Cellular Backhaul Opportunities Beckon for Cable MSOs

Print This Page

By Juan Pablo Conti

Juan Pablo Conti, the former editor of UK magazines The IET Communications Engineer and Cable & Satellite Communications International (CSCI), is an independent technology journalist based in Buenos Aires, Argentina

Mobile phone operators around the world are waking up to the realization that, after several years of patient waiting for the uptake of their 3G services, the success they're finally starting to enjoy could turn out to be their own worst enemy.

The singular threat they face is rooted in a combination of technological and economic causes. Technologically, the migration from 2G to 3G and now 3.5G data-intensive services means the requirement for backhaul traffic between base stations and switching centers is set to double in a few years.

Considering that backhaul costs already account for anything up to 30 percent of operating expenses (OPEX) in a cellular network, even a slight increase in demand - let alone a 100 percent increase - could seriously jeopardize the whole business case.

To make matters worse, competitive pressure is forcing cellular companies to offer these bandwidth-hungry data services on a flat-rate basis.

Historically, mobile phone providers have turned to local exchange carriers (LECs) for their backhaul needs. This was initially a very sensible thing to do: Not only were LECs the only type of network providers that had the expertise and technological tools to handle voice communications in the late 1980s and early 1990s, but they were also the only companies with the appropriate infrastructure in the ground to link up cell towers and central offices.

When the time came to address the backhaul element of initial 3G deployments in the early 2000s, cellular carriers went with their tried-and-trusted telco approach, i.e. using the same type of TDM-based T1 and E1 leased lines for connectivity. Now, though, as they start to plan for a future where both the voice and data portions of their offerings are unified in a single, packet-based, pure-IP network environment, Ethernet is clearly emerging as the de facto protocol for wireless backhaul.

This is leaving the door wide open for cable operators worldwide to enter a market that will generate $4 billion in 2008 and more than twice as much in 2011, according to research firm Heavy Reading. "Both voice and data traffic [in cellular networks] continue to grow, but the data traffic is growing more rapidly," notes Bill Dawson, VP of business development and product strategy with equipment vendor Arris. "So, ultimately, the demand for new backhaul capacity is going to be dominated by Ethernet kinds of connections - and many cable operators are very comfortable offering that."

Of course economics will play a major role.

One operator who is very comfortable offering that is Virgin Media, a nationwide cable operator in the UK. Boasting a 186,000km long network with a 20 Gbps MPLS IP backbone and over 300 Ethernet nodes, the operator is gearing up to grab a share of the country's cellular backhaul pie.

"In the UK there are approximately 30,000 base stations, with an average of maybe up to 10 base stations being serviced by a particular node," explains Julien St.John-Dennis, head of products at ntl:Telewest Business, Virgin Media's business arm. "We see the opportunity to provide backhaul services across our resilient network all the way back to original switching centers for the mobile operators. There'll be potentially thousands of Ethernet pipes."

St.John-Dennis says the cable company sees the cellular backhaul market as a natural progression for its Ethernet portfolio. Indeed, apart from Virgin Media's five million residential customers, ntl:Telewest Business has 100,000 business customers. Crucially for the operator's credentials as it targets this new market segment, among those business customers are some of the UK's biggest names in the communications service provider space.

"We've been using our national Ethernet products in the local loop unbundling (LLU) backhaul market in the UK, providing backhaul services for the likes of Cable & Wireless, Tiscali or The Carphone Warehouse," says St.John-Dennis. "We're offering a high-bandwidth, managed aggregation service for those DSL providers.

"When you look at it, this is a very similar challenge to the challenge that mobile operators face. So, from a products strategy point of view, we saw it bang on our roadmap to go into that market space." Although the MSO is yet to sign its first commercial deal with a cellular operator, St.John-Dennis is confident 2008 is the year when that will change: "We've tested equipment in our own labs, in some of the mobile operators' labs (unfortunately I can't disclose whose) and we are responding to a tender document that mobile operators are putting out at the moment," he reveals.

Virgin is not the only MSO setting its sights on these apparently promising new revenue-generating streams. "Without naming names, I know that a number of operators have contracts to deploy and we can foresee significant opportunities," says Dawson. "It's actually a surprise to me that this is only starting recently, that this is not something the cable operators have been doing already for many years.

"I don't really know why it didn't happen sooner. Because the cable operators have fiber deployed so deep… typically, when you talk about nodes being built to pockets of somewhere between 500 and 1000 homes, that means there'll be fiber end available fairly close to wherever there might be a cell site, as long as it's in a residential neighborhood."

The Fiber Dilemma
The distance between where the fiber network terminates and where the targeted mobile phone base stations are is a key point of consideration. While some argue that the coax part of an HFC plant could have a role to play, the widespread view among network architecture experts is that fiber is a safer bet and will be the preferred choice for operators.

Eli Baruch, senior director of product management with Arris, has little doubt about it: "I think, predominantly, it will be fiber," he reckons. "Fiber is the architecture that both the cable and the cellular operator feel more comfortable with. It also enables you to do easier redundancy."

If ntl:Telewest Business' plans in the UK are anything to go by, operators seem to be prepared to extend their fiber reach, at least in those cases where short distances are involved. "We feel that a high percentage of base stations are within 100 meters of our network, so we'll go with fiber," says St.John-Dennis.

Another MSO agrees that a fiber network is required for wireless backhaul. "It will probably be a similar network that will serve medium-to-large businesses in some countries. The key is does the cost of building and operating an overlay fiber/Ethernet network to cell sites justify the revenue potential," an executive with the company said. "It also must be realized that in city areas most cell sites are on the roofs of buildings and getting from the basement of a building to the roof is not trivial or cheap to do. In suburban and rural areas you would be surprised how many cell towers are off the beaten path."

The keys for successful backhaul for mobile operators, one opined:

• Ethernet services over fiber that start at 10 MB/s and can easily and quickly grow to 100 Mb/s
• Symmetrical services
• Low latency & QOS guarantees
• Redundant feeds & reliability (carrier grade)
• Ability to groom T1s/E1s for voice network onto the same fiber service

Still, with a price tag of $10,000 to $15,000 a mile for the construction of new aerial fiber - and many times that for the type of underground construction that is mandatory in Europe - some operators will understandably prefer to support their wireless colleagues with their coax infrastructure. It is expected that DOCSIS® 3.0 and its channel bonding techniques may be a valuable tool in many cases.

As for the exact amount of bandwidth that will be required, there is some consensus that the average cell tower will see its demand grow from the approximately 4.5 Mbps (or three T1 lines) it is currently using to around 10 Mbps by 2011. This means the average hub site in Europe will need no less than 100 Mbps.

Beyond that three-year timeframe, nobody quite knows how cellular backhaul requirements will evolve. "As a service provider we're unsure of that," St.John-Dennis admits, "and the mobile operators are a little bit unsure as well. It depends on how their own roadmaps are for rolling out the likes of LTE [Long Term Evolution, or 4G]. But our Ethernet solution is extremely scalable, and that's one of the principle reasons why we see the mobile operators adopting it: because of the full scalability and manageability."

Do you Circuit Emulate?
So, what specific technologies should a cable MSO looking to address this market have to invest in? "There are multiple technologies available," Baruch says. "The problem to solve at the [cell tower] sites is that there's an increasing need for Ethernet services but - at the same time - there's a large legacy infrastructure in place for T1s and E1s, and that's not going to go away anytime soon".

"Whenever a cellular operator comes to an MSO, there's more than one way to offer the traditional T1/E1 services and Ethernet. You could indeed go with a simple media converter. But the approach that is actually being mostly adopted (and which has now evolved to become a standard) is the use of circuit emulation. This technology provides a way for the Ethernet standard to support legacy T1/E1 services multiplexed in an Ethernet link.

"This gives you full flexibility of addressing legacy services while being able to use Ethernet to gradually increase the bandwidth as you need. This is the technology that we see catching up at the moment," Baruch notes.

Go with this approach, he says, and you stand every chance of not just stealing some market share away from the incumbent telco, but completely knocking it out of the way. "One of the things that differentiate largely the MSO from the telco is that, in order to [meet this expected explosion in bandwidth demand], the telco would need to go with a typical T1 to the site, and there's just not enough copper in the ground to do that," Baruch observes.

"However, if you do it with circuit emulation over fiber, then you can still have those TDM links, as well as the equivalent bandwidth over Ethernet for the predominantly data portion that is growing," he said.

It is a view that ntl:Telewest Business fully endorses: "The key thing for the mobile operators is, firstly, bandwidth and, secondly, single bearer solutions," St.John-Dennis states. "The single bearer solutions are based upon circuit emulation services (or Pseudowire over Ethernet), and we're working closely with a number of vendors on the synchronization of Ethernet under IEEE 1588 version 2, which we the service providers and the vendors are interested in."

It's all very nice to promise cellcos unlimited amounts of scalable bandwidth to help them with their increasing data backhaul needs. However, if cable operators are to convince the cellular guys that they have what it takes to handle their traffic, they will need to prove they're intimately familiar with the strict levels of delay, packet latency and jitter that time-sensitive voice communications demand. And with the increasing success of digital voice services offered by MSOs in many countries, many industry experts believe cable operators have proven they can deliver on these strict levels, both in terms of technical competence and operational experience.

To help with this process, the Metro Ethernet Forum has been busy drawing up a set of specifications (known as MEF-9, MEF-14 and MEF-18) that aim to make cellular backhaul requirements achievable via circuit emulation technologies. CableLabs and some operators are involved in the Metro Ethernet Forum.

"We were the first service provider in Europe to get MEF-9 certification back in 2006," points out St.John-Dennis. "And then last year we also got MEF-14, which really lays down performance, latency, jitter and packet loss requirements. The mobile service providers are very interested in the fact that we've got those accreditations."

In Arris' Baruch view, if anything, it should be the telco world again that should be worried about meeting cellcos' expectations for quality of service: "If you look at the topology of cellular backhaul - especially if we're talking over fiber - there is no issue of QoS. For point-to point connections, the fiber that we have in the ground is capable of running 5 Gbps worth of traffic. And, because it's a different kind of network with different kinds of frequencies, there is no issue whatsoever with QoS compared to the telcos.

"Actually, I would argue that, over an MSO infrastructure, because it's predominantly fiber (versus copper pair or bundled DSL), the quality that you'll get is superior to the one that you'll get through the telco, especially when it comes to reliability and performance," Baruch said.

Should cable indeed be able to match or exceed LCEs' standards for coverage footprint, data throughput, network reliability and performance, the deciding factor will be - as it always is - how much of a discount they will be prepared to offer.