May 1999
What It Takes To Join The Carrier Class
BY RANDY BRUMFIELD
With the emergence of a new IP/ATM WAN network architecture comes a new breed of
carrier products, referred to as "media-aware" switches. These products are
designed to support a whole host of circuit and packet-based services for converging the
WAN access and core. They have been specifically designed to address the stringent
requirements needed to maximize service revenues and network uptime, and they must meet
the same standards as today's circuit-based products. However, the concept of carrier
products goes beyond the words "carrier-grade" and Telcordia Technologies'
(formerly Bellcore) Network Equipment Building Standard (NEBS).
It becomes increasingly strange that the terms "carrier-grade" and
"carrier-class" are so loosely used in the industry. What's even more amazing is
how these terms are defined by many, especially the product and service categories they
fall into. The area largely affected is classifying which products fit into the telephony
segment.
DEFINING THE CARRIER CLASS
Over the past year, we've seen a big paradigm shift in the telecommunications market
regarding how the next-generation network will be designed and implemented. The market for
the new telco arena is projected to be large - read any recent market report to get a feel
for the billions of dollars that will be attributed to these new networks. Because of
this, vendors are targeting new and existing products to ensure, from a positioning sense,
that they can somehow differentiate themselves from the competition. However, most believe
that because their product has been certified NEBS-compliant (Telcordia specification for
what we call "shake-n-bake" test), that they are carrier-class.
All this means is that the product passed a stringent test for mostly environmental
conditions - kind of like getting the USDA stamp of approval! This is an important item,
especially for an ILEC, CLEC, or any carrier that requires collocation in central offices
or point-of-presence locations that require this certification, but should not be the only
item that would warrant this grand title.
Another area includes products that support DC voltage, or even some that are
CompactPCI-compliant with the tag line of "carrier-grade" - all attributed to
hot-swappable modules, or redundant power supplies.
Enough already! We rarely see a compilation of the other functionality and performance
requirements which need to be met to support this "holy grail" carrier-grade
term - a term that is increasingly being redefined as carrier-quality.
In addition to the environmental requirements, the other areas service providers should
take into consideration include:
- Performance: Low-latency, de-jitter, minimal delays attributed to packet processing.
- Quality: Toll quality, manufacturer quality (ISO 9001, ISO 9002).
- Signaling: SS7, Q.931, CAS, MF.
- Resiliency: High-availability and high-reliability, 99.999% available uptime.
- Scalability: In terms of thousands of ports within a rack space.
- Management: OSS, including billing (CDRs).
- Migration: Circuit to packet, circuit signaling, and packet interworking. Interworking
with existing voice switches, and circuit-switched services and features.
NETWORK REQUIREMENTS
Service providers will not tolerate downtime in the network. Any single point of failure
is catastrophic to the revenue generating aspects of their business. Telephony is the
"bread and butter" of all ILECs and IXCs in the market today - you can always
pick up a phone and make a call (unless of course there's a natural disaster), and
equipment in the network should not cause the circuits to become unavailable for use. The
current telephony infrastructure has been designed to help avoid such outages, both from
the equipment side and the network route end.
Today's network service providers and equipment vendors have put in place mechanisms to
notify end users of unavailable service. For example, on Mother's Day, if the end user can
not place the call through the network due to congestion, they'll be notified with a
"fast busy" or an announcement of "All Circuits Are Busy." These
conditions were put in place to ease end users' minds so that they still have connectivity
to their carrier or service provider. Granted, they don't like to hear this, but it does
reassure us in some strange sense that there is something moving through the other end of
the wire and the phone jack.
The other side of having high availability and fault-tolerant capabilities is the
carrier and service provider resources. The requirements for providing fault-tolerant
hardware and software are meant to lessen the need for rapid response to failures. Not all
sites are fully staffed, and require a good MTTR (Mean Time To Repair) cycle in order to
make physical changes to the equipment.
For telephony circuits, there are mechanisms in place to help identify and troubleshoot
lines and paths of the network. Proper detection of any fault within the network and
resolution must be capable of moving into the new generation voice-over-packet model. This
includes fault isolation on call setup, such as Continuity Test (COT), tone detection and
failure, and test line (including loopbacks). This is on the circuit and signaling side of
the path, and there is also the behavior and testing of the packet side of the network,
including packet performance, statistics, and test inject functions. The goal is to
isolate and remedy the problem as quickly as possible.
Signaling aspects also need to be looked at when evaluating your particular
requirements. Signaling System 7 (SS7) is the most common signaling access interconnect
between switches and networks today. It also provides the mechanisms to leverage the
existing Signal Transfer Points and Signal Control Points for Advanced Intelligent
Network/ Intelligent Network feature sets, including the requirements for supporting LNP
(Local Number Portability). LNP is key as it will become a mandatory function for any
carrier or service provider. LNP support allows end users to maintain the same telephone
numbers if they switch their local carrier or service provider. There are also the
interconnects between switches that range in the Feature Group D spectrum (bearer channels
and signaling) - Q.931, SS7, etc., and IMT (Inter-Machine Trunks).
Signaling is one of the most important items when evaluating the requirements for
integrating both circuit and packet-switched technologies. There are new frameworks and
requirements that are being worked on within the IETF, ITU-T, ETSI, and Telcordia
standards bodies to help define these interworking issues.
A NEW CLASS OF VOICE
When we talk about carrier-quality, one of the items that comes to mind is voice quality.
There's been a long-standing notion that migrating circuit to packet won't give users the
same voice quality as they are receiving today. Most of this controversy is based on early
implementations of voice codecs and digital signal processors (DSPs). The emerging media
switches are leveraging newer DSPs and are improving the packetization schemes to deliver
toll-quality voice. Earlier implementations of packetized voice attributed more than 150
ms of delay for an end-to-end voice call - this doesn't include any network delays, which
can be more than 100 ms and even greater when these technologies are used across the
public Internet or across international circuits. Since the public Internet is not
deterministic, you can never achieve toll-quality voice through it.
Carriers and service providers are starting to deploy private packet networks that are
optimized for carrying multimedia traffic, and in particular, voice. New media switches
are also being optimized to ensure less than 50-ms, end-to-end delay for voice
packetization and depacketization processes, including the use of dynamic de-jitter
buffers, programmable packetization intervals (the sampling of voice packets), and echo
cancellation technologies based on both G.165, and G.168 for tail-end echo.
Integrating with existing OSS platforms is also key for provisioning, billing, and
management (service, operations, engineering, costs, etc.) aspects. OSS operations are
important for the overall provisioning and maintenance of the end-user service - they're
also the vehicle for revenue collection. However, vendors will want to ensure that new
applications can be easily added to the carrier and service provider product portfolio,
through standard-based APIs.
Scalability is paramount, especially for carriers and service providers who want to
open new cities to services. The ability to start with a very economical entry point - 200
voice ports is a fair number - to supporting multiple thousands without a fork-lift
upgrade and new frame, or major equipment installs, is extremely important. The interface
scalability should reach from DS0s to higher-order interfaces including DS3s/E3s and
SONET/SDH interfaces on the access/trunking side, as well as high-performance
packet/broadband interfaces.
The other defining moment with new carrier-grade media switches is the collapse of
multiple services/functions within a single platform. The ability to introduce new
services without deploying additional equipment will give carriers and service providers
competitive advantages when developing new services for their customers.
CONCLUSION
It's a brave new world out there - the services that will, and can be offered, are
endless. They require that service providers look at solutions that will offer the best
migration strategy, while providing them with an open platform architecture for future
network and service requirements.
Randy Brumfield is director, product marketing for TransMedia Communications, Inc.
TransMedia is delivering a new-generation "media-aware" switch that gives
carriers a new edge - bridging the gap between high-performance Internet/packet (IP/ATM)
network and traditional public network communications (PSTN).
|