May 2001

Telephony Boards:
Just How Open is "Open"?

BY Tom Keating

By combining the voice, data, and other functionality from multiple systems into a single platform environment, converged systems eliminate integration and interoperability concerns. Since common interfaces can be used to manage voice, data, and other communication types, system installations as well as maintenance are simplified.

Similarly, there has been a trend towards open systems, open standards, and open programming platforms. Examples include Linux, Java, H.323, Voice XML, SIP, TAPI, and S.100, just to name a few. Indeed, convergence and open standards are the mantra of Communications Solutions, as well as our sister publications. It's what we write about, it's what we test in TMC Labs, and it's what we showcase at our tradeshows.

When I learned about Dialogic's announcement, called the Intel Converged Communications Platform, it certainly caught my attention. This standards-based, application-ready platform enables value added resellers (VARs) and system integrators to consolidate many telephony and business communications applications, such as unified messaging and speech dialer solutions from multiple vendors, onto a single system. This, in turn, allows small to medium-sized companies to rapidly and cost-effectively expand their business communications capabilities at any time. You don't need an ACD right now? That's fine, add it later. And when you do, in theory, you can choose the "best of breed" ACD or any other application written to the Intel CCP. Could this spell the death knell of the traditional PBX?

Of course, I should point out that currently there are only 17 ISV partners who are developing applications that will be ready for deployment on the platform when it ships later this year, yet there are hundreds of products designed to work with traditional PBXs. However, if enough products are written to the CCP, then you can have "best of breed" communications products running all on the same hardware platform, sharing the hardware resources, and thus saving the customer money. Some of the 17 ISVs include Artisoft, Sound Advantage, CTL, Elix, Envox, and TriVium. For a more detailed list, see Dialogic's Web site.

I'm excited that Dialogic has provided an open platform to allow several third-party developers to write applications for the Intel Converged Communications Platform. This certainly allows for some interesting possibilities. With that said, I do have one reservation about this recent Dialogic announcement. Why is it that, in an era of open standards, I still cannot choose any vendor's hardware telephony board I wish? That is, why can't I choose between Aculab, NMS, Dialogic, MusicTele.com, BICOM, or others, and be assured it will work with all telephony applications?

BRIDGING TWO WORLDS
After all, the most important thing telephony boards do is provide a bridge between the PSTN and the PC, which contains the telephony boards. Thus, incoming analog lines, ISDN lines, or incoming E1/T1 lines from the PSTN "cross the bridge" (so to speak) once the trunk lines are connected to a telephony board(s) placed in a PC. Thus, with analog, ISDN, and T1/E1 standards, shouldn't I be able to plug in any vendor's telephony board to terminate the trunk lines? Also, with the fairly recent advent of PC-PBXs, telephony boards also provide the "bridge" between the PC and the employee's desktop phones.

In most cases, phone extensions are terminated to a punch-down block, which is also terminated to a PC-PBX containing one or more telephony "extension" cards. Because of the PC-integration, applications such as unified messaging and desktop call control can be performed. The advent of the PC-PBX certainly opens up some potentially interesting communications solutions. In addition to my first idea of being able to use any telephony board for the trunk lines, shouldn't I also be able to use any vendor's telephony board to terminate the phone extensions? No such luck!

Unfortunately, when we talk about the importance of open standards, the telephony board industry has been slow to adopt this mantra. For example, why is it that if I want to expand our Artisoft PC-PBX system to allow for a second incoming T1 line, I am forced to buy a Dialogic T1 board? Why can't I bargain hunt for cheaper alternatives from competitors, such as Aculab, BICOM, Natural MicroSystems, Brooktrout Technology, or MusicTele.com? Not that any of their T1 boards may necessarily be cheaper than Dialogic, but I still like having choices.
I should point out that MusicTele.com analog boards are also supported on the Artisoft PC-PBX platform because of their universal API which allows developers to "plug-and-play" their boards into applications that were originally developed for competitor's products, such as Dialogic. Apparently, this universal API maps Dialogic API calls to API calls understood by the MusicTele.com boards. But why must competitors resort to "trickery" to be "Dialogic-compatible," or API-compatible with any other board for that matter?

Of course, there has been some progress to make telephony boards more open. Standards such as S.100, H.110, and others have been created to try to do away with proprietary API calls for each telephony board. Many of the telephony board manufacturers have embraced these standards. There have even been test "bake-offs" demonstrating multiple vendors' boards working in the same chassis. So why is that I still can't choose any board in compliance with S.100 and H.110 and use it, for instance with an Artisoft PC-PBX or Sound Advantage's voice portal product? Both support Dialogic's CT Media, which in turn supports the ECTF S.100 Revision 2 standard.

COMMODITIZE TELEPHONY BOARDS
One day while sitting at my desk, I pondered why -- after the supposed "bus wars" (MVIP vs. SCSA) were over -- the telephony hardware industry still hadn't become "commoditized" as the PC industry has. If one were to look at the PC industry, hardware is very much commoditized. For example, I can buy a PC with an Intel, AMD, or Cyrix processor, and software applications will work just as well on all three. Similarly, I can buy a sound card from Creative Labs, Turtle Beach, and others, and I can be assured they will work just fine with any of my software applications, particularly games. Can you imagine having to buy a separate sound card for Quake, another one for John Madden Football, and yet another sound card for Half-Life?

Fortunately, with Microsoft's DirectX open-API, you can be assured any sound card you buy will work with any software application. Why shouldn't it be the same way in the telecom world? The answer might be so simple that my technical mind just looked right past it. The simple answer is: partnerships. Let's face it, companies ally themselves to one particular telephony hardware vendor for two reasons:

1) Companies had to choose one telephony board vendor, since it wouldn't behoove them to hire multiple engineers to learn multiple proprietary APIs for each board manufacturer. This was especially true in the old days before H.110 and S.100 helped to alleviate this concern. So certainly learning multiple APIs was cost prohibitive, and often the customer didn't care what hardware was used, as long as the solution worked. But today, customers are much more savvy -- they want open systems and choices. They don't like being tied to a single vendor. One only needs look at IT personnel's backlash against Microsoft and their support of Linux to see that customers really do want choices!

2) Companies such as Dialogic often offset the marketing, exhibiting, and advertising costs of their partners -- but at a price. The partner must exclusively support Dialogic hardware only to receive financial support. If I were a startup looking to develop a product and market it, having an exclusive deal in exchange for financial support would certainly be beneficial. Of course, I shouldn't just pick on Dialogic. They are not alone in doing this. All the other board manufacturers have done the same thing. Dialogic just did it better, which is probably one of the major factors for its strong market position today. Interestingly enough, I should point out that Microsoft also has had exclusive agreements, which is one of the reasons it got itself into trouble with the Department of Justice.

THE HELL YOU SAY!
What I want to see in the near future is the ability to go out and buy any telephony board you want, plug it in, and you're ready to go! Of course, I'm sure some telephony board manufacturers would argue, "The S.100 standard is incomplete" or "We have some specially optimized API calls that S.100 doesn't support" or "Our converged telephony board platform supports not only voice, but also VoIP, voice compression codecs such as G.729a, and speech recognition. These require proprietary APIs."

I say nonsense! Voice codecs are standardized, and often the board manufacturers license the codecs from the same third party. This is also true of speech recognition algorithms. As long as S.100 adds support for IP telephony, speech recognition, and the like, there is no need for proprietary APIs. I should point out that standards such as SIP and H.323 should take care of the Internet telephony APIs rather than S.100.

IT'S ALL IN THE PERFORMANCE
Of course, some board manufacturers will do a much better job at speech recognition, add less latency when compressing voice packets, and have other performance advantages over their competition. And that's my point. I should be able to choose the best telephony board based on performance (as well as value) and not based solely on name brand or be forced to buy a particular brand due to some partnership agreement.

So am I just dreaming? Do you think the telephony board industry will ever become truly open and become commoditized like the PC industry and PC boards have? Or will loyalties and exclusive partnerships, and hence higher costs and fewer choices for the customer, win out? Tom Keating .

[ Return To The May 2001 Table Of Contents ]

I recently discovered a very cool utility that I thought deserved mention in my column. High Criteria (www.highcriteria.com) designed a utility called Total Recorder that has some very interesting features. I found this utility on the Internet when I was frustrated that the Windows Sound Recorder would only record 30 seconds of audio before stopping. Often I would conduct phone interviews, and I would want to record it using my PC, but unfortunately the Windows Sound Recorder was too limiting. 

After doing some research, I found Total Recorder and it had just the feature set I was looking for, including the ability to record sound being played by other computer programs (e.g., RealPlayer, Windows Media Player, Quick Time, WinAmp). It could even record live Internet broadcasts, thus I could now record streaming media to my PC for later listening. This very cool feature is unique to Total Recorder and is not available in any other sound recording program that I am aware of. I was also excited to learn that it also records Internet telephony conversations -- even full duplex. I tested it with NetMeeting and it performed flawlessly. Other features include:

• Convert different sound formats to WAVE formats (plain or compressed) or MP3 formats, providing you have a player capable of playing a particular format. 
• Record from any sound card input line. For example, you can record from CD or you can hook up a cassette player and record from it.
• A built-in scheduler lets you schedule future recordings and playbacks. Or you can use Total Recorders command line options with an external scheduler such as Microsoft Task Scheduler.

In TMC Labs, we love our PC console switches (also known as KVM switches), which allows us to hook up several PCs to a box and use a single keyboard, monitor, and mouse, saving a tremendous amount of space on the lab tabletops. As such, I was intrigued when I heard about a new KVM switch that actually works across the LAN -- now thats convergence!

Avocents new KVM over IP Technology provides administrators complete access to an unlimited number of servers or devices in their network, from multiple locations. With the DS1800, instead of carrying signals to screens and from keyboards and mice in the analog domain, it converts them to digital packets that may be relayed to clients over any IP network.