Archive for August 2012
Some readers may have noted that Ericsson sold its fibre access product division to Calix, a loss-making North American manufacturer that provides “last mile” active components that connect communications services providers (CSPs) and their customers.
The move will cost Ericsson SEK400m ($61m) and 60 staff. It didn’t say how much it got for the division, but as Calix has not made a profit in five years, it couldn’t have afforded much. That might tell us just how badly Ericsson (and its competitors) want to get out of the fixed line access market. ITU figures show that it’s a market in decline, and besides, there are six times more mobile subscribers than fixed line subs.
Calix, in its most recent SEC annual filing, spells out the problems and risks fixed line CSPs face in graphic detail. I reproduce it below in its entirety, subject only to anglicising the spellings. It should be required reading for anyone about to spend taxpayers’ money upgrading their local broadband network.
CSPs compete in a rapidly changing market to deliver a range of voice, data and video services to their residential and business subscribers. CSPs include wireline and wireless service providers, cable multiple system operators, or MSOs, electrical cooperatives, and municipalities. The rise in internet-enabled communications has created an environment in which CSPs are competing to deliver voice, data and video offerings to their subscribers across fixed and mobile networks. Residential and business subscribers now have the opportunity to purchase an array of services such as basic voice and data as well as advanced broadband services such as high-speed internet, IPTV, mobile broadband, high-definition video and online gaming from a variety of CSPs.
The rapid growth in new services is generating increased network traffic. For example, Cisco Systems estimates that global IP traffic will grow at a compound annual growth rate of 32% per year from 2010 to reach approximately 80.5 exabytes per month in 2015. We believe that increased network traffic will be largely driven by video, which is expected to account for over 90% of global consumer traffic by 2015.
CSPs are also broadening their offerings of bandwidth-intensive advanced broadband services, while maintaining support for their widely utilised basic voice and data services. CSPs are being driven to evolve their access networks to enable cost-effective delivery of a broad range of services demanded by their subscribers. With strong subscriber demand for low latency and bandwidth-intensive applications, CSPs are seeking to offer new services, realise new revenue streams, build out new infrastructure and differentiate themselves from their competitors.
CSPs typically compete on their cost to acquire and retain subscribers, the quality of their service offerings and the cost to deploy and operate their networks. In the past, CSPs offered different solutions delivered over distinct networks designed for specific services and were generally not in direct competition. For example, traditional wireline service providers provided voice services whereas cable MSOs delivered cable television services. Currently, CSPs are increasingly offering services that leverage internet protocol, or IP, thereby enabling CSPs of all types to offer a comprehensive bundle of IP-based voice, data and video services to their subscribers. This has increased the level of competition among CSPs as wireline and wireless service providers, cable MSOs and other CSPs can all compete for the same residential and business subscribers using similar types of IP-based services.
Access networks are critical and strategic to CSPs and policymakers
Access networks, also known as the local loop or last mile, directly and physically connect the residential or business subscriber to the CSP’s central office or similar facilities. The access network is critical for service delivery as it governs the bandwidth capacity, service quality available to subscribers and ultimately the services CSPs can provide to subscribers. Providing differentiated, high-speed, high quality connectivity has become increasingly critical for CSPs to retain and expand their subscriber base and to launch new services.
Typically, subscribers consider service breadth, price, ease of use and technical support as key factors in the decision to purchase services from a CSP. As CSPs face increasing pressure to retain their basic voice and data customers in response to cable MSOs offering voice, data and video services, it is critical for CSPs to continue to invest in and upgrade their access networks in order to maintain a compelling service offering, drive new revenue opportunities and maintain and grow their subscriber base.
Access networks can meaningfully affect the ongoing success of CSPs. Governments around the world recognise the importance of expanding broadband networks and delivering advanced broadband services to more people and businesses. For example, in February 2009, the US government passed the American Recovery and Reinvestment Act, or ARRA, which set aside approximately $7.2bn as Broadband Stimulus funds for widening the reach of broadband access across the United States, a portion of which includes broadband access equipment. These funds, distributed in the form of grants, loans and loan guarantees, primarily target wireline and wireless service providers operating in rural, unserved and underserved areas in the United States.
Many CSPs have actively pursued stimulus funds and have submitted various proposals to receive assistance for their broadband access infrastructure projects. Awards for these projects have been issued between December 2009 and September 2010.
The timetable for completion of funded projects varies between the two agencies administering the awards. Projects funded under the Broadband Technology Opportunities Program (BTOP), which is administered by the National Telecommunications and Information Administration (NTIA), must be completed by September 30, 2013. Projects funded under the Broadband Initiatives Program (BIP), which is administered by the Rural Utilities Service, must be completed by June 30, 2015.
Limitations of traditional access networks
CSPs rely on the capabilities and quality of their access networks to sustain their business and relationships with their subscribers. In the past, subscribers had little influence over the types of services provided by CSPs. Today, subscribers can be more selective among CSPs and they are increasingly demanding advanced broadband services in addition to basic voice and data services.
In general, access networks are highly capital intensive and CSPs have historically upgraded capacity as technology and subscriber demands on their networks changed. CSPs will increasingly integrate fibre-and Ethernet-based access networks to enable the delivery of more advanced broadband services at a lower cost while at the same time enabling the continued delivery of basic voice and data services.
Thus far CSPs have taken an incremental approach to capacity upgrades in their access networks. As a result CSPs face multiple challenges concerning their access networks, business models and service delivery capabilities, including:
A complex patchwork of networks and technologies—In order to upgrade their access networks CSPs have typically added networks for new residential or business services that they deliver, such as digital subscriber line, or DSL, data over cable service interface specification, or DOCSIS, GPON or Gigabit Ethernet on top of existing networks. This led to an overbuild of access technologies and an unnecessarily complex patchwork of physical connections between the central office and the subscriber.
In addition, CSPs have generally begun to expand the penetration of fibre into their access networks, thereby shortening the length of the subscriber connection through other lower bandwidth media types (such as copper-based or coaxial cable-based networks). CSPs have also attempted to evolve their access networks to enable more efficient packet-based services by adding Ethernet protocols on top of existing asynchronous transfer mode, or ATM, and DSL protocols.
In addition, CSPs have often deployed separate equipment to facilitate the delivery of Synchronous Optical Networking, or SONET, Gigabit Ethernet and 10 Gigabit Ethernet transport which connects CSP central offices with their access networks, further increasing the complexity and the cost of their networks. This approach has left most CSPs with disparate architectures, features, functions and capabilities in different parts of their networks.
This increasingly complex, patchwork approach to deploying access networks and delivering new services to their subscribers has created potential complications for CSPs within their access networks. These potential complications limit data transmission capability, increase the cost of operation and maintenance and can negatively impact the subscriber experience.
Limited capacity from legacy access architectures—Legacy access network architectures were designed to address earlier generation communication demands of wireline telephone, cable television and cellular services. Such access networks have physical limitations in their ability to scale bandwidth, avoid latency issues and deliver advanced broadband services, which subscribers demand today and are expected to increasingly demand in the future.
In addition, CSPs understand the need to add fibre to their networks to provide the bandwidth required to scale advanced broadband services. However, it is costly and complex to integrate fibre-based technologies into legacy access networks.
Inflexible technologies increase network switching costs—Legacy access networks were built around a narrow set of technologies. For example, traditional voice calls use circuit switching technology to allocate a fixed amount of network capacity to each call, regardless of whether such capacity is fully utilised.
The emergence of packet-based technologies, primarily IP and Ethernet, has significantly improved the ability to transmit data efficiently across networks as bandwidth is only consumed when signals are actually being transmitted. Most legacy access networks do not allow circuit- and packet-based technologies to co-exist or to evolve from one technology to another.
Inefficient service roll-out constrains subscriber offerings—Legacy access networks were designed to support a narrow range of services and as a result, they limit the ability of CSPs to provision the advanced broadband services increasingly demanded by their subscribers.
Packet-based networks are more flexible and efficient than traditional circuit-switched networks. For example, to provision additional business services in a legacy access network, a CSP would typically deploy additional physical connections and equipment, whereas packet-based infrastructure allows a CSP to change or add services virtually, without the presence of a service technician or the installation of new equipment.
In order to deploy these services quickly and efficiently, CSPs must be able to utilise their existing infrastructure while upgrading the legacy access network to packet-based technologies.
Highly reliable access products are difficult to engineer and manage—Given the critical nature of access networks and their typical deployment in remote and distant locations, access infrastructure products must be highly reliable. Unlike most other communications equipment which is deployed in environmentally controlled central offices or similar facilities, most access equipment is deployed in outdoor environments and must be specifically engineered to operate in variable and often extremely harsh conditions, as well as fit into smaller spaces, such as on a street corner, near office buildings or on the side of a house or cellular tower.
Since the access portion of the network is broadly distributed, it is expensive as well as difficult to manage and maintain. CSPs require access network equipment that can perform reliably in these uncontrolled environments and be deployed in a variety of form factors, thereby adding significant engineering and product development challenges as compared to most other forms of communications infrastructure equipment. In addition, some portion of the access market is supported by government initiatives and products sold into this segment require additional government certifications and approvals in order to qualify for deployment.
Expensive to deploy and operate—As a result of deploying multiple networks with discrete functions, legacy access networks require a wide variety of equipment to be installed, maintained and ultimately replaced, thereby placing a significant and recurring capital and operating expense burden on the CSP. Once installed, this equipment occupies valuable space inside a central office, requires frequent labour-intensive maintenance and consumes meaningful amounts of power.
Moreover, the lack of integration across protocols and fibre- and copper-based network architectures negatively impacts network performance. Inferior network performance diminishes the subscriber experience and increases network operating costs by increasing service calls, the number of required support staff and the frequency of equipment upgrades and replacements.
As broadband network availability and quality are becoming more critical to subscribers, lack of network reliability can be materially disruptive, expensive and ultimately increase subscriber churn, thereby negatively impacting the CSP’s business.
Given these limitations of legacy access networks, CSPs will increasingly emphasise fibre- and Ethernet-based technologies in their access networks, thereby enabling the rapid, cost-effective deployment of advanced broadband services. Such technologies reduce overhead expenses, simplify network architectures and seamlessly integrate legacy and next-generation networks. We therefore believe that successful CSPs will be those that evolve from providing basic subscriber connectivity to providing the most relevant services and subscriber experience.
Freshly triumphant from a successful Olympics (with more to come from the Paralympics) culture secretary Jeremy Hunt has been outlining his ambition to have not only the “best broadband in Europe” but also the fastest.
In a speech this week week he said, “We simply will not have a competitive broadband network unless we recognise the massive growth in demand for higher and higher speeds.”
He added government is considering how to allocate the £300m from the BBC licence fee earmarked for broadband investment. “In particular we will look at whether we can tap into to this to allow those able to access superfast broadband to be even greater than our current 90% aspiration,” he said.
Is that good news for rural communities? Not necessarily.
Hunt quoted Ofcom’s recent finding that average UK broadband download speeds had hit 9.1Mbps. However, the small print in the Ofcom release said the SamKnows study “includes only ADSL customers within 5km of the exchange and in geographic Markets 2 and 3 and in the Kingston-upon-Hull area”.
So the 9.1Mbps does not include Market 1, roughly two-thirds of geographic UK where BT is the sole supplier. Had it done so, the average download speed would be more modest.
Hunt is clearly hoping that the mobile network operators will fill in the gaps. The government has allowed Vodafone to pick up Cable&Wireless Worldwide’s national fibre network, essential for backhaul. Ofcom’s decision to allow mobile operator Everything Everywhere (EE, aka T-Mobile UK and Orange UK) to refarm its 1800Mhz spectrum for LTE will help extend high speed access. This assumes EE and Three, which is buying the chunk of spectrum EE is forced to sell to comply with its merger conditions, can lay their hands on affordable consumer connection devices. Barring legal challenges, this may come ahead of the auction of 800MHz and 2.6GHz bands sometime net year, which stipulates in-house coverage of 98% of UK homes.
Once that is done, the mobile operators can apply for £100m to reach into rural areas, and compete with fixed operators for a further £150m to upgrade broadband speeds in selected cities.
With Hunt confident that DG Competition will clear the way to release £530m in state aid for county broadband projects, it is hard not feel buoyed by the prospect of so much money pumped into networks.
However, there is a danger that the impending investment will deepen the existing digital divide between town and country communities.
Hunt will do well to make sure that every penny of the £530m is spent outside the suburbs, so too the £100m for mobile coverage, and the BBC’s £300m. If he could find a way for communities like B4RN to invest their own money with confidence, he could unlock more millions, and ensure Britain gets the broadband network it needs. That would be a legacy no-one could take away from him.
It is interesting to compare the Australian approach to a national broadband network with the UK’s apparently strategy-free approach, if only because the projected costs are within spitting distance of each other, give or take a few billion pounds.
Analysys Mason’s original estimate for a national point to point fibre to the premises (FTTP) network for the entire UK was £28.8bn; for a GPON-based FTTP service, £24.6bn; for a GPON FTTP network with a 93% population coverage rate, about £19bn.
Following the release of the business plan for the Australian government-owned NBN Company this week, NBN now plans to spend AU$37.4bn (£25.2bn) to build a wholesale-only, open-access, fixed-line broadband network over 10 years, ending in June 2021. This is six months later and 3.9% more expensive than originally planned.
NBN will deliver a GPON-based FTTP service to 93% of the population or some 12 million of 13 million-odd homes, flats and offices. Fixed wireless will cover 4% of the rest and two satellites the last 3%.
NBN will buy the voice and data traffic from the former incumbent Telstra and second carrier Optus, and rent Telstra’s passive infrastructure.
The taxpayers’ contribution via the government’s equity stake in NBN will be AU$30.4bn (£20.5bn).
In contrast, the UK is relying on the private sector (i.e. BT and Virgin Media) to provide a GPON fibre to the cabinet (FTTC) service to two-thirds of the UK’s 26 million homes, and is offering £530m in taxpayers money, plus “matched funding” from local county councils and some funding from various European Union development budgets.
Further comparative figures emerged in a Linked In discussion group started by Nguyen Duc. As Duc notes, geo-demographic factors determine costs.
The UK has 62.7 million people in 244,000 square kms, with 80% living in urban areas. Almost 90% of Australians live in a necklace of towns and cities on the coast, leaving the vast majority of the continent’s 7.7 million square kms to goannas and kangaroos. As a result, the UK’s population density of 257 plays Australia’s 2.9 persons per square km.
The economies of scale available for servicing a population density almost 100 times higher clearly work. On a per household basis the bill for a 100% P2P FTTP UK would work out at about £112 per household, the Australian are paying £2,100, which is one reason the NBN is controversial.
However, NBN is finding cheaper ways to connect homes, and the communications services providers are getting behind it.
NBN reports it has resale deals with more than 40 telephone and ISPs who represent 94% of the fixed broadband market. Similar deals are in the pipeline for the satellite and fixed wireless services.
Consumers can choose between more than 500 NBN retail price plans. Fibre prices compare “favourably to existing ADSL packages (but with the potential for higher speeds and greater capacity over the NBN)”, the firm claims.
The increase in data traffic, up 80% in a year on both fibre and copper networks, is way beyond what was assumed originally.
NBN says if this trend continues, NBN will be able to reduce wholesale prices faster than originally anticipated.
It has already committed to keeping the wholesale price for key products fixed for the next five years, and to keep later increases below the rate of inflation.
“Should actual data usage growth exceed the assumed growth rate, NBN would be able to reduce wholesale broadband prices more quickly and still maintain the same overall return (7.1%) for Australian taxpayers,” says CEO Mike Quigley.
The Australians will pay more for their network, amortise it over a smaller population, expect to reduce wholesale prices, and still expect a decent return on their money. The UK would have paid £6bn less, had a bigger market, seen at most a 12 year payback, and still chickened out.
Edward Phelps, secretary general of the new Digital Policy Alliance, gives his first on the record interview following the DPA’s formation in the wake of the House of Lords report that condemned the government’s broadband plan for lacking a strategy.
The government has said the BDUK process will go ahead. Isn’t the DPA initiative too late to affect the contract awards and disbursements that are now pending?
Phelps: The DPA will continue to ensure decision makers are made aware of the problems and possible solutions. Broadband solutions will remain critical long after BDUK money is spent. Currently few MPs or local authority decision makers understand the issues involved, if they did they would probably be more likely to put pressure on government to support innovative solutions.
What can the DPA say and do that will persuade the government to adopt HoL report recommendations such as dark fibre, commercially realistic open access, and community hubs?
The DPA does not necessarily support all of the HoL report recommendations but will be working on a set of recommendations to government based on industry consensus.
How can the DPA persuade Ofcom to change its apparently tolerant treatment of BT on matters such as PIA?
DPA will not advocate one technology or the business model of any one company. But we can increase awareness of the possible solutions and ask questions as to why they are not being used as a means of putting pressure on agencies such as Ofcom.
How can the DPA speed up regulatory processes such as the LLU costs issue now before the Competition Commission?
If this is a particular issue that our working group feels we should address we will ensure the relevant people and organisations are briefed accordingly and put pressure on the government.
Will the DPA take on issues such as the “fibre tax”, which leads to price distortions in the market, not least in comparison with Europe (except Ireland)?
This will be an issue for the working group to decide in the coming months.
The government’s failure to open up the market to innovative competitive solutions to providing easily accessible, affordable high speed broadband has prompted the formation of a new lobbying group, the Digital Policy Alliance, in the wake of concerns raised by peers on Tuesday.
Formerly known as Eurim, the DPA provides neutral ground where leading firms including Google, Microsoft, IBM and Cisco can discuss issues of national interest with politicians, civil servants, regulators, customers and NGOs.
The new alliance says it will press for more action from government on issues of technology uptake and broadband, which it says are essential for kick-starting the economy, as well as improving services like schools and hospitals, the DPA said in a statement.
“Industry has expressed anger that the government’s broadband strategy has failed,” it said. It is now working to provide expert solutions to broadband roll-out through the sharing of existing local authority infrastructure.
It is likely it will include attempts by the Conservative Technology Forum to enlist the energy industry’s attention via the estimated £4bn network needed to implement the national smart meter project that will see a communication link into every home and office building in the country by 2017.
Lord Erroll, a peer with a lifetime’s experience in the technology sector will chair the group. He will be helped by former Downing Street advisor Sean Worth, with former Eurim director Edward Phelps becoming secretary general.
Phelps said, “Government urgently needs to review its strategy for broadband roll-out. Government has failed to inject competition into the broadband market, leaving many innovative providers unable to offer their services. The current approach would appear completely at odds with the localism agenda.”
Separately, would-be bidders who have so far been frozen out of the BDUK broadband procurement framework are discussing how to cooperate to mount a challenge to BT for county council broadband projects.
The DPA’s new website is www.dpalliance.org.uk. It will be fleshed out in the coming weeks.