A s Charles Darwin discovered, evolution is a process designed to ensure a smooth transition and long-term benefits. As networks become ubiquitous multimedia, multipurpose utilities, the move among service providers to a complete
IP Multimedia Subsystem (News - Alert)
(IMS) architecture in order to provide fixed and mobile multimedia services bears careful consideration. Some believe that evolving IMS standards have yet to adequately address all the features necessary for carrier-class operation and do not adequately support applications that don’t require
Session Initiation Protocol
(News - Alert)
(SIP) — such as IPTV, video on-demand (VoD), voice-over-IP (VoIP), Internet gaming, and more. IMS has come a long way in the last five years since its inception in the mobile standards groups and appears to be poised to become a viable platform for SIP-based multimedia services. Since that time, the approach has been adopted and adapted by the wireline standards organizations, such as the European Telecommunications Standards Institute (ETSI) and by CableLabs in its Packet Cable 2.0 draft specifications. Prudent service providers should, however, focus on developing networks that support both IMS and other services so they can offer the broadest range of services possible to their customers and to improve operational efficiencies as fixed and mobile networks converge.
The Prevalence of Non-SIP Applications
As a control layer for SIP-based applications, IMS was originally conceived before the emergence of VoIP by the Third Generation Partnership Project (3GPP) as an architecture that would allow mobile carriers to run data services alongside traditional mobile voice services. Today it is being promoted as the architecture of choice for mobile and fixed multimedia services of all kinds. Wireline and cable operators are also interested in IMS, and organizations such as the ETSI and CableLabs are creating standards to support both IMS and non-IMS applications. The International Telecommunications Union (ITU) through its various study groups are also focused on a more complete Next Generation Network (NGN) architecture that takes into consideration both IMS and non-IMS application support.
As an example, services such as bandwidth-on-demand and IP virtual private networks (VPNs) are being delivered today and are generating income for providers without IMS architectures. Other consumer services such as VoD don't necessarily require IMS either (as evidenced by the large number of non-IMS
IPTV (News - Alert)
trials going on today), but some vendors are investigating whether these services can also be delivered effectively in an IMS environment. Providers have a range of choices regarding if, how and how quickly they migrate to IMS. Therefore, the key is to build an architecture that provides flexible support for both traffic types and easy migration from one approach to the other over time. This is hard to do in traditional service provider networks today.
Ongoing Support of IMS and Non-IMS Applications
The various IMS architectures promise significant benefits, not the least of which is a common, open, industry-standard environment across vertical markets that equipment vendors can support and that allows service providers to more easily interconnect. By providing for the consolidation of service delivery platforms for voice and data, IMS creates one environment instead of multiple ones, thereby theoretically reducing the amount of IT and network integration work necessary to bring new services to market. This capability and others will give service providers entry into new markets and allow them to capture increased service revenue from existing home and enterprise network customers.
But IMS is not a one-size-fits-all solution. As previously noted, today VoIP, IPTV, online gaming, web access, email and other applications are non-SIP and therefore non-IMS. Service Providers see immediate revenue opportunities for these types of applications and so are left with a dilemma — should they wait until such applications are ‘ported’ to IMS/SIP or should they proceed immediately to enhance their revenues and protect their customer base? In the short term, most service providers will likely have to support both IMS and non-IMS services in order to offer these revenue-generating services today and other applications that customers will want. This flexibility in delivery of both IMS and non-IMS consumer and business services forms the basis for an IP next-generation network. Also, supporting both IMS and other applications is wise because there are a number of important network features not yet covered in the IMS specifications.
Verizon (News - Alert)
Wireless announced an initiative in conjunction with the company’s five largest equipment suppliers to add some of these missing features. The initiative aims to introduce more robust security services, packet-flow optimization for more efficient asset management, and the ability to manage both SIP and non-SIP applications. Verizon plans to submit its work as a set of enhancements to the 3GPP2 Multimedia Domain (MMD) architecture, which is related to the IMS standards. Leading network equipment vendors have also addressed operational and security gaps in converged multimedia network design in the next-generation network architecture and technologies.
Security Concerns with IMS-Only Environment
As a network overlay, IMS does not rely on underlying IP transport technologies for security. Instead, it relies on SIP for access and infrastructure signaling. Instead of addressing security on an end-to-end network basis, the IMS standards have basically left security to a link-by-link and layer-by-layer approach. Posture assessment — the ability to admit to the network only those endpoints compliant with the software policy — is not provided. Integration of intrusion detection, real-time attack response, and other security and control layer intelligence capabilities are also not currently addressed in IMS standards.
Supporting Both SIP and Non-SIP Applications
The current IMS standards outline a methodology for the control of SIP-based services only, although some standards bodies are well on their way toward expanding IMS, as previously mentioned. A more flexible environment would allow for support of non-SIP applications as well. Such a solutions environment should consolidate features such as accounting, authorization, authentication, policy controls, service control, application interaction, roaming, peering, quality of service, and management for both types of multimedia applications. These services should be adaptable enough to support SIP and non-SIP applications from a third-party provider and blends of SIP and non-SIP applications.
Enabling Packet Flow Optimization and Session Border Control
Another important feature not yet considered in IMS is packet-flow optimization, a technology solution for the intelligent allocation of network resources for application flows to maximize return on investment for service providers. This technology lets service providers inspect and classify network packets at session- and application-level support in order to gain visibility into their network activities, optimize application traffic to reduce operational costs, and improve network performance, as well as identify and mitigate malicious activities. Packet-flow optimization also assists service providers in managing third-party applications and provides quality of service for those applications that do not provide direct policy control interfaces.
Defined in recent ETSI TISPAN specifications, the session border controller — a device used in VoIP networks to control signaling and media streams related to setting up, conducting and taking down calls — is another key component of efficient converged multimedia networks.
Choosing the Right Protocol for the Right Job
With many service providers spending significant amounts to build IMS control layers for SIP-based applications, the temptation is to reuse the framework for other types of applications as well. Many multimedia applications involve sessions, and SIP can be used to support some of these applications either natively or encapsulated. But native sessions can require substantial changes to SIP, and the encapsulated solution can increase the amount of messaging over access links and slow the network down. The best approach is to choose the right protocol for the right job.
Leading network equipment suppliers support the emergence of IMS. Some have chosen to support both IMS and non-IMS applications, providing the greatest efficiencies and choice for customers. An ideal solution would allow service providers of all types to control customer access and use of services without limiting the kinds of applications that can be deployed.
Access-independent and open approaches support today’s evolving IMS standards and help network operators achieve better visibility and control of networks. These new “service control frameworks” provide enhanced network and service intelligence and accommodate SIP and non-SIP applications and both IMS and non-IMS environments. The service control layers in these next-generation network frameworks deliver new levels of subscriber and application awareness to let service providers know who their customers are, where they are, how they are using their authorized services, and when policies that govern that use are applied. With support for IMS and greater granular visibility and control, these frameworks enable service providers to deliver differentiated applications more securely and profitably today, for the best services that network technology and architectures can offer. This gives service providers the greatest degree of choice and flexibility in their service offerings, letting them compete and generate revenue by creating new services rapidly, with better control, while reducing costs through converged network efficiencies.
Peter Clarke is Director of SP Strategy in the Service Provider Routing Group at
Cisco Systems (News - Alert). For more information, please visit the company online at www.cisco.com.