One is volume. The other is value. One tremendously determines the top-line.
The other greatly influences the bottom-line. Voice and data, they do co-exist
in different ways in the networks of different carriers. The present, itself is
confusing. The future, as envisioned, a heady mixture of success, uncertainty
and risk.
The decision-making for the operators is certainly tough. Do they go in for
complete fork lifts to implement an all-new IP network that has been optimized
for data services? Or do they stay on course and add more tandem trunks to their
switches, so that they can handle the surge of increasing level of data?
The consensus, after a lot of debate, seems to be the traditional Indian
belief–the middle path. In simple terms, putting in place an efficient
migration system that seamlessly transforms a traditional TDM network to an
IP-based network.
The Evolution of Packet
Packet-based
telephony has evolved quite rapidly since the time it was introduced in 1995.
During its adoption phase, it emerged merely as a hobbyist’s pastime when the
PC was used as a cheap option to make telephony calls in an unregulated market.
But today, when one tracks the evolutionary trends in packet telephony, it is
seen that on the services front, the development curve is reaching almost
vertical levels.
So the question is: where is the infrastructure that supports such services?
Surprisingly, while the scope of services that can be offered on a packet
network is skyrocketing, on the infrastructure front, technology is reaching
more mature levels and is approaching a plateau stage. This signals the fact
that carriers no longer need to rely on complete fork lifts to provision
IP-based services in their networks. Thanks to the efforts made by some vendors,
who have assessed the volatility of the situation and have come up with
technologies and solutions that help carriers seamlessly transform traditional
TDM networks into the packet domain.
Market Drivers
Before entering into the era of ‘Next Generation Networks’ (NGNs) let us
have a look at the prime market drivers.
n Surge in
bandwidth demand: One of the prime drivers is the surge in the demand for
bandwidth by the end users, as they consume more and more bandwidth and demand a
growing portfolio of data and voice services. This has led many service
providers to manage more than 6—7 separate networks deploying special multiple
purpose networks. But then the overheads in managing these networks also
increase in such a scenario. This has led some of the more aggressive players to
look at options of deploying multi-service networks that they claim will save up
to 70 percent in terms of cost.
n Deregulation:
Deregulation has been taking place in most of the regions and this in turn is
increasing competition. And increase in competition means putting pressure on
the service margins of operators. It is because the only differentiation that
service providers can make at this point is the price. Therefore, it is
imperative that service providers deploy newer services much faster than the
competitors. Therefore, the NGN must offer integrated network management to make
it easy for operators to run the networks efficiently and sensibly, and deliver
premium levels of service and responsiveness with minimum operational costs.
n Importance
of QoS: Service providers are signing SLAs that require high quality-of-service
(QoS). It is because many multimedia applications require quite stringent QoS in
the network elements. Therefore, NGNs must preferably offer end-to-end QoS
control via one distributed management system. Further, insights into the market
reveals that there is no such thing as a homogeneous serving area. It is because
of the varied demand pattern of customers, so it is important to be able to
satisfy them, without increasing network complexity. Thus, the NGN must be able
to support multi-protocol service architecture, so that it would be possible to
provision different levels of service in one managed network.
n TDM networks
not optimized for data: Another major factor, which is a market driver for NGNs
is that the traditional TDM networks are not optimized to carry data. But the
surge of data has forced operators to add more telephony trunks to match the
onslaught of data traffic. But this is a costly and inefficient way to carry
data. While TDM switches have been inherently profitable for operators to carry
voice and huge investments have been made to build these networks, calls for the
operators to choose a smooth transition path to the multi-service ‘Next
Generation Network’. The scenario is especially true for incumbent’s network
where there are numerous small switches in the network. Given such an
environment, they involve a huge operational cost to operate and provision these
trunk groups. The other issue with TDM networks is that new service introduction
usually requires service upgrades to be made in all the switches.
Best of Both Worlds
The more prosperous service providers in the future will be the ones who are
able to retain the best of their existing networks while transforming them with
the best of the new networks. With judicious and well-timed investments, service
providers can convert their existing voice-optimized networks into a completely
multi-vendor compatible, multi-service network, optimized to carry IP and
end-to-end services over a variety of architectures. The network transition
process is critical, as they still have to carry on with the existing services.
In the existing TDM-based networks, significant savings can be achieved by
consolidating several layers, as this could translate into capex and opex
reductions.
Consider asynchronous transfer mode (ATM) for instance. With the right ATM
architecture, service providers would save a big chunk of change eliminating
cross connects and network elements, while still exploiting the efficiencies of
ATM to consolidate routes with switched virtual circuits and slashing trunk
administration costs. Therefore for serving a customer base that demands various
levels of service, ATM turns out to be a clear choice.
Services: Closer to the User
A discerning look at NGNs reveals that it means provisioning services on a
‘distributed network intelligence’. Distributed network intelligence means
that the intelligence resides in the standards-based servers that reside on the
edge of the network, unlike legacy networks where the intelligence resided on
the switches.
Under a distributed intelligence architecture, we would be entering into an
‘Open Programmable Era’, and this is what a NGN is all about.
It is because the intelligence residing on the servers at the edges allow
custom development for service providers and third parties, so that they can
configure services based on their requirements. Whereas at the core, packet
telephony application is agnostic for the core connectivity, therefore it is
immaterial if the core backbone is IP- or ATM-centric. So the issue really
resides in the edge of the network. The NGN is based on a distributed rather
than on a nodal architecture.
The new network moves access and services out of the switch-based backbone to
the servers residing on the edge. In this process, network and service control
move into the hands of service providers and ultimately to the subscribers.
From core to edge |
|
![]() |
The emerging paradigm is that of distributed network intelligence |
![]() |
Here, the intelligence resides on the standards-based servers |
![]() |
This is unlike legacy networks where the intelligence resided on switches |
![]() |
Service providers and third parties will be able to configure services based on their requirements |
As the convergence of voice and data networking advance, the most successful
service providers would be those who implement a rapid cost-effective
transformation strategy that reuses the key elements of the TDM network to
provision next-generation services. While the introduction of an ATM switch can
be the first step to commissioning a NGN, a generic NGN architecture would
comprise of a multi-service delivery mechanism using multi-service gateways to
carry traffic onto a high-speed broadband ATM switching fabric or on MPLS core.
With such architecture, the definition of a ‘central office’ takes a new
dimension unlike that of the TDM networks. Conceptually speaking, it assumes the
nature of a distributed architecture consisting of multi-service gateways, ATM
switching platform or MPLS switches at the core, and finally, the server based
intelligence to pull it all together. It means that many of the functions of a
TDM central office, such as SS7 connectivity, call control and a unified view of
network management, are now provided through the ‘new central office’, the
next-generation network.
Therefore, in this transformation scenario, familiar and profitable services
can still continue uninterrupted because the integrity of the SS7 signaling,
which is essentially a TDM functionality, and the value is preserved even as
connection control is passed to the ATM or MPLS devices. In the logical
extension of this transition strategy, TDM components would be slowly phased
out, paving the way for applications and services to migrate to network servers.
And moving these components to the network servers would mean that custom
services can be developed by the service provider and will eventually move down
to the subscriber.
Finally, should we say that a rapid transformation into a packet network
should be the one that reuses the components of the TDM network, which is the
logical business course for any service provider to enter into ‘an open
programmable era’ of NGNs. Logically, it does makes business sense to preserve
the value, in terms of services and revenues of the TDM network, protecting the
huge investments made, while transitioning into a data-centric future.
Sandeep Sharma, Manager (product marketing) VoIP and intelligent Internet,
Nortel Networks