several of our offices, with broadband connections ranging from "business class" SDSL to residential cable modem and ADSL connections.
Our results? Well, the results were good and bad. The feature set offered was, compared to a comparably-priced POTS service, considerably higher. On a strictly dollar basis, today's telco small business offerings can't touch the price/feature performance of the Packet8 offering. On a feature richness basis, we got almost all that is reasonably to be expected of a small business offering. Our only annoyance was the lock-in we'd incur in buying Packet8-specific gear - we'd rather invest in equipment we can port from provider to provider.
Where our experiences fell a bit behind our expectations was on the performance side of things. We had a few bugs with ATAs and phones acting up and needing occasional reboots – but that's certainly not the end of the world, and is the kind of thing we expect can be updated over time with firmware updates and hardware revisions.
What did worry us most was related to the bandwidth and QoS side of things. Calls transferred through the auto attendant feature (as opposed to directly dialed) were of low quality. Packet8 addressed this by increasing the bandwidth allocations of such calls dramatically. And that's where our problems really hit home - we ended up requiring as much as 400K bit/sec of symmetric bandwidth to handle conference calls through an external bridge. And on upstream-limited SOHO broadband, that was essentially taking up all the available bandwidth we had.
The lack of a QoS mechanism for prioritizing our voice calls was really the Achilles' heel for this type of solution. Because we couldn't prioritize voice, and because we had some calls running back and forth through centralized servers for PBX functions, we had to "throw" bandwidth at the problem. This approach "worked" but it wasn't a satisfactory solution for our needs.
If we did not have the upstream bandwidth issues, the Packet8 solution would be a great solution for us (and an end of our long quest for a suitable multi-location business class IP phone service). Other small businesses may however have such bandwidth, and be willing to trade some bandwidth for the features and cost savings offered by the Virtual Office solution. We loved the features, but couldn't quite make the business case for ourselves without the missing QoS element. It's pretty bad when you have to tell a client you'll call back on a non-IP phone connection…it sounds 'cheap'.
If you're a carrier reading this and feeling pretty secure in your circuit-switched offerings, don't. As we've said, some small businesses will probably gladly trade the occasional glitches and bandwidth hit for the cost savings and feature increases offered here. And systems that can help prioritize voice packets are on the streets and cheap - $60 SOHO routers from folks like ZyXEL which don't offer end-to-end QoS, but do help alleviate bottlenecks at the local router, for example. A business-oriented ISP could easily incorporate this kind of service into their DSL or T-1 network on the cheap, and provide an alternative to today's RBOC and CLEC voice offerings without the shortcomings we've just mentioned. QoS will be solved; it's just not there yet.
If you think you have a solution that can be crowned the Dream Small Business IP Phone Service, let us know - we're still looking.