One ray of sunshine amongst the generally depressed economic picture in recent years has been the continuing success and growth of Carrier Ethernet deployment worldwide and especially in the APAC region, as reflected in the region's surge in MEF membership - the industry body largely responsible for defining and promoting Carrier Ethernet - and the provision of services certified to MEF standards.
Initially it was suggested that Carrier Ethernet's value proposition was the key to maintaining sales in a shrinking economy -it offers the cost advantage of flexible, granular scalability that allows bandwidth to be added, or reduced, on demand and at short notice. But the MEF always made a point that cost saving was only one factor in its success. The real benefit would turn out to be the new business opportunities being generated by Carrier Ethernet's performance, features and flexibility - and nowhere are these more apparent than in the dynamic Asian markets.
As the economy begins to recover, we see new business models emerging around Carrier Ethernet services, with organizations taking advantage of the new connectivity opportunities, already reaping significant benefits.
Three new developments stand out in this respect: the creation of Ethernet exchanges, Ethernet for mobile backhaul and global interconnectivity. Ethernet Exchanges already feature in this supplement, so this article focuses on the second two factors.
According to industry analyst Michael Howard of Infonetics Research: "Ethernet is the only solution for next generation mobile backhaul networks. When you take into account the popularity of iPhone, Android and the new wave of feature-rich mobile devices, you can see the operators are under pressure to deliver a lot more bandwidth in a highly cost-competitive market. Legacy TDM and ATM backhaul solutions cost too much and don't offer anything to match the scalability and other advantages of Carrier Ethernet."
Backhaul costs are the biggest challenge facing wireless service providers today. Companies that had invested heavily in the 3G spectrum auctions were under pressure to reduce further capex spending and so opted for leasing backhaul infrastructure. However, narrowband TDM turns out to be a poor choice because, as T1 reaches near capacity, the network performance degrades by over 40%. Sonet isn't suited for demand variance and TDM field capacity scales at 1.5 Mbps, 45 Mbps, and 155 Mbps, committing the customer to investing in big, costly jumps in capacity.
TDM only supports one Class of Service at fixed QoS, while IP allows for CoS/QoS according to a packet and application priority weighted against the willingness to pay. TDM signaling is static while IP traffic is dynamic. Although TDM guarantees bandwidth, Ethernet is designed to scale with demand in ranges from kbps to Mbps. Although DS1/E1 when cabled, tested, and commissioned over Sonet meets the incremental needs of TDM, it is the least suited technology for TCP/IP.
Such inflexibility is a major handicap in a fast-evolving business environment, and the rise of the smartphone has driven surging demand. Operators are expected to keep prices low while deliverng vastly increased bandwidth to serve new, demanding applications. And it isn't just the bandwidth - it's the unpredictability of traffic levels that makes the flexibility of Carrier Ethernet so attractive.
In place of TDM's one-style-for-all, Carrier Ethernet lets you create new services based on bandwidth guarantees combined with additional temporary bandwidth needed to satisfy unplanned real-time demand deviations. Priority routing by application is now possible, because the carrier dictates how packet handlers and schedulers manage the overall traffic. Carriers can still offer delay and jitter service guarantees for VoIP and video, while common data can be managed by a lower cost service connection.
Carrier Ethernet offers the best of both worlds: a perfect TDM solution that is also geared up to address emerging technologies for next generation wireless architectures such as Wimax and LTE. It suits providers wanting high quality Ethernet connections to their own Circuit Emulation technology; and it equally suits providers needing a high quality managed Circuit Emulation product with separate Ethernet for data applications.
MEF 22 (Service Implementation Agreement (IA) for Carrier Ethernet services in mobile operators backhaul networks) was launched last year to accelerate the transition by outlining ways to apply existing MEF specifications and industry standards to meet the needs of backhaul. It provided a practical migration path to 3G and 4G services, while preserving legacy services revenue.
This opens up an exciting business proposition for mobile backhaul service providers across two rapidly expanding markets - mobile backhaul applications and Carrier Ethernet - as well as an opportunity for fixed network operators to redeploy their existing infrastructure to provide wholesale services to mobile operators. As a result, the number of cell sites worldwide served by Ethernet backhaul is expected to rise tenfold since 2009, to 1,061,000 million (a third of all cell sites) by the end of 2013.
Between 2000 and 2009, Ethernet had extended from a local area technology through the MAN and the WAN to embrace the world. An enterprise connected via MEF certified Carrier Ethernet becomes truly global, with the distance between its New York and Beijing offices measured purely in terms of the speed of light.
In business terms it was not quite so simple. Unless the service provider happened to own a complete network infrastructure linking all the global sites, the global network would require connections to be made between two or more service providers. Although each carriers' services could be fully MEF certified, establishing each connection would require a host of parameters to be aligned. The very flexibility of Carrier Ethernet meant more options in terms of SLAs and QoS parameters, each needing to be brought into line, as well as standardization of contractual agreements for each connection. It was achievable, but it typically required months of negotiation to establish these links.
So 2009 saw the launch of a new phase of activity by the MEF, aimed at fully scalable worldwide operation. The MEF describe this as "Global Interconnect" ?the interconnection of autonomous Carrier Ethernet networks to enable standardized and streamlined delivery of MEF-certified Carrier Ethernet services with end-to-end Class of Service, management and protection.
It began with the launch of MEF 20, automating the management of the link between the user network and the WAN. MEF 21 addressed the Link OAM (Operations Administration and Maintenance) - equipment certified to MEF 21 will replace costly truck rolls, on-site diagnostics and troubleshooting with remote management. Meanwhile basic business templates were created to streamline the local ordering and provisioning of Ethernet access and standardize the process of interconnection.
Then in January 2010 the MEF announced MEF 26 - defining Phase I of E-NNI, specifying how to join a network to another, external network - typically one operated by a different carrier. The issues include OAM, CoS and SLAs to ensure that; for example, one carrier's "premium business service" will not be routed through a lower quality service.
E-NNI is more than just another specification; it has become the impetus for an entirely new wholesale Ethernet market. Take CoS as an example. Where two service provider networks meet, the E-NNI preserves priority information across the boundary of the networks allowing both high-priority traffic such as VoIP calls and low-priority traffic such as overnight server backup to run over the same network while preserving priorities.
The MEF's Global Interconnect Program is bringing all these steps together to cut the time and cost of Ethernet service deployment, to reduce capital and operating costs by replacing hardware with software solutions, to empower providers to extend their coverage and generally increase business efficiency.
Nan Chen is president of the Metro Ethernet Forum