The network conundrum for a Web services world

The era of static and predictable service traffic has been consigned to history. The era of dynamic and unpredictable services traffic is now. Carrier networks, still designed with the operational requirement for accurate prediction of service traffic, have only seen incremental improvements rather than step change. Network operators must now consider radical options to support future bandwidth demands.

During the last decade internet and web services have become the de facto medium for communications, entertainment, information, and work. Service growth has followed exponentially, fusing spectacularly to the extent that it is difficult, or indeed impossible, to differentiate the service delivered and/or to distinguish the role that the consumer is performing.

Yet despite these advancements in services, there has been no significant development or change to the primary optical switching and transport network since their introduction.

Networking mantra appears to have been to build the biggest optical pipes possible, at the lowest capital cost and fill them as full as possible. No serious consideration however has been given to the underlying services, resulting in ‘hub and spoke’ architectures that are too big and costly and that do not provide the direct connectivity between the service providers and where the services need to be.

The impact is significant increases in OPEX and TCO with low levels of asset utilisation. Network operators now need to fundamentally re-think their approach to meet future business demands.

Meanwhile, new market dynamics have created an environment where the pressures and demands on carrier networks and operations are growing exponentially and at a much faster rate than revenues causing profits to shrink dramatically.

Whilst innovation and investment has been made in IP routing and consumer devices, very little has changed in the underlying packet switching and optical fibre network that interconnect the IP based servers and routers to the IP enabled consumer devices.

Web services by their nature are dynamic and unpredictable and are increasingly including a rich set of video features – demanding ever increasing bandwidth per service and guaranteed quality of experience.

However, the dilemma is that current carrier networks are managed and provisioned based upon predicted future patterns of usage from historical data that has taken days, weeks, and months to collate and is then analysed as a basis to predict and formulate future network changes. It is well acknowledged that accurate forward CAPEX investment planning for the network based on historical data is an increasingly impossible task to achieve, whilst being an exceptionally slow and costly process.

In addition, carriers have been unable to align their business and operational models to keep pace with the market and consumer demands – and therefore revenue opportunities. This discontinuity has created an overly complex, static, and unresponsive network infrastructure that continues to be disconnected from the services value chain.

The largest fixed cost within the carrier business is typically the network infrastructure which has remained largely the same – hierarchical and asymmetrical from the core through the metro to the access, whilst new services are symmetrical in profile, and hosted from a variety of distributed rather than centralised sources.

Whilst service providers have strived to reduce operational costs through functional consolidation, process improvement and workforce reduction – fixed costs such as power and real estate remain very high and are directly correlated to the size and complexity of the network.

In the face of rapidly declining margins on traditional products and services, the service provider has been prevented from responding to the competitive market pressure through service innovation and monetisation of new product revenues.

This is not unsurprising as the current legacy network is hugely inflexible; it is unable to support introduction of new services and pricing models due to the lack of correlation of services carried to network resources consumed. This fact alone directly prevents the innovation and monetisation of new revenue models.

The legacy of tactical investment coupled with a lack of step change innovation by the traditional vendors has resulted in a network architecture that is a barrier to business. The network architecture is connection oriented – yet is being tasked to deliver connectionless traffic.

As a result, the current network technologies can never match the business economics to deliver the required level of service and associated lowest level of TCO – for any given service – resulting in a fundamental destabilisation of the business model.

Hub and spoke networks (by definition designed for connection-orientated traffic) are struggling because web services (connectionless protocols) have became the primary delivery vehicle for all services – including the rapid growth area of online video. Meanwhile, network operators also face the challenge that new demands for HD video need to be met with guaranteed bandwidth and low latency only afforded by connection-oriented networks.

Carriers now need to consider adopting transformational innovation versus continuing the deployment of repackaged hybrid products of today. The capability of this innovation must deliver automatic and dynamic response to any mix of incoming service flow traffic requirements – allocating network resources appropriately that are based upon the quality of service tiers required. To achieve this requires a fundamental change and reduction in the complexity of the network.

John Dunne co-founded Intune Networks in 1999 with Tom Farrell following EU-funded research work at University College Dublin. Since its foundation, John has been responsible for the commercial vision of how the now global networking company's core technology could be applied into the telecoms market. He is regularly invited to speak at conferences on innovation, network technology and the future trends in the communications industry. John has a first class honours electronics degree and a PhD in the field of tunable lasers from UCD and has spent his career working on the application of these devices into telecommunications systems. He is a co-inventor on some of Intune’s core intellectual property on network systems and is published internationally in the field of optoelectronics.