Mobile operators are gambling on how often they can keep loading the infra-structure  before it goes bust. Perhaps help can come from an unlikely quarter, says Stephen Rayment, CTO, BelAir Networks.

Anyone who's ever been to a casino – or even on a camping holiday in the rain – knows how to play pontoon; it's a simple game of chance where players stake on how long they can keep adding to their hand before they go bust. It's a game that seems to have found its way into mobile data as MNOs start to gamble on how long they can hold out on their existing networks before the amount of traffic overloads the individual cells.

With many MNOs planning to implement LTE in the next three to five years, the goal is to string out the lifespan of existing infrastructure until 2014. At the same time, carriers want to secure their share of the rapidly-growing community of smartphone users today. All the time, the MNOs are continuing to up the ante. Twist: an exclusive deal with a smartphone manufacturer. Twist: a tie in with a popular social networking site. Twist: a new app store. But how long can the carriers say twist before the infrastructure busts?

Poker face
2009 was not a great year for the public perception of MNOs with half a dozen highly publicized network outages in Europe alone. With the press seizing on the outages, a publicity battle ensued on an international basis along the lines of "my network's better than your network".

When you dig beneath this, however, there lies a bigger issue as to what's actually causing this downtime. Have some networks already hit "bust" in some locations?

Certainly there can be other explanations for a network outage, besides capacity constraints, such as damage to a fibre cable. However, in the absence of any other wire-based explanation, it might be fair to assume that outages can represent early sign of cracks in an overloaded area of the network.
But, with reliability and availability featuring high on the tick list of mobile shoppers, the MNOs are likely to keep pokerfaced about how close to capacity their networks might be.

Full house
In the history of 3G networks, overload is a relatively new concept. From the spectrum auctions at the turn of the Millennium until the launch of the iPhone, high levels of mobile data traffic were largely the stuff of optimistic carrier and vendor predictions made at trade shows.

The impact that Blackberry had in securing a steady stream of email traffic across networks cannot be underestimated but nothing seems to have launched as many new users into the world of constant data connectivity as the iPhone. And iPhone users tend to be much more data hungry than those carrying Blackberries. Few business users choose to download PowerPoint presentations to their handhelds as they're hard to interact with, so a device that's email centric doesn't typically use a lot of bandwidth. Orange, for example, states that 100 emails downloaded without attachments will use about 1MB of data. Similarly, people with feature phones, rather than smartphones, will, according to Strategy Analytics, use only around 5 to 10MB of data a month.

Smartphone users, on the other hand, who are more likely to engage with whizzy applications and media streaming services, will consume 10 to 20 times that amount of data, which might be one reason that O2, with its exclusivity on iPhone sales in the UK, commented last year that it saw an 18-fold increase in the amount of data going across its network.

All of this still pales when we look at dongle use. According to Phil Kendall at Strategy Analytics, dongles are overwhelmingly responsible for mobile data traffic load, pulling, on average, 1.5GB of data across the networks every month. That's still around ten times the amount that a smartphone user consumes.

No wonder then, that in the 12 months to May 2009, Orange reported a 4,125 per cent increase in the amount of data traffic on its network caused by dongle use.

So, over the past ten years, we've seen data use dawdle for the first half of the decade, gradually rise in the third quarter and then take off like a rocket in the fourth. It comes as no surprise that this might have impacted some mobile networks.

Hit me dealer!
In spite of this, competition to gain new data customers is fierce. While we can't know how many customers the other networks may have lost to O2 when they won the exclusive deal on the iPhone, it's clear that they accepted the impact the phone could have on customer retention. Attractive new data services plans became a top priority for all the major networks.

This involved some smart thinking and innovative marketing. New music-sharing services and applications stores were launched and all of the carriers fell on the handset manufacturers to secure deals on the next new "iPhone killer".

As a result, more of us than ever are carrying smartphones. Vodafone now has 18.7 million users, of which, more than a quarter have web-enabled handsets.

High rollers, low returns
Competition has taken its toll on profitability though. Consumers have become used to an "all you can eat" pricing structure for buying data capacity. The 56% of UK households (ONS) that have a broadband connection have become used to paying a monthly fee to upload and download as much content as they want. This mentality crossed over to mobile data consumption and it is now normal for a mobile data user to get 3 – 10GB of data as part of their package.

So, for a 2,000 per cent increase in data traffic, a user will usually pay around 25% more on their bill.

Stick
It's for these reasons that many carriers are looking for ways to optimise their networks to increase capacity. But it's simply not economically viable to be pouring too much money into an infrastructure that won't deliver a return in the short term.

LTE is looming large on the horizon and is the focus for investment for the majority of carriers in the coming years. Many are already hugely committed to the task of building their LTE networks with some, expecting to cover their current 3G footprints as soon as 2012.

The GSMA has predicted that there will be 87 million LTE subscribers by 2014 – a figure bumped up to 100 million by analyst firm Juniper Research. New networks will make way for more bandwidth, as well as the opportunity to review pricing models.

But what happens until then? And, in reality, what happens at the same time? Whilst the GSMA is suggesting wide scale uptake of LTE, it's also predicting continued growth in HSPA users – 1.5 billion by 2014. This isn't an issue that's instantly going to go away when LTE arrives.

Aces high
There are a number of ways to optimise current networks. At least one operator invested heavily last year to upgrade the speed of its backbone and, by doing so, increased capacity by 30 to 50 per cent. Similarly, load balancing across the network can help to allocate bandwidth where it's needed most, minimising the impact of usage spikes.

However, these approaches answer the problem at the backhaul level only. They don't take account of overloading at the individual cell level, which is a much more pressing concern. Most people have experienced the difficulty of getting a good signal at a conference or concert where large numbers of phone users are all crowded together in a relatively small space. This is because we're all sharing the capacity of a single cell.

Now, multiply that use by 20 to account for the data traffic and you can understand the pressure on that single cell to support its users. And, as areas where high data users converge tend to be places like city centres, business parks and financial districts, it's the high-value customers who feel the pain first.

Split your bets
Rather than embarking on the time-consuming and costly effort of establishing a new cell site, an operator can choose to split the traffic to get some of it off of the cell. Peter Cochrane, BT's chief scientist and futurologist, recently suggested the solution is for carriers to split some traffic onto other MNOs' networks, that is, to share cells. It's not unheard of: O2 and Vodafone do this in some places. It does, however, raise issues of control, ownership and differentiation for the networks.

Femtocells represent another alternative to split traffic off. A common femtocell business model is to have customers pay for their own hardware, which they also pay to power. They also pay for connectivity as the femtocell accesses the network over the user's home broadband line. In return they can register a number of SIMs to connect to the device and achieve uninterrupted service in areas that may have previously experienced coverage issues.

From the operator's point of view, the benefits are obvious as consumers carry the cost of the network upgrade. However, it remains to be seen how many will be willing to do this. It also fails to address the issues of the millions of roaming users passing through all those locations "where people congregate" who won't benefit from this and may continue to experience service challenges in areas of high user concentration.

The other solution is for carriers to add functionality to their networks that allows them to respond in a more cost-effective way. Rather than adding cellular capacity or urging users to create private wireless networks, MNOs can deploy carrier-class Wi-Fi networks that allow them to hand off data traffic from the cell.

Trumps
This kind of network can add the same capacity as adding a new cell for what is typically less than a third of the cost. It also allows MNOs to add capacity on a targeted basis, building coverage in high-density usage areas today, whilst allowing low-level areas to grow on the existing infrastructure until LTE arrives.

And, because it's only the data that's handed off to Wi-Fi, there is no danger of diminishing the quality of voice services that still lie at the heart of the carrier's business.
Essentially you can think of Wi-Fi as a giant offload point for wireless data traffic.

The winner takes it all
Which brings us back to the stick and twist scenario that we began with. Should the operators stick with the networks they have or twist and try something new? In reality, none of us can afford to stick. We don't know what's in competitors' hands, as they play them close to their chests, but we know that it's probably good. So, if the carriers don't want to be aced, I think we'll see a new twist on Wi-Fi in 2010, with carrier-grade expectations placed on the ubiquitous license-free technology once shunned by operators.

Operators need to meet the high levels of QoS demanded by users, or they risk losing  business. But what steps do they need to take to ensure a quality customer experience? 

Throughout 2008 mobile broadband saw dramatic growth in subscription levels.  In fact, at one point during that year UK telecoms operator 3 reported a 700 per cent increase in data traffic resulting from the launch of its mobile broadband packages.  They were by no means alone in experiencing this growth.  It appeared that a true alternative to WiFi had entered the marketplace and the future for mobile broadband seemed assured.

Today however, as we enter 2010, the success of mobile broadband is coming into question.  In October 2009 The Carphone Warehouse CEO Charles Dunstone noted that the company was seeing a decline in customer interest for mobile broadband and a corresponding rise in sales for fixed-line offerings.  He attributed this to the slower speeds and lower quality end user experience provided by mobile broadband.  A report by Thinkbroadband.com mirrored these results, showing that 76 per cent of mobile broadband users were unhappy with the speeds provided while 60 per cent thought the coverage was poor.

Two years ago when the outlook for mobile broadband was much more optimistic than today, many had not taken into account just how successful touch-screen smartphones such as the iPhone would be.  The iPhone has achieved a ubiquity not seen for a device in the mobile phone space since the 1990s.  It has provided an easy method of accessing 3G mobile broadband services and has driven consumption of these services to new levels.  Operators have frankly started to struggle under the weight of demand and the end-user experience has further suffered.  The situation led to the CEO of O2, Ronan Dunne, apologising to customers for the poor performance of its mobile data network in the second half of 2009.  

Mobile broadband, it seems, needs to improve its user experience if it is to remain a revenue-generating proposition for carriers.  But what exactly is the problem, and how can operators address it?

A look at North America, where users of mobile broadband services enjoy a much higher quality of experience, is useful in understanding the challenges facing European operators.  From the outset European operators focussed their efforts on marketing mobile broadband services.  This led to an exceptionally rapid deployment and market penetration of 3G ‘dongles' and datacards. It was essentially a hype-cycle that continued to be fuelled as more and more subscribers took up the services promoted so heavily by the major operators.

In North America, however, operators approached the rollout quite differently.  They concentrated first on ensuring that the backhaul they had in place was sufficient to cope with the rise in traffic at the cell that  mobile broadband would generate.  With hindsight this appears to have been the best approach to take, but at the time it was something of a chicken and egg question – do you create a market and then upgrade the backhaul when you know the market is there, as in the European model, or do you upgrade the backhaul and hope the market will meet expectations as in the North American model?  Neither is risk free, and it is primarily due to the fact that the marketing of mobile broadband was so successful in Europe that the networks are now struggling to cope.  It is clear that European operators must now focus on improving backhaul if mobile data services are to deliver on their promise. 

GET TO GRIPS WITH BACKHAUL
The first step is to replace legacy TDM backhaul used for data, with Ethernet.  One of the biggest issues operators face with mobile broadband Quality of Experience is ensuring there is enough bandwidth getting to the cell.  Key to this is to have a platform that can dynamically manage the bandwidth to guarantee that there are no bottlenecks in the backhaul network, the primary cause of poor mobile broadband quality.  Ethernet provides the best solution to this as it delivers both high-capacity connectivity at a low cost and has the intelligence to manage bandwidth effectively. 
Ethernet, however, presents its own challenges to operators.  Ethernet backhaul for mobile data requires a large number of Ethernet Virtual Circuits (EVCs) to be configured at the cells.  EVCs are highly complex to configure correctly and require a completely different approach to testing than operators are used to in the TDM environment.  Correctly configuring the EVCs and testing to ensure for optimal performance is vital in ensuring that there are no bottlenecks in the backhaul network.  So how can operators do this?

To ensure that the configuration, testing and management of the EVCs are conducted correctly, operators need to address three areas: circuit turnout, SLA monitoring and troubleshooting.  Once these three core areas have been tackled, the quality of the overall mobile broadband experience will improve dramatically. 

Circuit turnout testing needs to be carried out to ensure the EVC configuration is correct and that the data bursts run effectively.  There is greater complexity here than in circuit testing for voice which it transmitted in a steady and more predictable flow.  For all mechanisms to be tested thoroughly, around six and a half hours need to be devoted to circuit turnout. 

MONITOR THOSE SLAS
The second foundation stone of high-quality mobile broadband service delivery is SLA monitoring.  For this to be effective monitoring needs to be conducted on the service layer as well as on the physical layer.  As with TDM, multiple EVCs require highly granular service monitoring in order to be able to effectively calculate availability.  Operators must be aware that the way in which the Operations Administration and Maintenance (OAM) mechanisms are architected is just as important as the service and infrastructure build.  Each plays an essential role in allowing operators to provide high levels of service quality that can be measured against SLAs.

Effective OAM is an integral part of delivering mobile broadband.  Operators need to manage mobile backhaul solutions, both network elements and services, end to end in an integrated fashion. The right OAM solution is an essential tool for element and service provisioning, network maintenance, and fault management, with service visualisation to enable the identification of customers or services affected by network troubles.  By managing network elements and tracking individual services across multiple products, operators will be able to manage broadband services more effectively and ensure their customers get the quality they require.

TROUBLESHOOT
Finally, trouble shooting is another core requirement.  Operators need to be able to look into the network to ensure packet prioritisation is taking place.  Without this ability, operators have no way of understanding why a data service does not run at the required speed and jitter-free.

Forward looking operators across Europe are making the move to an all-Ethernet backhaul infrastructure, but the migration route must be planned carefully.  Network timing and synchronisation will become increasingly important over the next few years.  Timing and synchronisation is mission critical to stop handover and interference within cell sites, and new technologies, currently in development, will mean that this important area of mobile broadband delivery will function more effectively than ever before. 

With any telecoms service, there is always room for unforeseen disruptions to the network.  With voice the damage was always limited to a dropped call and slight inconvenience to the end user.  Mobile broadband is a different proposition altogether, however, and results in much greater disruptions for the end user – loosing email, social networking tools, or video downloads can be a lot more stressful that a dropped call. 

Network resilience and protection are critical to applications based on Carrier Ethernet services, and a full feature suite is required for resilience and protection of wireless backhaul using 802.1ag, Y.1731, G.8031, and other emerging related standards. Combining resilience and protection methods with OTN-based performance monitoring and fault sectionalisation in the metro network optimises mobile backhaul availability and guarantees uninterrupted service for end users.  

However, the backhaul networks do not operate in isolation. To work efficiently and leverage Operational Expense (OPEX) advantages, the mobile core network also must evolve as the access network migrates to Ethernet. 

Building a Carrier Ethernet network infrastructure (using standards defined by the Metro Ethernet Forum) provides operators with a long-term, low-cost strategy to replace their existing SDH infrastructure while maintaining carrier-class reliability. 

As operators have invested heavily in their current mobile networks, they cannot afford to simply tear out and replace current legacy equipment. It's crucial that their mobile backhaul and core network strategy still supports legacy traffic and services while allowing them to gradually transition to next-gen infrastructures that are more scalable and economical.

By combining a hybrid electrical/optical Reconfigurable Optical Add/Drop Multiplexer (ROADM) with Ethernet aggregation and switching in a single chassis, operators can enjoy more cost-effective metro packet/optical transport. The resulting platform enables the convergence of mobile backhaul with residential and business services traffic on a unified metro network, allowing operators to deliver high quality mobile broadband services while gaining huge savings in OPEX. 

Fixed telecom operators are already benefiting from the migration from TDM-centric to next generation Ethernet centric networks. Mobile operators must now also manage the transition to next-generation Ethernet to maximise the investment in existing mobile and network infrastructures while maintaining a quality of service that minimises subscriber churn.

In summary, demand for mobile broadband is growing, with consumers requiring support for rich multimedia and higher bandwidth. Yet consumers remain price sensitive and are not willing to pay more for higher speeds. At the same time, consumers view current speed and quality as the main disadvantages of mobile broadband compared to fixed broadband solutions.

The challenge facing operators in the mobile space is how to reduce the total cost of running the network while increasing the bandwidth (capacity and speed) and offering superior quality of service.
This highlights the need for operators to focus their investments in specific areas of next-generation network infrastructure to improve existing offerings, accommodate increasing bandwidth demand – while offering more competitive services and maintaining profitability.

This article was co-authored by: Vinay Rathore, senior marketing director, Ciena,
Reza Vaez-Ghaemi, Americas Market Management and Technology Research Manager,  and Jay Stewart, Director Ethernet Service Assurance,
both JDSU.