Voip danmark

A DSL represents arguably the most complex market development in mass market fixed line telecom since the deployment of PSTN networks. The economics, implementation, rollout, marketing, and customer services all present challenges that operator business models need to address. There are significant implications for the core networks involved in a broadband domain and the performance of a mass market broadband business and access network needs to be carefully monitored. Tele Danmark has taken one of the most aggressive leaps of any Western European operator into the DSL business, and presents an appropriate case study for any operator wanting to engage in the business. Telecommunications[R] International interviewed the executive responsible for Tele Danmark's fixed line and ADSL business, Dan Sorensson, in his office in Aarhus, Denmark to discuss the business development of the project.

Tele Danmark made early strides in a core network upgrade to handle IP traffic and the company started to create an IP network to handle dial up Internet traffic in 1997. "In 1999 we conducted ADSL trials, initially with a single vendor, explains Mr Sorensson. Tele Danmark embarked on this in a relatively low-key way, but even in the absence of a major marketing campaign the ADSL business "exploded" in the middle of 2000. "We went out for new vendors to handle the whole chain: ADSL, VoIP, ATM, and IP," he recalls, "and because of some vendor problems, we chose Ericsson as the second vendor in February 2001. We found that the flexibility of the Ericsson ADSL system was good. Ericsson ADSL has a cascading capability, so we could make a case by case deployment. The cascading capabilities of the Ericsson system to my knowledge were quite unrivalled."

The contract commits ADSL uptake within a given timeframe. "The Ericsson DSLAM technology is now being rolled out and we've built a concentration network to support it. We are looking for architectural benefits between DSL technologies between the two vendors," he says. "We have a very good take up for our own brand ADSL services of between 10000 and 12000 new customers per month, and that stretches the ability of networks to extremes. Currently our marketing department does not see a weakening of user demand."

There are particular market segments that the operator is targeting. There is an obvious need in the business market for broadband access. Within the private residential market there is a teleworking segment and we are developing an ADSL solution for that. The Danish government is also encouraging and funding teleworking for government employees to underline the trend.

Mr Sorensson emphasizes the need for an effective customer service programme as users coping with ADSL technology frequently have many questions generally concerned with the CPE. Tele Danmark has been administering a busy hot line that users can access. Tele Danmark also administers its own installations.

There is also a substantial wholesale market for ADSL in Denmark and Tele Danmark supplies both dry copper access to competitors in unbundled local loop and bit stream access with complete ADSL infrastructure for other service providers to use. In the marketplace, Tele Danmark also faces substantial competition in broadband Internet access from cable TV companies in the country.

But complications remain, particularly in marketing and pricing. ADSL pricing is a complicated issue both in terms of the competition and in terms of the tariffing regime - flat rate or metered access - used. "I think there is a general feeling in the industry that ADSL service pricing is too low. Perhaps the industry will have to eventually abandon the flat rate ideas [at least for ADSL]," he comments. He acknowledges, however, that if processes are in order for example with automatic provisioning throughout the system then the costs can be driven down." Cost elements come from a variety of directions. There is significant manual work involved in provisioning because the process is not yet fully automated. Network debugging can also be costly.

Implementation is another issue. "For ADSL in the contract, we have clauses about the vendor expertise at our disposal and what is required. In the case of Ericsson, for example, the vendor installed the first 2 PoPs by themselves. Our main preoccupation today is ensuring all parts of the system work together."

Rollout progresses. "We have a plan to grow out the architecture of the ADSL network and we are committed to a rollout that covers some 95% of Denmark's population by July 2002 with a 256kbps downstream service to subscribers although this is not a regulatory obligation imposed on us," Mr Sorenson points out. "We are in the early stages of rolling out fibre networks to enrich the subscriber copper network with the intention of placing access within 500 metres of customer premises."

"Strategically we can now say 'we are now building a very flexible ADSL network based on ATM transmission so what possibilities does that give us for repositioning the access network as a whole'. Now we are trying to ascertain what it can be used for in other areas such as VoDSL or even the flexible provisioning of new lines from the ADSL modem. Even at this stage we have a pretty good ADSL business. If you ask customers are you happy with this, the overwhelming answer is 'yes'. I think we will see a breakthrough in premises wireless networks in the next 18 months. It could be a big market."

Nevertheless, he sees ADSL itself as a moving target. "We have been reading about ADSL technology for nearly 10 years but it is still immature." Compared with telephony products he regards this as still immature in terms of scaling capability: "We can have perhaps 200 subscribers per DSLAM compared with 100,000 subscribers for a telephone exchange under one roof. I look forward to the day when we have DSLAMS supporting 10,000 customers. Volume means everything in the new types of network. I think the IP network is scaleable but the access network is not at least in the same way."

There are implications for the core network too. "We have to build both the ATM network up to the IP network and we also have to expand the IP network." This rearchitecting has significant implications in terms of capacity and network growth because the multiplexers must be added frequently. Further up the chain there are also load considerations on remote servers. It's a huge intake of traffic compared with what happens traditionally," he concludes. "We are building a lot of complexity on top of IP today," he says, pointing out that the IP network should be capable of being isochronous even with millions of customers in the same network. ADSL traffic up to the IP network will be carried on ATM supplied by Ericsson. "Ericsson has a product line which is excellent in my view," Mr Sorensson states.

Operators even with deep knowledge about customer behaviour and traffic loads can also be surprised. "The dynamic people of the Internet world sometimes forget to tell us that there will be heavy traffic loads because, for example, of a major media Internet-based event. The quality of the CPE is also crucial because even with long mean time between failure a mass rollout means significant failures with the attendant problems of reputation and service quality."

More changes will be necessary. "We have interesting challenges. We are working on a visible and clearly expressed migration path but we don't have it yet. We have some sketches of what it could be. I expect a roadmap for our next generation networks early in 2002 in Tele Danmark. The most important part of our strategy is to 'sit on' and focus on the access business because the next battle will be in terms of what customers install on their own premises. Most of the money lies in the access segment."

All of this will be pragmatically handled. "Our CAPEX compared to other operators is relatively low," Mr Sorensson comments. "I expect a movement towards more open platforms for incumbent operators in general which means you open up the value chain more clearly than we have done so far and this is of great interest internally. The network costs need to be visible to our organization."