| Coming To Terms: Naming CTI
Products In any industry, there are always those that wish to
coin a term. Many have tried, and many more have failed. The same holds true for our
beloved CTI industry. While short acronyms and catchy phrases have their place, they often
cause more confusion, especially when a particular type of product has its acronym changed
many times. We believe there is a time for coming up with new or better acronyms, and
there is a time to settle on using one consistent terminology.
Recently, there has been a barrage of terms to describe "PC-based PBXs."
Everything from UnPBX, communications server, telecommunications server, CT server, and
CTI server has been proposed. Far be it for us to stop marketing personnel from using new
terms. We do take exception, though, when arbitrary new terms cause confusion or a
disservice to the CTI industry.
The explanation for creating new terms goes something like this: "Our product is
more than a PC-PBX. We do unified messaging, call control, and voice mail, too." This
may be true, but what people are mainly looking for when buying a PC-based PBX are the
switching capabilities of the PBX, plus the openness of the PC. Additional features such
as unified messaging or call control are secondary to the main functionality of a PC-based
PBX. This main functionality is the PC-PBX's "openness," and its ability to
incorporate additional CTI features inexpensively and more easily than a traditional PBX.
We don't know of a single PC-PBX vendor that is selling their product without
additional functionality such as call control or voice mail. Thus, our readers already
know that the product being sold has functionality that goes beyond a PC-PBX's inherent
switching capabilities. Let's not hype CTI products with vague marketing terms such as
UnPBX or communications server. If you want to distinguish your product as having more
features than other PC-PBXs, then put your feature list in your press releases or
advertisements.
In this vein, we recently held a vote on our Web site regarding the naming of the
PC-based PBX product, after Tom Keating's CC: column addressed this issue in the January
1998 issue of CTI� ("UnRaveling the UnPBX"). The final tally: UnPBX, 9.5
percent; Communications server, 30 percent; PC-PBX, 60.5 percent.
-The Editors, CTI � magazine
[return to the top]
Polycom Enhances Video Conferencing With
Voice-Activated Camera
Polycom, Inc., tapped the brain power of the Ivy League to enhance their award-winning
ViewStation video conferencing unit. The Brown University Research Foundation, which
manages the inventions made by Brown faculty and research staff, were granted U.S. patents
for technology that allows a camera to focus automatically on the current speaker during a
video conference. Under a licensing agreement with the university, Polycom gains exclusive
worldwide rights to implement the technology as a feature of its ViewStation group video
conferencing system, as well as in future video conferencing products.
The voice-activated camera works using an array of microphones. Sound source locations
are estimated 35 times per second, allowing the camera to train itself on the person
speaking, even if that person moves around. "Partnering with the original source for
camera tracking technology will allow us to further enhance our products and protect this
critical piece of intellectual property," stated Brian Hinman, Polycom's chairman and
CEO.
Polycom has also licensed a second U.S. patent from Brown University, which provides
support for this voice-activated camera tracking technology. For more information, contact
the company at 408-526-9000 or visit their Web site at www.polycom.com.
-- Dara Bloom, CTI� magazine
[return to the top]
DSL Modems Put T1/E1 Service Over Single-Pair
Copper Wires
Been waiting for your local telco to draw a T1 line to your doorstep? Your wait just
got shorter, thanks to a line of high-speed extension MSDSL (Multi-rate Symmetrical
Digital Subscriber Line) modems from WaiLAN Communications, Inc., that allow telcos to
deliver high-speed T1/E1 links using ordinary, single-pair copper wires. And they're doing
it without any performance degradation.
Called the AGATE 250CO (Central Office) and AGATE 250CP (Customer Premise), the modems
support T1/E1 line capabilities for either data transmissions or multi-channel voice
across the PSTN. The key is that they work without the normal T1/E1 repeaters needed every
4,000 feet, and can be drawn up to 5.5 miles. Telephone companies can double the number of
subscribers compared to T1/E1 circuits with the same number of copper telephone wires
used, and lower the cost of providing T1/E1, frame relay, ISDN PRI, or other digital
services. All the telephone company has to do is to join up the wire pairs at the
switching office.
The AGATE 250CO has a T1/E1 interface that links to the PSTN at the telco central
office, and uses only one unconditioned copper twisted pair. The AGATE 250CP interfaces to
the telco SDSL line on one end, and to a V.35 interface (on a router, for example) on the
other. Both AGATE 250CO and AGATE 250CP are required in order to implement a T1/E1
replacement or extension.
The AGATE 250 modems are available beginning this June, with prices ranging from $1,499
to $1,799. For more information, contact the company at 408-452-8081 or visit the
company's Web site at www.wailan.com.
-- Dara Bloom, CTI� magazine
[return to the top]
Hypercom Divvies Up The Bandwidth
You can never be too rich, too thin, or have enough bandwidth. When it comes to
Internet telephony and indeed Internet "anything," most problems revolve around
bandwidth: There's either not enough of it, or you spend too much to overprovision your
network. Hypercom's recent announcement of a new technique that allows reallocation of
bandwidth between voice and data calls during peak usage times -- Dynamic Rate Change
(DRC) -- goes a long way to overcoming those obstacles. In fact, Hypercom's Dynamic Rate
Change addresses two key challenges:
Reliability: If the network gets congested with voice and/or data calls, DRC
reallocates the network bandwidth in real-time to ensure that voice calls are not dropped
due to wide-area network "traffic jams."
Overpaying for bandwidth: Without DRC, users would have to pay for much more
bandwidth than they need on a regular basis, to accommodate those brief periods of peak
network traffic.
In a nutshell, here's how the DRC technique works. DRC is accomplished via variable
rate compression performed by DSP software in Hypercom's Integrated Enterprise network
(IEN) devices, such as the IEN 6000. The DSP adjusts the amount of bandwidth given to each
call according to user-defined priorities and in response to network congestion. For
example, a voice call may be allocated 9.6 kbps of bandwidth when it is initiated. If the
network encounters heavier traffic during the call, the IEN device automatically scales
back slightly on the bandwidth allocated to that call (in increments of 800 bps) as
required to distribute available bandwidth among all the active calls and prevent any from
being dropped.
The IEN 6000 is a high-density, multi-service access gateway that supports up to 1,024
voice or fax channels. For more information, contact Hypercom at 602-504-5000 or visit the
company's Web site at www.hypercom.com.
-- Greg Galitzine, CTI� magazine
[return to the top]
CyberTranscriber's Boon For Road Warriors:
Internet Transcription Service
While traveling, I used to call into my voice mail to leave myself a message with
meeting notes when it wasn't convenient to jot them down on paper. Speech Machines has
completely altered that routine with CyberTranscriber, which they bill as an
"automatic Internet-based dictation and transcription service."
The way it works is simple, but the results are powerful. Record your dictation by
calling their toll-free telephone number, speaking into a hand-held digital recorder that
uplinks to a PC, or using a microphone attached to your PC. Speak your punctuation as you
dictate. For example, say "period" at the end of every line, and "new
paragraph" when you want one. You can also pause the recording, collect your
thoughts, and then resume your dictation. When you end the call, the recording is
encrypted and automatically sent to the CyberTranscriber service center. CyberTranscriber
uses a 65-bit encryption algorithm, but will implement other security options for users
who require them. They'll also send you an e-mail receipt to let you know that your
document has been received. I'd love it if Speech Machines added an approximate turnaround
time to this acknowledgment; in the meantime you can track the precise status of your job
on their Web site.
The speech recognition technology for the system was developed in collaboration with
the Speech Research Unit of the UK's military equivalent of the U.S. Department of
Defense. Using speech rec, a computer-directed correction process, and a few good human
editors, a surprisingly accurate transcription is generated. Mike von Grey, Speech
Machine's president and CEO, told me that ideally, the human editors are handling two
percent or less of the transcription -- CyberTranscriber does the rest. The service center
encrypts the document and sends it back as an e-mail attachment. When you receive the
return e-mail, just double-click on the attachment. The company promises a one business
day turnaround time, but I had my transcription back in exactly an hour.
I dictated a letter I had already written down so I could compare the original typed
document to the transcript returned by CyberTranscriber. I was impressed -- most glitches
were the result of incorrect commands on my part. The learning curve for the commands is
quite short, though, and those minor formatting errors certainly didn't affect the
accuracy of the content. In fact, CyberTranscriber had no problem with words I'm
constantly spelling out for people -- such as my name. It didn't stumble over an abundance
of acronyms, either. CyberTranscriber is sophisticated in its transcription; it recognizes
homonyms and chooses a word's correct spelling based on context, for example.
Von Grey also noted that while individual subscribers can access the service directly
from the Web, Speech Machines is really a business based on resellers, not end users. For
PCS providers, CyberTranscriber is yet another enhanced service to offer customers. Speech
Machines has deployed their service with Cellnet, a UK cellular service provider, and
they're in discussions with several U.S. wireless vendors and even some traditional
telcos. For more information, contact the company at 650-568-1500 or visit their Web site
at www.speechmachines.com and sample the
service for free.
-- Dara Bloom, CTI� magazine
[return to the top]
BittWare And NASA -- The Sky's No Limit
BittWare Research Systems is going for a ride on the Space Shuttle Columbia. (Not the
whole company, mind you, just one of their SHARC-based DSP boards!) Their Snaggletooth 40
MHz ISA board will be the main voice recording mechanism for NASA's Virtual Environment
Generator (VEG), which in turn is part of NASA's Neurolab experiment. The Neurolab Sensory
Motor Experiment 136 will use the VEG system, a space-qualified 3D graphics processor,
wide-field view helmet mounted display, and joystick designed to provide a controlled
interactive virtual environment. Neurolab is a set of life-science experiments that test
human adaptability to zero-gravity environments. Developed by Lockheed Martin, the VEG's
headset microphone tracks the astronaut's comments, and simultaneously archives and
downlinks the voice data to Earth in real-time.
BittWare's Snaggletooth board, together with an AMBE-1000 Vocoder (from Digital Voice
Systems, Inc.), provides the main voice compression engine for the VEG, compressing at a
constant bit rate of 4 Kbps. Voice is compressed on-board, and then downlinked to NASA's
Johnson Space Center Mission Control, where it is decompressed by another BittWare
Snaggletooth in real-time.
The VEG platform, based on a 200 MHz Pentium Pro with 256 MB of RAM, includes a
13-board passive backplane in a custom ruggedized drawer. The Snaggletooth SHARC board is
configured with an 8-channel analog audio I/O mezzanine with Digital Voice Systems' voice
compression, and 16 channels digital I/O. The system further includes 5 cards for virtual
reality graphics generation (1 accelerator card and 2 cards each for left and right eye
generation).
According to Trent Mills, VEG Systems project leader for Lockheed Martin, the BittWare
DSP board was chosen because of its ability to achieve multichannel
compression/decompression at the 4-kbps rate. For more information on BittWare, call
800-848-0436 or visit the company's Web site at www.bittware.com.
-- Greg Galitzine, CTI� magazine
[return to the top]
Quallaby Redefines Service Level Management
I recently had the opportunity to speak with John Kostak and Jeff Gillespie of Quallaby
Corp., an international provider of adaptive service level management software. They told
me all about their new PROVISO product family. Neat stuff. The PROVISO offers a whole new
class of customer-defined network management capabilities. The solution allows NSPs
(network service providers) to actually customize service level agreements, customer by
customer, and even allows NSPs to let their customers handle their own online
provisioning.
One of the keys to the success of PROVISO is that the software employs third-generation
metrics to process customers' performance information based on SNMP, RMON, RMON2, or other
information; it applies predefined business rules to that data; and presents real-time or
historical reports back to the customer. PROVISO supports both frame relay and ATM, and
allows NSPs and major corporate users to create business driven service level agreements
(SLAs) based on four factors: availability, throughput, latency, and utilization.
According to John Kostak, "SLAs are hot, providing uptime and performance levels
across the IT infrastructure. With customer-defined service level management, service
providers -- for the first time -- can give customers self-service network management
capabilities to define and control their networks."
To create these customer-defined services and management, Qualaby also offers PROVISO
Studio, an object-oriented data services warehouse. Using the Studio software, service
providers can create service classes based on data maniplation and multidimensional or
formula-based metrics. Hardware requirements for the PROVISO software are simple: the
product will run on a Microsoft NT Server (3.5 or higher), HP UX, or Solaris. The product
is in beta testing with some folks who have fairly large networks. I'm not sure if I can
mention them now, but one's a mammoth ISP and one is a large utility company. Product
availability is scheduled for the end of the first half of this year, with pricing for an
NT or Unix server at $15,000 and $25,000 respectively. For more information, call Quallaby
at 781-935-4700 or visit the company's Web site at www.quallaby.com.
-- Greg Galitzine, CTI� magazine
[return to the top]
CallManager Changes The Paradigm For Least Cost
Routing
How can you argue with a product that makes their profit off of a percentage of your
cost savings? Joe Cline, the president of CallManage, calls this concept "paying for
performance." If their least costing routing system doesn't save you money, you don't
pay.
CallManage's flagship product, CallManager, is a software technology that works with
your PC, PBX, or Key system and modem to find the least expensive rate for each call.
Based on the company's patented Intelligent Selection technology, CallManager monitors up
to 30 parameters, including time of day, call destination, long-distance carrier specials,
and current industry rates to find the best value for each call, providing a "per
call" optimization. Cline says the system is also set up to handle Internet telephony
calls, although this functionality is not currently being deployed.
CallManager performs a rate look-up and leverages carrier access codes to transparently
direct your call to the best available long-distance company. The software automatically
updates itself (that's why you'll need a modem) to ensure you're getting the most current
industry pricing. By the time you dial the final four digits of your phone number,
CallManager has already determined how the call will be routed. It is completely
transparent to the user.
The company has a vision of the "CallManager Ready" logo becoming as
ubiquitious as "Intel Inside." Several strategic partnerships with third-party
vendors (including Packard Bell Platinum PCs and Artisoft's TeleVantage software) are
setting the company in the right direction. Currently, about seventy percent of
CallManage's revenue comes from business customers, who in beta tests of CallManager saved
an amazing average of sixty to sixty-five percent off of their long-distance bills each
month. CallManage charges twenty percent of your call savings for this service. There is a
residential plan, too, which has a flat-fee billing system. For more information, contact
the company at 203-351-0880 or visit the company's Web site at www.callmanage.com.
-- Dara Bloom, CTI� magazine
[return to the top]
Mockingbird Introduces Intelligence To IP
Networks
Mockingbird Networks' Nuvo 100 PSTN-IP Telephony switch, designed to seamlessly
interface PSTN voice circuits with IP networks or backbones, allows carriers and ISPs to
interface to local exchanges and offer Telco-quality telephone service to their customers,
while bypassing traditional long-distance networks. One of the more significant
innovations here is a new class of Intelligent Network (IN) SS7 provisioning, which allows
Nuvo switches to imitate standard PSTN voice switches while converting PCM voice or fax
circuits to IP packets.
Interfacing to the IN also allows the Nuvo switch to support standard features of SS7
such as:
- Instantaneous call connection (no long delays as when using Internet telephony gateway
products).
- The utilization of SS7-based Call Detail Records (so charges can appear on the
customer's phone statements).
- Full support of IN features that phone customers have come to expect, including caller
ID, call forwarding, and multi-point conferencing, among others.
The highly scalable Nuvo switches are designed with built-in fault tolerance with fully
redundant hardware elements, including MVIP switching. A single Nuvo 100 switch supports
96 full-duplex voice channels, and can be easily expanded to support up to 10,000 channels
by adding additional Nuvo switches. The cost per port takes a steep drop in large-scale
POPs, since the SS7 provisioning software and IN hardware is only required in the primary
switch. This makes Mockingbird's switch a good deal for carriers who want to start small
and then increase switch capacity as demand for IP telephony grows.
Nuvo switches incorporate state-of-the art DSP technology to provide real-time voice
compression. PCM voice channels are compressed from 64 Kbps down to 4-6 Kbps, then routed
to 100BaseT Ethernet ports. The Nuvo 100 dedicates its DSPs only to
compression/decompression while using specialized hardware ( | 
