August 2004
Session
Controllers and the Complexities of Connecting VoIP
BY
DANIEL C. DEARING
Voice over IP (VoIP) is rapidly
gaining traction among competitive and incumbent carriers because of the
compelling economics of packet-based over circuit-based transmission. While
up to 90 percent of international voice traffic today is packetized for at
least part of the transmission, more carriers are offering packet voice for
local and in-country calling � and consumers are signing up for service.
However, for VoIP to meet the full potential of consumer demand, carriers
must solve critical interoperability issues that exist because of the way
VoIP implementations have evolved in the industry.
To understand these
interoperability challenges, one must review the ever-changing protocol
standards and their evolution to support more robust applications while also
providing operational efficiencies. Session Initiation Protocol (SIP), the
Internet Engineering Task Force (IETF) approach for VoIP, is centered on
making voice calling more Internet-like. The SIP protocol is modeled after
other Internet protocols, such as HTTP and SMTP, where flexibility and
simplicity are key attributes. The International Telecommunications
Union-Telecommunication Standard-ization Sector (ITU-T), on the other hand,
developed H.323 to ensure interoperability among systems that deliver
real-time communications over packet-based networks. Despite the fact that
H.323 and SIP address similar requirements, the mechanics of how they
perform call setup, media negotiation, and call tear down, make them
incompatible, preventing direct connectivity between SIP and H.323
endpoints.
Session controllers provide the intelligence to securely interconnect
carrier networks with other SIP and H.323-based networks. This capability,
SIP/H.323 interworking, is one of the important ways session controllers are
helping carriers simplify their networks and profitably introduce and expand
VoIP offerings and other real-time services over cost-effective IP
infrastructures.
H.323 Versus SIP
The H.323 and SIP protocols were developed with similar intentions, but with
different orientations and techniques � to enable real-time services across
packet networks.
H.323 came out of the voice-oriented, circuit-switched world. This protocol
was designed to support video conferencing among multiple IP networks; it is
the first version of the ITU specification, and dates back to the early
1990s. Because H.323 provided some key call-control and
gateway-administration functionality, carriers chose the protocol as IP
telephony gained market traction. Though interoperability issues created bad
buzz for H.323 even early on, there exists a significant base of users. Not
being able to afford being left behind as more enterprises moved to VoIP and
other IP-based services, incumbent carriers invested heavily in H.323.
Application developers and, especially softswitch vendors, derided H.323 as
overly cumbersome and complex, lumbering and memory-hogging � a desperate
creation of the Time Division Multiplexing (TDM) world that was ill-suited
for the emerging, all-IP future, especially in the case of VoIP. SIP arrived
late in the decade, spawned by the IETF and hailed as an enabler of not just
of IP telephony, but as a key underpinning of the entire converged,
application-oriented, next-generation Internet. The protocol defines methods
for initiating, modifying, and terminating interactive user sessions.
Uncovering the need for a new tool
The H.323/SIP eternal battle proved bad for business. Established carriers
that heavily invested in H.323-based platforms found it operationally
difficult and expensive to interconnect with new providers and enterprises
committed to the new SIP standard. Because various SIP-based systems were
engineered at different stages in the protocol�s definition, not all
SIP-based systems shared a common interpretation of the standard�s details
(in the areas of message format, timing, and sequence, for example). The
same can be said for H.323-based platforms. The ramifications have caused
problems with signaling interoperability, network address translation (NAT),
and firewall traversal.
TDM peering, with softswitches and media gateways, gave carriers an approach
to overcome the SIP/H.323 interoperability issues with media gateways linked
back-to-back, converting voice traffic from VoIP to TDM and then back to
VoIP at each carrier interconnect. This solution was a workable but
cost-prohibitive strategy, requiring significant operational and capital
expenditures. So, in the absence of real SIP/H.323 interworking, the
introduction and scaling of VoIP and other real-time service offerings
remained largely on hold.
Then came the session controller�
Normalizing at the edge, standardizing in the core
A Layer 5 technology, session controllers enable carriers to move to VoIP
peering � far simpler to engineer than Time Division Multiplexing (TDM)
peering and reliant on none of the expensive DSP resources. Small carriers
offering wholesale VoIP services were the early adopters of session
controllers.
Today, as session controllers have evolved, the entire carrier community is
making a fundamental shift in the infrastructure model for supporting VoIP
and other real-time, packet-based services. The industry�s best-of-breed
edge session controllers operate on three planes: signaling, media
processing (transcoding among various codecs), and media routing (NAT and
firewalling).
Session controllers operating in the network core provide end-to-end
signaling capability, acting as call control engines to provide intelligent
routing, and traffic management. Edge session controllers operate in points
of presence (PoPs) to handle interconnections with enterprise customers� IP
Public Branch Exchanges (IP PBXs) or carrier partners� IP networks. The
media processing and media routing functions of edge session controllers are
as important as signaling compatibility to carriers, because G.711, G.723,
and G.729 codecs approach media compression in different ways. The edge
session controller negotiates between incompatible codecs.
The softswitch, meanwhile, provides the mechanism for IP-enabling traffic
orienting from the circuit-TDM-based, Public Switched Telephone Network (PSTN).
Core session controllers treat the softswitch and its associated media
gateway as just another IP endpoint. In turn, ports on the expensive TDM
backbone switch and media gateways that were previously used for
VoIP-to-VoIP traffic can be harvested and cost-effectively re-assigned for
TDM-to-VoIP traffic.
Making the transition from TDM peering with softswitches to VoIP peering
with session controllers alone can cut a carrier�s costs by 50 to 80 percent
(depending on the capacity of the network interconnection), spurring
broad-scale rollout of more competitively priced services. With session
controllers delivering true SIP/H.323 interworking and media processing,
carriers are freed to take advantage of the capital and operational cost
efficiencies of standardizing their VoIP core technologies. Regardless of
which flavor of H.323 or SIP or media compression used in the carrier�s own
network core, the market�s leading edge session controllers enable
cost-effective interconnection with VocalTec, Clarent, Cisco or any other
standards-based IP network.
Moving to end-to-end real-time control
With session controllers shouldering the responsibilities for signaling
interoperability, network security, call admission control, and service
quality control at the network edge, carriers are now increasingly realizing
their benefits in core deployments.
The technology�s very granular and highly programmable routing capabilities
are especially valuable in the network operations center (NOC). Session
controllers can assess the multitude of peering points a carrier might have
with carrier partners and enterprise customers and leverage policies in
routing and managing authorized traffic (e.g., there might be a policy to
limit the amount of calls originating from a particular carrier partner via
several ingress points.) The session controller ensures that traffic is
routed to optimum peering points, boosting call-completion rates.
VoIP billing is another responsibility that is more frequently relying on
core session controllers. Demand for VoIP and other real-time, packet-based
services (video hosting and other multimedia opportunities that will
ultimately prove country-specific) is intensifying as prices drop and
availability widens. Continuing to seek new levels of network simplicity,
carriers are coming to regard session control as a multidimensional
technology delivering cost-effective control end-to-end across their IP
networks. As a result, core session control is rapidly increasing in
popularity.
The promising equation of Internet telephony � basing voice services on
less-expensive equipment that would equal lower prices and greater demand �
failed to add up to wide-scale rollout until recently. The incompatibilities
among SIP- and H.323-based VoIP systems comprised one of the primary
obstacles, but those complexities have been solved with session controllers
normalizing traffic at the network edge. Via VoIP peering, interconnection
is seamless, so carriers can cost-effectively expand infrastructures and
service footprints.
For the foreseeable future, there must be accommodation for both SIP and
H.323 in carrier IP networks. Session controllers protect space for both of
the protocols, through SIP/H.323 interworking.
Daniel C. Dearing is vice
president, marketing for NexTone Communications. For more information,
please visit the company online at
www.nextone.com
If you are interested in purchasing reprints of this article (in either
print or HTML format), please visit Reprint Management Services online at
www.reprintbuyer.com or contact a representative via e-mail at
[email protected] or by phone at
800-290-5460.
[
Return
To The August 2004 Table Of Contents ]
|