Getting VoIP [This is my (somewhat stale) response to Oliver's original post]
As both a huge VoIP advocate and senior networking consultant, I’ve been meaning for some time to assemble a robust response to Oliver’s earlier questioning of the burgeoning enthusiasm for VoIP. At the risk of vying with the prophet Zarquon for tardiness, here is my attempt to shed some light on some of the salient questions Oliver has raised, and thereby dispel some of the confusion that I still see surrounding VoIP in the corner offices. And in deference to our colleagues over at the tech support board at Sundance Communications, I’m cross-posting this as a comment to Oliver’s original thread which was picked up there.
Before I go into why I think VoIP is such a great thing, I think first we need to clarify just what we mean by "VoIP" in the first place. There are several ways to interpret the term, and unfortunately I think that the idea that the term can mean different things in different contexts is not at all well understood, especially by executives (I know of a CIO who, already having various wrong-headed ideas about voice technology in general, banned the use of EVDO-capable cell phones because he thought they were using VoIP, and that VoIP was insecure). There are five basic ideas (that I know of) the term can refer to:
1. The various signalling and transmission protocols used to encode analog voice signals and then transmit them over packet data networks using the Internet-standard IP protocol for processing by different endpoints.
2. The use of asynchronous packet data networks to carry voice signals, as opposed to the old skool method of isochronous circuit-switched TDM networks. (Wikipedia is your friend if that– or indeed anything else I’m writing– doesn’t make much sense).
3. The use of a single homogenous packet switched network to carry both voice and data traffic
4. The use of the particular packet switched data network called the Internet to carry voice traffic
5. An "Enterprise 2.0" buzzword referring to enterprise voice solutions that enable sophisticated unified messaging, flat-rate long distance, free international calling, and follow-me services, regardless of the technologies used to provide those services
While I end up touching on each of these concepts throughout this article, I will attempt to identify which definition I am referring to at a given point. Clarifications now having been made, let’s move on to my response to Oliver’s implicit question of "What was the point of VoIP again? I seem to have missed it." By way of elucidation, I will claim that there are really two key advantages to voice/data convergence (and I would also contend that anything beyond these are chimeric), of which VoIP (in the sense of meaning #3) is the primary enabling technology:
1) We will call the first advantage "unified infrastructure", again per meaning #3. You only need to worry about a single information delivery infrastructure within a building. This is, IMNSHO, the biggest and best point about VoIP. No more split plant, no more having to run the voice wires separate from the data, etc. And, the same skillsets used to maintain the wiring for the computers work for the phones, too. Another big win from doing this is that you can take your VoIP phones wireless, and get rid of the cabling altogether if so desired. "We’ve had wireless phones for decades", you say? Yes, but here again, it’s not the concept so much as the implementation that is an order of magnitude improvement. Unless you’ve been toiling in the IT boiler room for a good part of your career (as I have had the dubious honor of doing), the advantages of being able to jettison half of the extant physical communications infrastructure may seem somewhat underwhelming. But believe me, this is a huge win.
Building on this point, WiFi phones can then be connected far more cheaply, easily, and reliably than their older analog or digital counterparts, and with a far richer feature set. It’s important to keep in mind that, say, a Cisco or Aastra WiFi SIP phone differs substantially from the typical home cordless, or even a cell phone. The plethora of features available on legacy PBX phones (call forwarding, voice mail, conferencing, do-not-disturb features, and especially multiline appearances) are present on a WiFi phone. This is not the case with a mere cordless phone, most of which simply take one or two "analog lines" and make those available on the handset with the PBX features unavailable or hidden in cryptic button sequences that not even the PBX engineers can remember. Sure, the big name legacy PBX vendors like Avaya and NorTel make "cordless PBX phones", but these things cost the earth, have limited range, and their roaming capabilities around, say, an office park campus are a joke. None of these limitations apply to WiFi phones (assuming a properly designed wireless data infrastructure– a topic for another post, to be sure).
2) We will call the second advantage "unified services". There are several aspects to this, but VoIP magic isn’t about applying Nyquist algorithms to acoustic signals; as Oliver notes, that’s old hat. VoIP is about using open encoding and transmission formats that can finally be "gotten at" by any developer who wants to take the time to dig into them. The idea of third parties having access to the digitized voice streams in legacy PBX systems was all but unheard of. Not that the VoIP PBX’s themselves aren’t a healthy swath of proprietary moat-and-castle architecture (although see further observations below), but the idea behind using open standards for encoding and signalling is that anyone who bothers to can make hardware or software that can interact with the PBX and add value. There’s been some low hanging fruit already tackled because of this:
You can finally get your voice mail in our Outlook in-box as wave files. Called "unified messaging", the fact that the voicemail is recorded in a portable file format, and that the file is then exported from the PBX, is another advantage of "VoIP" (meaning #4 this time). Could this have been done with a legacy PBX? Yes, but, going proprietary on the systems and resisting open standards for CTI (Computer Telephony Integration) made the vendors more money– or so they thought.
As noted by Oliver, the same funky Dell servers that run the corporate apps now run the PBX. Seems like a trivial point, but actually its a huge advantage because for the first time the process for making and managing backups of the PBX and voicemail configurations and databases is straightforward; indeed, it’s standard industry best practices for data backups. Previously, most PBX’s relied either on local backups in proprietary formats, or, for the rara avis that did have a tape or floppy drive, the backups either required excessive babysitting and/or (all too typically), didn’t really back up all the data in the system. Anyone who has had the painful experience of dealing with an NEC PBX knows all too well what I’m on about.
I said above that the VoIP magic is about open call signalling and transmission formats. It can be argued that the old "Bellcore standards" implemented in the legacy TDM networks were open; there’s a thriving ecosystem of data and telecommunications end points out there because of it. But this is an extremely telling point– because that wasn’t always the case. Before the AT&T monopoly breakup in the early 80’s, that rich ecosystem of "third party" telecommunications products– and associated services– didn’t exist. The original deregulation ushered in an unprecedented boom in new telecom and datacom technologies.
It’s my contention– and that of other VoIP advocates– that the open, LAN based signalling standards of the VoIP revolution are going to engender the same type of transformative dynamics in IT. I mentioned above that currently the corporate PBX is still surrounded by a moat of proprietary protection; this is as true for Cisco and Avaya "VoIP" PBX’s as it was for their legacy NorTel and AT&T predecessors. But there are two aspects in which this landscape is changing before our eyes:
First, as noted above, while the switch software itself may be proprietary, because of the open signalling standards, it is possible to connect non-proprietary endpoints to the proprietary PBX’s, such as fax adapters, desk phones, and the noted WiFi phones. Second, and of far greater import, the proprietary PBX is itself under siege, chiefly from an open source project called Asterisk. Asterisk is nothing more nor less than an open source PBX, a freely available counterpart to Cisco and Avaya’s monolitic call processing software. Admittedly, Asterisk doesn’t– yet– have the sophistication of its commercial forebears, but this is changing fast. And not only is the number of devices it supports vast (any VoIP phone supporting the SIP protocol can be used, even Cisco phones), but (just like with the AT&T breakup) a thriving ecosystem is forming around it, offering an ever greater array of new products and services. Anyone at all familiar with the rise of open source software should realize just how large the import of such a project and its effects are. And for the SMB market, Asterisk offers the same tantalizing promise for telephony as other open source projects do for other areas of IT: all the key features and functionality needed from the high-end, financially burdensome commercial products, but at a mere fraction of the cost.
It is critical to understand that both the unification of the infrastructure and the development of open telephony infrastructure are both made possible solely because of the open, nonproprietary nature of the many signalling and voice transport standards residing under the "Voice-over-IP" umbrella (meanings 1, 2, and 3). Just like the telecommunications revolution that followed the AT&T breakup (and launched my career in data networking), so I (and many other pundits) expect that a similiar revolution is already in progress following the opening up of the corporate voice infrastructure to developers far and wide.
More discussion welcome,
–Paul
Paul !