All, For a better understanding of why open programmable switches are important to VoIP, and to enhanced services in general, see this month's Telephony Magazine's focus article on the topic at:
internettelephony.com
Nortel's DMS is mentioned throughout [as they too strive for openness in progammability] with references made to others, sometimes generically, including Summa Four. The article is posted below for posterity, but you'll need to go to the actual web site for graphs and charts.
Enjoy, and Regards, Frank C.
==========================
July 27, 1998
Programming the future network
Open switches' roles are expanding, but a full network turnover
is nowhere in sight
WAYNE CARTER
The beginning of the end for circuit switches is at hand.
Carriers acknowledge that trying to keep up with mushrooming
demand by using traditional circuit switches is a losing battle. To add
services and features to the giant Class 5 switches, carriers must turn
to switch vendors for new applications, and incorporating such new
features into the big switches isn't a simple operation.
Enter programmable switches, which have been in use since the late
1980s as adjuncts to large switches. Carriers can simply connect a
programmable platform to an existing switch to immediately add
enhanced-services capabilities. The platforms can accept
programming from a variety of vendors, adding flexibility in choosing
enhanced-services applications.
The programmable switch itself is one of three pieces. There also must
be controlling software, known as the host program, and external
media to provide speech recognition and other enhanced-services
functions.
Under the control of the host program, programmable switches
process calls and perform all kinds of intervention functions, such as
play prompts and perform credit-card checks, says Peter Carlino,
business development director for Summa Four.
Summa Four and Excel are the clear market leaders in programmable
switching. They have established themselves as strong providers of
adjunct solutions, but they are now looking to expand their switches'
roles in the network.
"The programmable switch at one time was used in a very niche
market," says Paula Foster, marketing director for Excel.
"Programmable switches are moving into the networks themselves,
providing enhanced services and network routing."
Although programmable switches have mostly worked with larger
switches, they can act as stand-alone switches as long as the load is
not too heavy. As smaller competitive local exchange carriers and
wireless operators spring up, the opportunities for relatively
inexpensive, very flexible and scalable switching platforms are growing
proportionately.
"Looking at the computer industry, this is not a new concept," Carlino
says, adding that computer hardware and software companies have
figured out their core competencies, and through open systems, users
can get the best possible solution at the lowest possible price.
A good parallel for programmable switches and traditional switches is
personal computers and mainframes, says Probe Research Analyst
Hilary Mine.
"If you need flexibility and to add applications quickly and cheaply, a
PC is good," she says. "On the other hand, if you need to process
billions of transactions, a mainframe is cheaper. The same arguments
apply [to switching]."
That's how programmable switch vendors hope to move their
products into the space of primary central office switches: targeting
carriers that don't need the call processing power of a Class 5 switch.
A start-up carrier can even choose not to deploy a programmable
switch at its full capacity, then easily scale the platform to meet
demand as its business grows.
That type of switch, one that lets the user quickly and easily expand
capacity and change services and features, will eventually become the
type of switch for which carriers predominantly look.
"These switches themselves are going to be the basis for the next
generation central office switching systems," says Ken Kelly,
Dataquest senior analyst. "Programmability, scalability and
survivability, combined with packet switching, that's what you're
looking at as the next generation of switches."
However, it will be some time before the new generation usurps the
old guard's dominant position in networks. The traditional switches
still function well, and carriers aren't going to toss away a huge
investment that hasn't lived out its useful life.
"There's a heavy embedded base of circuit switches, and nobody can
afford to throw them away," Kelly says. "We'll need the big switches
for some time to come."
Carlino agrees, adding that he wouldn't suggest anyone throw away
Class 5 switches.
Target audience remains
That doesn't mean the outlook isn't changing for switching vendors.
Circuit switch bastion Northern Telecom is already more than a year
into a program that adds programmability to its stalwart DMS
platform. The idea came from Nortel's large customers.
"All of them were using third-party programmable switches--in some
cases, more than one manufacturer's [programmable switches]," says
Mike Doerk, senior brand manager for Nortel's programmable
switching platform. "The common theme was, 'We'd like Nortel to
develop the ability on the DMS platform to perform the same
functions.'"
Nortel developed software that allows programmable-switch
capability to be activated on in-service DMS switches, but the
company also allowed for the solution to be deployed as simply a
programmable switch (Figure 1).
"It can act as a stand-alone programmable switch and do nothing but
programmable-switch functions under the control of an offboard
[host] program," Doerk says. "We're bringing to the programmable
switch market a platform with the same levels of reliability,
redundancy and robustness that the DMS platform has demonstrated
over the years."
Still, the Nortel platform is better-suited for a different target market
from the market the smaller programmable switch vendors have in
mind. In comparing the DMS platform to other programmable
switches, Doerk points out that current programmable switches
typically support 1500 to 2000 trunks, whereas the DMS can support
tens of thousands of trunks on a single platform. That's not the kind of
capacity that small start-up carriers typically would need or want to
pay for right off the bat.
"I think what will shake out is that Nortel is more cost-effective for
larger carriers," says Mine.
Yet the fact that Nortel is turning its attention to the matter and
opening its core platform to applications from other vendors is a telling
sign that programmability is becoming increasingly important for
switch operators. It's not a stretch to expect other major switch
vendors to follow suit, Mine says. In the traditional switch world, all
intelligence resides on the switch itself. Switch vendors must develop
new applications, and that will happen only if enough demand exists
among users for a new application.
"What Nortel did is say, 'Let's leverage the intelligent network, open
the DMS platform application programming interface and allow others
to program for it," Mine says. "All switches are programmable. It's a
question of who can program for them."
Nortel's entry into the market is less a threat to individual competitors
than a boost to the overall market, says Excel's Foster.
"The fact that Nortel decided to introduce a programmable switch
architecture is really great," she says. "A company in the traditional
switch market recognizes the need for programmable capability. It's
similar to IBM holding back on PCs until the market was made by
other companies. [Nortel's presence] will only continue to validate the
market for open programmable switches."
As with personal computers, the more open the platform, the greater
the number of choices the user has in setting up a station or network.
PC users can choose a variety of hardware and software options to
custom-fit their needs. And as programmable switches become more
prevalent and more vendors develop applications for them, carriers
will have similar flexibility in building networks.
"The network is moving toward the convergence of voice and data,
voice over IP," Foster says. "As long as we have a platform that's
open, we can easily add those new services."
As more companies develop the necessary applications to provide
those services, the easier it will become for carriers to take advantage
of programmable switches, says Rob Rich, senior vice president of
telecommunications research at The Yankee Group.
"[Switch operators] can get intelligent hosts with call-processing
software to drive the host, and it essentially becomes a switch," he
says, adding that increasing scalability will allow programmable
switches to handle significant traffic volumes. "Certainly [a
programmable switch] would be a lower cost than a tandem switch."
The advantages of programmable switches are already being
exploited overseas and in budding PCS networks, Summa Four's
Carlino says. It will be crucial for switches to keep pace with the
increasing complexity of the network as Advanced Intelligent
Networking standards evolve.
"There's no such thing as a basic network. There can't be," Carlino
says. "All calls are going to be intelligent to some degree."
Dataquest's Kelly says the growing need to move traffic other than
voice, and the intelligence required to do so, will drive demand for
programmable switches.
"We're seeing things become more data-centric," he says. "You have
to be able to handle that data, and the current generation of switches
isn't able to do that. [Switches] must also be programmable for rapid
change to keep up with the transport area."
Architecture plans differ
Programmable switch vendors may be driving in the same direction,
but they're taking different roads. While Nortel is building on its core
platform, Summa Four and Excel are evolving their platforms in
opposite directions with regard to architecture.
Excel is embracing a distributed architecture with its Open Network
Expansion architecture platform (Figure 2). It uses Excel's
Expandable Switching System hardware and software and Excel's
patented Programmable Protocol Language to allow carriers to build
a distributed switching architecture that is open to integrate unrelated
resources, such as interactive voice response, voice recognition and
voice-over-Internet devices, into the switch. The system is designed
to work with any host platform, and applications can be written
independent of the network protocol. The system also allows for new
technologies to be incorporated via daughter card processors,
eliminating the need for system redesigns for such implementations.
"Our story of openness applies to every piece in the architecture, and
it's not just programmability," Foster says. "Operators can change the
software that runs the switch itself. As we move more and more into
the network, ONE architecture allows us to provide network routing
and management services required in the network infrastructure
environment. We can now position ourselves as not just a
programmable switch. Very often, we're seeing ourselves being
compared with traditional switches."
With its new architecture, Excel plans to offer a platform that can
support small carriers' needs and grow to a very high density.
"[Customers] can purchase as few as a couple hundred ports and
grow to meet capacity requirements up to 30,000 ports in a single
switch," Foster says.
While Excel's thrust is for a distributed architecture, Summa Four is
developing an integrated programmable switch with its Project Sigma
initiative. The idea is to integrate the switch, the media (the elements
that provide various capabilities such as voice, fax, speech recognition
and voice/fax over IP) and the host into a single box (Figure 3).
The project is still being developed, but Summa Four has already
clearly defined how the system will work. It will consist of three
products: a media server designed to meet intelligent peripheral
requirements; a voice-over-IP gateway that supports up to 2000 fully
redundant voice-over-IP connections and a variety of compression
algorithms for voice and fax; and a transport platform that scales up to
16,000 non-blocking ports and is designed for use as a
programmable Class 4 or 5 switch or, for wireless, a base station
controller, mobile switching center or wireless local loop controller.
"Media and switching are now fully integrated," Carlino says. "You
can move calls around and provide all the services you need. It's a
core switching platform suitable for CO deployment."
The intended small footprint for the Project Sigma devices should
make it appealing in terms of real estate costs as well as its lower
initial cost to deploy. By reducing the space a single platform requires,
carriers can scale their switching centers to greater densities without
having to expand their actual floor space to do so. Those factors
should be especially appealing to newer carriers that are just building
their networks, and a network starting from scratch is an ideal
opportunity for programmable switches to make inroads, Carlino
says.
"Newer networks aren't tied to [legacy equipment] constraints," he
says. "It allows them to build intelligence from the start and to be
scalable."
Changing with the times
With all the new competition cropping up, established carriers are
beginning to change their thinking to match that of their new
competitors, says Alex McCarthy, business manager of telco network
services at Dialogic, which makes media for programmable switches
and has worked closely with Summa Four on Project Sigma.
"A lot of Bell [companies] are setting up internally and behaving
entrepreneurially," McCarthy says.
Part of that entrepreneurial behavior involves offering new services to
customers. However, offering a new service doesn't guarantee
success. Different areas of the same city may have different demands
for services. Business districts and residential areas are evidence of
that. The question for carriers becomes one of determining what
services will do well in which areas.
Nortel's Doerk says that integrating the programmable switching
functions into a primary switch allows carriers greater flexibility in
trying out new services.
"A number of enhanced services have different busy hours," Doerk
says. "Very often, carriers have excess capacity on certain trunks. At
a cost much lower than a third-party programmable switch, [carriers]
can have programmable switch functionality, and they don't have to
add hardware or trunk capacity to try an enhanced service. If it takes
off, they can add additional resources to handle demand. If not, they
can move on to the next enhanced service."
As programmable switches become more important to network
operators, the switches themselves will have to evolve, as will the
vendors and their strategies. The programmable switches enjoy
several key competitive points. They include providing the same
capabilities and reliability that current CO switches provide, as well as
supporting industry standards such as ATM, frame relay and
eventually, voice over IP. But the crucial element is building
programmable switching into a standards-based function.
"The future for programmable switching lies in being established as a
valid, standard architecture for switching in the future," Excel's Foster
says. "We want to build awareness of programmable switching and
have standards for the future."
As such standards evolve, it will become easier for more companies
to write applications for programmable switches, leaving carriers in an
interesting quandary, says The Yankee Group's Rich. The more
applications vendors there are, the more difficult it will be to separate
the wheat from the chaff.
"It ultimately opens up to a much larger group of applications
developers," Rich says. "Having it open to more [programmers] gives
more choice and you probably pay a little bit less, but [carriers]
obviously don't want any Tom, Dick or Harry who's a Unix hack"
writing applications for their switches.
But, as Dataquest's Kelly points out, those days are still far away.
Despite the advances in programmable switching, such devices aren't
ready to take over for the bigger switches, and will remain still in a
specialized--if slightly more expanded--role, he says.
"They're lacking the services to use as a core switching system. The
feature content in big switches is humongous," Kelly says. "The good
news is, one size doesn't have to fit all. They can be used in rural
areas where you don't need a ubiquitous switch."
It is clear that while programmable switches will continue to grow in
importance, the growth curve will be shallow. Open programmability
may become a more common feature among switch vendors, but an
absolute change to programmable switches is not going to happen in
the foreseeable future.
"There's no one type of switch that's going to win out," says Mine.
"CLECs are increasing demand for smaller, more flexible platforms,
but there will be continued demand for large-scale networks."
Programmable switch vendors agree. Despite the company's belief
that its platform is a viable stand-alone switch, Summa Four's Carlino
says the company doesn't expect its switches' roles to entirely change
from service-provisioning adjuncts to CO anchors.
"It's a dual role," Carlino says. "We continue to see growth on both
sides." |
