|Africa and the Middle East 2004
|Convergence in the Access Systems
|Marconi Middle East
Mr Andrew Furner, the Managing Director of Marconi Middle East, has over twenty years experience in telecommunications, information technology services, software and management consulting in a variety of senior management roles. Before assuming responsibility for Marconi’s operations in the Middle East, he served as head of Technology Strategy for Marconi Corporation. In this role, he led the high level team that defined and developed Marconi’s Next Generation Telecommunications Network vision and Marconi’s technology investment strategy.
Throughout the world, fixed line operators are seeing voice revenues decline and the voice networks are showing signs of obsolescence. In many markets, even the basic, fast, Internet service offered by DSL is not generating sufficient revenues. Operators are discovering a real business need to make convergence happen. Operators working towards convergence find they need fewer edge platforms and have reaped major operating expense reductions. The need to plan evolution to packet voice technology is compelling.
Convergence presents a new window of opportunity for telecommunication operators to explore as they seek greater business efficiencies and strive for new levels of growth and profitability. Throughout the world, fixed line operators are seeing their traditional voice revenues decline and their existing voice network investment is showing signs of obsolescence. Residential broadband over DSL (Digital Subscriber Line) is a big success in developed markets and its growth requires substantial investment. However, in many markets, the basic fast Internet service offered by DSL is not generating sufficient revenues on its own to justify network investment or to create a powerful business case. As a consequence, operators have two major objectives: To achieve a substantial reduction in ongoing network operational costs and obtain incremental capacity at a much lower cost-per-bit Find new service revenue opportunities. Since it is unlikely there will be any single ‘killer application,’ this means many new services must be deployed. Operators are at different stages in the process of emerging from corporate reconstruction and debt reduction programmes and many are in the process of privatisation and deregulation. Moreover, operators’ views differ on the appropriateness of basic network infrastructure. Nonetheless, growing numbers of them are beginning to consider new programmes of capital investment to reduce operating expenses and provide multiservice flexibility. These investment programmes will have major consequences both in network infrastructure and operating support systems (OSS) architecture. Operators are discovering a real business need to make convergence happen. Convergence provides operators with the opportunity to reduce the number of edge platforms deployed in their networks and to cut the number of service-specific overlay networks (‘stovepipes’). At the same time, operators need to find new revenues, to build on sales of basic fast Internet access service and create new profitable solutions for small businesses. The new services must be delivered with appropriate quality and will need usage-based charging. Many operators are now considering how to shape an investment programme to address these issues, whilst migrating highly profitable voice services onto their IP-based data network, as legacy equipment becomes obsolete. Multiservice access nodes (MSANs) As operators work to enable convergence, they see a reduction in the number of edge platforms, deployed in high volume in their networks, as a major opportunity for operating expense reduction. In addition to delivering new revenue generating services, they see the need to plan their evolution to packet voice even where no compelling short term reason for implementation exists. In this context, ‘first generation’ DSLAMs (Digital Subscriber Line Access Multiplexers) are being replaced by much more flexible multi-service access nodes (MSANs), capable of terminating a wide variety of access technologies. MSANs will continue to support the DSLAM functionality. They increasingly require terminating Ethernet bearers over copper, fibre and even fixed wireless and can also support Ethernet backhaul (transmission from a remote to a central site) on the network side. They will support ‘combo’ POTS/DSL (POTS – plain old telephone service) line cards as well as packet voice gateways, allowing a graceful evolution from old to new voice networks without on-site intervention. A key enabler for new service revenues is the ability to support ‘session-based services’ – that is, charging for content streams based on guaranteed levels of quality and usage. There will also be a need for a ‘media firewall’ that provides flow identification and policing functions for packet-based access to multiple services. The predominant trend today is the deployment of MSANs in copper wired networks together with existing switch remotes and DSLAMs. Many operators are also considering the possibility of electronic nodes on the street as an evolution bringing fibre nearer to the customer and allowing higher bandwidth (e.g. VDSL). This can be accommodated within an MSAN architecture using small remote line shelves and aggregator nodes. Finally, there is obviously scope for bringing together MSANs with edge transport functions. The likely outcome is an architecture allowing for ‘remote line shelves’ subtended from aggregation nodes. MPLS and ATM Most profitable data services today are carried over Asynchronous Transfer Mode (ATM) networks. Moreover, the most important new source of bandwidth in operator networks is residential broadband, which is aggregated using ATM. Hence, ‘rumours of the death of ATM have been greatly exaggerated.’ Most operators have substantial and profitable ATM networks which they will not abandon quickly and which they will require to evolve without service disruption. Whilst best-effort IP traffic has huge volume, most such traffic does not support profitable services. Operators recognise that obtaining such profitability requires service delivery with appropriate and dependable quality. In the context of real attempts at network convergence, including voice and video, differentiated Quality of Service (QoS) is completely essential. However, Quality of Service is not completely assured by Class of Service technologies which, although they can ensure that some packets are treated with priority, cannot ensure that there is sufficient network capacity to provide appropriate quality for the delivery of any particular service flow. This requires the reservation of resources and admission control. MPLS (Multiprotocol Label Switching) is intended to solve this problem by integrating a hierarchical ‘virtual circuit’ technology with IP forwarding that supports QoS. As currently standardised, MPLS does not fulfil that promise completely. Scalable QoS support requires either changes to existing gateway protocols or the adoption of a scalable QoS-aware alternatives, which addresses these problems. We believe these problems can be solved, but they are not solved yet. Current generation routers are not appropriate and probably cannot be extended to handle these functions. We think, therefore, that MPLS is the most likely medium-term solution to a converged core network, but that the necessary preconditions are not yet in place. When MPLS is deployed for that purpose, it will need to interwork effectively with the ATM installed base. Multimedia softswitches Packet voice technology has been “emerging” for a long time in public networks. Whilst there has been substantial deployment of IP PBXs (Internet protocol switchboard equipment), most operators have not made major investments in public network systems. There are several reasons for this: Existing voice systems are stable, dependable and do not require major new investment Basic fixed-voice demand is not growing. There has been growth, in particular because of dial-up Internet access, but this has been addressed where necessary by specialised bypass solutions and is now disappearing because of broadband substitution Whilst basic packet voice technology is not in doubt, there are still major concerns about quality, security and scalability for the public network and the evolution issues are complex. In our view, this situation is now changing. Operators are beginning to have real obsolescence concerns about their voice infrastructure and they recognise that any replacement strategy will take some time to deploy. Moreover, it is becoming clear that they cannot survive on their traditional revenues and therefore need a flexible service architecture. The challenge in designing a public network packet voice solution is to balance several issues: Internet-derived technologies, such as SIP, give great flexibility for rapid deployment of new services. Operators must be able to support new kinds of services and new media if they are to invest in a whole new service infrastructure A totally ‘transparent’ IP network gives operators no opportunity to provide appropriate quality and dependability for a public network offering and no way of charging for their services Commercial service depends on universal interoperability with existing fixed and mobile networks and hence must interface to current standards and feature sets and support current regulatory and government requirements. Once again, we believe that this technology is now becoming sufficiently mature to address these concerns. State of the art softswitch systems consist of a set of components connected together by standardised interfaces. They offer: Controlled access to resources and other networks based on ‘media firewalls’, which police and translate between multiple IP address spaces and also manage QoS for media streams on the boundaries of an operator’s IP domain A full range of gateway protocols Controlled access to SIP (Session Initiated Protocol – a signalling protocol for IP telephony, among others) servers and devices. Full interoperability with existing network features and protocols Flexible service intelligence Service interworking with 2.5G and 3G mobile networks using CAMEL & Parlay (applications interfaces for networks) High dependability and network wide fault tolerance Flexibility in size and location, allowing deployment at sites convenient to the operator. Convergence in services, applications and devices is largely driven by market demands and the need for operators to maintain their competitive position. Cost optimisation considerations as well as new service delivery will drive network convergence. Whatever the driving considerations, operators ignoring the benefits of converged network architectures may do so at their own peril. The same technology is available to all players so, if an operator fails to adopt next-generation network architectures, their revenues will eventually fall to an alternative service provider. This alternative service provider will offer new and legacy services at a lower cost than their ‘stovepiped’ competitors. The combination of multiservice flexibility and low operating expense is a combination that no network operator can afford to ignore.