Singapore's NGN proposals process: a legion of devilish details
David Kennedy
24/12/2007

The Singapore government has kicked off its tender for a national next-generation network (NGN) by imposing a tough operational separation regime on the project. 'Passive' access infrastructure (the 'netco') will be separated from the 'active' layer 2/3 switching and routing functions (the 'opco').

The government's briefing makes it clear that this separation will be strict, possibly even involving separate ownership. Broadband retailers would also be required to be separate from the opco, but these requirements would be less strict. The tender for the netco will be launched first, and will provide up to S$750 million of funding. Netco proposals are due in March 2008, while opco proposals are due in 3Q08.

It was always understood that wholesale and retail operations would be operationally separate on Singapore's NGN, but these additional requirements raise important issues for potential tenderers.

If 'passive' infrastructure is limited to fibre/copper cables, conduits, ducts, poles, splice boxes, etc., then it would be workable to separate this off into a separate company. In fact, something like this has been done by both NTT and the Amsterdam metropolitan NGN network. In these cases, only the fibre and ancillary infrastructure are held by the netco, and access to dark fibre is provided to an opco.

In contrast, separating the active part of transmission infrastructure (i.e. SDH/WDM, optical switches, and so on) from layer 2/3 data switching would make little sense given the convergence underway between these two areas. This convergence is not just hype, but is impacting products today, and is affecting other things such as R&D organisation and the structure of telco network divisions.

It is therefore most likely that Singapore's netco will only be responsible for building and providing access to dark fibre and ancillary infrastructure. In contrast, the opco will have full responsibility for lighting the fibre.

What has motivated this move? The documents we have seen do not make this entirely clear, but there are hints. The government seems to envisage multiple opcos; i.e. multiple providers of layer 2/3 functions. These functions involve more than just directing traffic. They involve all sorts of things that directly impact the quality of services, such as monitoring SLAs, throttling down peer-to-peer traffic and so on. Control over these would make it possible for multiple opcos to differentiate their wholesale broadband offerings.

Naturally, this requirement will rule out certain network technologies and architectures. Netco tenderers will need to devise proposals that complement the requirements of the opco, and vice versa.

Apart from the technology implications, this will also affect the economics of the network. One of the consequences of the shift to NGN technologies is that the economies of scale and scope in the network are also shifting. The Singapore government seems to have concluded that its separation model does not disrupt any significant economies of scope between access and switching. Whether this is the case will depend on whether the line between opco and netco is drawn correctly.

The government's objective is to promote quality-based competition between multiple opcos in the provision of broadband services, and this is a worthwhile objective. In contrast, it is clear that the netco will remain a monopoly. Without some regulatory intervention, the netco will be inclined to act like a typical monopolist with control of a network resource. Some form of ongoing access regulation will be necessary, and this requires clarification before March 2008 when netco proposals are due. The final regulatory arrangements will probably bear a fair resemblance to the typical regulatory regime for unbundled local loop (ULL)-based broadband: regulated access to the access hardware monopoly (ULL/fibre) and a competitive market for broadband services.

- David Kennedy is a research director at Ovum responsible for broadband and wireline research.

advertisement