Convergence, Not Integration, Is The Goal
BY DAVID M. FRIEDMAN
In writing this piece a simple fact dawned on me: rather than
integration, the real goal of CTI (computer telephony integration) is
easily implemented and seamless convergence of media applications. The
idea of "integration" came from a mindset that knew
only traditional computer and telephone systems, and that recognized the
need to coordinate data and voice. Logically, they figured that
integrating the two would be the solution. Instead, this approach has
proved to be more short-sighted than comprehensive and long-term.
While these developers were on the right path with integration, their
solutions were created separately. Consequently, standards have evolved from
various perspectives with different requirements, and in the end only proved to be impediments
to the whole process of bring voice and data together.
CONVERGENCE ASSUMPTIONS
Before ubiquitous convergence can be accomplished through the use of
common standards, there are certain statements or truths which should be accepted throughout the
industry. The following assumptions set the groundwork:
- The modern PC (personal computer) is a standard hardware object that
-- solely through software -- offers virtually limitless information and
applications tailored by and for an individual user.
- The hardware resources of a PC, while standard and dynamically
accessible to the various applications running, can be optimized in
configuration to best meet the budgetary and performance needs of the
user (i.e., memory, graphics, other media processing resources).
- There is an ability to develop applications according to robust and standard
APIs (Application Program Interface) such as those supported by and
under Microsoft Windows.
- A PC is better than having a separate appliance for each of these
multitude of possible applications.
- In multi-user environments, servers have resource
configurations optimized for more limited tasks. Desktops then can draw upon the services of these servers
dynamically, as needed, rather than performing all required functions
internally.
- Having all requisite telephony hardware resources bundled into a
server with flexible configurations and a robust API is the only
logical way to seamlessly embed this type of media into enterprise
applications.
- CRM (customer relationship management) players are scrambling to
integrate different media processing and contact center applications
under their respective umbrellas.
Until your telephony infrastructure racks up next to your file
servers, routers, and hubs and communicates with your applications over
your network backbone, there is no convergence. While conventional PBX
vendors will be selling systems for a long time to come, contact centers
and other "new world" enterprises will not be fooled.
Proprietary telecom systems are at death's door and with them,
conventional CTI. The future lies in true media convergence, rather than
computer telephony integration through computers.
ALL ABOUT STANDARDS
The evolution of adopting open standards
has been occurring for a while now and we are far closer to convergence
than we have ever been before. There are three general types of standards in the CTI industry:
hardware/platform, communication protocols, and APIs. Having the right
combination leads toward highly effective convergence of telephony with
CRM and applications. However, I submit that with the exception of the
API, the others are either foregone conclusions or irrelevant.
Hardware/Platform. Having the right kind of hardware is an
integral part of achieving convergence. Your enterprise systems should
comprise the operating system and form factor for media processing and
communications interface components. The following are some industry
hardware standards:
- ISA (Industry Standard Architecture)
- PCI (Peripheral Component Interconnect)
- MVIP (Multi-vendor Integration Protocol)
- SCSA (Signal Computing System Architecture)
- H.100
- Windows NT
- Unix
- Linux
In choosing between all the different types of hardware, what's most
important are adequate capacity and a decent range of
functional components and manufacturers under one standard. Personally
speaking, my preference
is SCSA, with H.100 coming on strong (backward compatible with SCSA). With
both of those, I can build my generic telephony server.
Communication Protocols. In addition to setting the
hardware standards, establishing the format for transmission of data
(data/voice/video) -- or communication protocols -- is as equally important.
TCP/IP and ATM are really transports upon which other protocols can
"ride." H.323, SIP (Session Initialization Protocol) and WAP
(Wireless Application Protocol ) are geared toward voice and data transmission
over IP or wireless networks, while G.723 and G.711. are "coders,"
dictating compression of voice over an IP network.
As long as all the systems that have to
talk to each other use the same language, it doesn't matter
what the language is. TCP/IP is the only important one here -- most of the
others ride on top of it and will change and evolve over time. As long as
the conversation sounds good, who really cares what coder is being used (besides the
licensor)? Things like coders and media transports should be
re-configurable without modifying your applications.APIs.
Applications Programming Interfaces are standard sets of commands by a third-party application to control or permit control
of another system, such as a generic telephony platform (or conventional
switch). Some examples of telephony APIs are:
- CSTA (Computer Supported Telephony Application)
- TAPI (Telephony API)
- TSAPI (Telephony Services API)
- CT Connect
All of those specifically mentioned have the glaring
limitation of only addressing interfacing with a PBX or ACD.
There is no material support for the advanced call handling and voice
processing applications that add so much value in the contact center.
Hidden agendas of their respective proponents often, but not always,
impose
restrictions on adopters such as a single operating system or limited
hardware platform support. As for vendor-specific proprietary APIs,
they're the cause of the problem, but may also hold the solution.THE
HOLY GRAIL OF CONVERGENCE
Comprehensive telephony control -- anything and everything
you could ever need to do in telephony -- may be available from only
one or two vendors in the world, yet this type of API (in conjunction with
the generic telephony server previously discussed) is the holy grail of convergence. If our goal is to achieve true convergence,
there is no doubt that standards will play a part at many levels. Seeking
and choosing the "right" API seems to be the last remaining
element in the quest for true convergence and the end of CTI. It may
entail taking a risk and changing the way you view and implement
technology. After all, nothing worth achieving comes without risk or hard
work.
There are
still a few more obstacles in the road to convergence. Getting multi-vendor hardware and software to seamlessly
work in conjunction with one another has not yet been completely realized. While many systems
support integration,
true media convergence is still the goal. From the looks of the state of
the industry, the best
is yet to come.
David M. Friedman is the vice president of marketing & sales at Buffalo International, Inc.
Since 1989, Buffalo International has provided tools and systems to
telephony enable e-business and customer relationship management (CRM)
applications.
|