Hundreds of Columns Sorted by Topic
Ethernet-based Access in Fiber-to-the-Home
by Thomas Eklund, VP of Marketing and Business Development
6/23/2009
The fiber access build-out tsunami is swelling. The promises of fiber-based access are huge and over the coming years millions of ports will be installed worldwide. The old copper- based infrastructure will be replaced with a modern fiber-based infrastructure ready to deliver advanced real-time applications like HD quality video, IPTV and online gaming.
Today more than 750 million households connect over copper and only 20 million over fiber. While it will take years to make a full transition to a fiber-based infrastructure, in the modernized world (U.S., Europe and Southeast Asia) a rapid transition is in progress. In Japan, NTT has started to build out a nationwide Fiber-To-The-Home (FTTH) network using EPON with over 11 million ports already installed as of 2007. In Scandinavia, Telenor (B2) and TeliaSonera are building out nationwide FTTH networks based on Active Ethernet. In the U.S., Verizon and AT&T have started to deploy a nationwide FTTH network based on GPON. The technology deployed is dictated by user demographics, local regulations and the economics of the legacy deployments. As a result, there will be a mix of several technologies (EPON, GPON, 10G PON, WDN-PON, Active Ethernet, LTE, HFC) fighting for Ethernet-based access deployments.
Fiber-to-the-Home Deployment Models
The Ethernet access based services associated with FTTH have been delivered over a variety of deployment architectures. The most common architectures include:
Fiber-to-the-home (FTTH) -- the fiber path is terminated in the home living space or office space
Fiber-to-the-building (FTTB) -- the fiber path is terminated at the boundary of the private home or office space or the basement of a multi-dwelling unit. The last access will then be carried over another medium such as copper (VDSL+/2)
Fiber-to-the-curb/node (FTTC) -- the fiber path is terminated in a distribution point off the Central Office (CO) or similar location using fiber from the CO to the distribution point (Remote DSLAM, ONU, ONT)
Replacing older media and transitioning to truly high-speed access creates new opportunities and challenges for carriers. While FTTH signals the end of access network upgrades for the foreseeable future, this means carriers must focus on service revenue creation to secure and migrate or substitute existing services with high value services on the new FTTH infrastructure. Carriers need a scalable and flexible solution that supports the functions previously only required for the core networks; high reliability, rapid provisioning and protection mechanisms. Network equipment vendors (NEVs) must offer scalable platforms having flexible deployment options enabling the carriers to provide FTTH at the right cost, low power and high throughput performance. All of this can be accomplished through a proven metro technology called Carrier Ethernet once it is extended into the access network.
The Benefits of Carrier Ethernet
Cost and performance are the primary drivers for adoption of Ethernet-based fiber access. Because Ethernet has already proven itself as a superior technology for both enterprise and Metro Ethernet, it is now the key foundation for driving down capital expenditures (CAPEX) and operational expenditures (OPEX) in most carrier deployments today.
Carrier Ethernet is a ubiquitous, standardized and carrier-class service defined by five attributes that distinguish it from the all too familiar LAN-based Ethernet. These services create an end-to-end Ethernet service for business and residential users.
- Standardized Services: E-LAN, E-line, E-tree
- Quality of Service (QoS): Advanced traffic management, end-to-end service level agreements (SLAs), bandwidth profiles, class of service, committed burst size, excess burst size, excess information rate, committed information rate
- Service Management: 802.1ag, 802.3ah, ITU1731, Carrier Class OAM, link and service connectivity management, rapid service provisioning
- Scalability: In terms of number of users, spans access and metro to national and global services over a wide variety of physical infrastructures implemented by a wide range of carriers, and scalability in terms of bandwidth from 1 Mbps to 100 Gbps.
- Reliability: High service availability, detect and recover from incidents within 50ms, node and link redundancy
The Metro Ethernet Forum (MEF) is the industry alliance that drives the worldwide adoption of Carrier-class Ethernet networks and services. Lately, the focus of the Forum has been on extending service models to the first mile. New features have been added to Carrier Ethernet to better manage, service and operate the network, commonly referred to as Carrier Ethernet 2.0. For example, Ethernet-based OAM, improved QoS capabilities for sensitive real-time applications such as IPTV and VoIP, flexible service provisioning and synchronous Ethernet for wireless backhaul.
Carrier Challenges
Carriers need to plan for future growth, and to do so, they must be able to introduce new, revenue-generating services. IPTV has proven itself to be the killer application for FTTH deployments and it is vital that the high standards for user experience are maintained as networks transition from today's television broadcast over-the-air and HFC networks to FTTH networks. Consequently, NEVs must position themselves with the wirespeed distribution and switching equipment required to support this new network.
Another challenge for carriers is the maturity of the Carrier Ethernet standardization. Regional differences in deployment create a price and technology bar that could be overcome if the right solution could be programmable at the right price. This allows carriers to future proof their networks so they can begin mass volume deployment in advance of formal standardization, and be better positioned to take market shares.
Carriers are looking for technologies that can last several decades and are future proof in terms of performance, scalability and the flexibility to introduce new services over time. Today, 20 percent of the cost per installed FTTH port is related to the hardware cost. Programmability is a vital component of extending the lifetime of the equipment as well as increasing NEV interoperability.
Carrier Needs
As carriers plan for the growth of new Ethernet-based FTTH deployments, they must balance competing priorities between ubiquitous deployments, service revenue growth and service quality, while optimizing total cost of ownership. As in the early stages of other technology deployments, subscriber-line cost has had a domineering influence on the FTTH technology selections and service deployments. There has been little regard for the need for a scalable access infrastructure having flexible platforms capable of multiple deployment scenarios. In addition to high density platforms, new multifunction platforms having the ability to be deployed in different tiers of the network and with a combination of Ethernet, PON and DSL interfaces are required to optimize deployments of FTTH. These flexible access products extend the useful life of the network and promote vendor interoperability.
The new FTTH infrastructure must support the consumer-oriented services where user expectations for quality are well established. Carriers believe that the extension of Carrier Ethernet into the first mile provides a robust and resilient foundation for these deployments. This enables them to flexibly provision new services and provide SLAs while making use of the service management and provisioning platforms built for Carrier Ethernet.
Efficient delivery of high quality IPTV and high-definition video is essential as one of the key applications driving FTTH but it needs to be done with high user experience (HDTV quality). Initial deployments have given rise to a multitude of revenue-generating service opportunities that rely upon carriers to enable an interactive user experience in combination with real-time conferencing and gaming, media services and social networking applications along with access to the Internet. These new services will likely have a broad range of requirements, but as a minimum will require platforms to flexibly manage the QoS of the traffic flows within the network and intelligently manage distributed content caches. As result, a flexible and distributed traffic management capability must be integrated into both the upstream and downstream traffic flows.
While the economics of access deployments demand that subscriber-line cost remain a key consideration, FTTH business cases increasingly rely on total cost of ownership arguments given the opportunity that FTTH creates new revenue-generating services and the ubiquity that a flexible access infrastructure brings.
An Ethernet-based FTTH Solution: View from a Network Equipment Vendor
In response to the carriers need, NEVs seek an Ethernet-based access solution that scales over different media and aggregation points. The new Carrier Ethernet-based fiber access networks drive GE performance and bandwidth to the home with 10 G as uplinks in the access nodes. The increased bandwidth poses new scalability challenges to be taken into account when designing new access platforms. A key component of next-generation Ethernet-based access equipment is to incorporate new technologies into the design of the system that meet the carrier's need at the lowest cost.
NEVs are driven to an architecture that shares the same processing, switching and traffic management subsystem with the media options demanded by the carriers (Ethernet over Sonet, SDH, PDH, Fiber, EPON, GPON, WDM, VDSL, LTE, HFC, etc.). This type of architecture can be adapted to fit into either pizza box or chassis-based solutions, and as a result, minimizes the cost of development by heavily reusing the same Typically, designs of this type save time-to-market and have very efficient product development cycles.
Requirements for the processing, switching and traffic management subsystem are demanding and in aggregation systems having 10 G uplinks can require up to 500,000 flows of different priorities be individually managed. The quadruple-play services being planned for the carriers demand advanced traffic management that provide fine-granular user flow management and control to ensure IPTV and video-on-demand are delivered with high quality. Similarly, there are stringent requirements for a fully programmable processor and switch that addresses a full range of data plane applications and guarantees deterministic wire speed performance at all packet sizes. At the application level, this subsystem must support the evolving Carrier Ethernet service and management standards, including the resiliency requirements for restoration, rerouting, node redundancy and in-service upgrades.
Programmable Solutions
In addition to the chassis design and hardware architecture, the flexibility of new platforms designed for the FTTH market relies on the NEVs ability to make programming changes to the data plane and incorporate new features into these platforms. NEVs must offer a programmable solution to carriers that meet the growing demands for capacity, new services and investment protection. These programmable devices are at the core of the system and represent a substantial recurring cost to the carrier and development cost to the NEV. Below is a brief description of a variety of programmable solutions as well as a table comparing alternative solutions for building an Ethernet-based FTTH platform:
Network Processors
FPGA
Fixed function Ethernet Switch ASICs
Multi-core Processors
Programmable Ethernet Switch
Hybrid Fixed solutions of fixed function Ethernet Switch ASICs with FPGA
The matrix highlights the Programmable Ethernet Switch as the ideal solution for NEVs who are demanding carrier class components that intelligently integrate functional elements of the system, yet maintain the flexibility to support features and applications unanticipated by the original design. The integration of an Ethernet switch with a network processor not only reduces the number of components on the line card, but provides new capabilities which enable NEV designers the ability to customize the switching functionality in the same way they program applications for the network processor.
Summary
The rapid transition to fiber infrastructure is underway and carriers need to create a more flexible deployment model. New FTTH network architectures will be based on a Carrier Ethernet service model for running quadruple play and business services to consumers and business users very efficiently. Carriers also need to design a network that can efficiently deliver IPTV, Video-on-Demand and HDTV into a cost-effective, common Ethernet-based access infrastructure.
On the other hand, NEVs need to design a highly flexible design that fits all of these next-generation Ethernet-based designs to increase the carriers' time to market while maximizing their total cost of ownership.
Programmable Ethernet Switches offer a unique solution to deliver these new Ethernet-based FTTH network designs at the right cost, power, performance, scalability and flexibility. All other architectures based on fixed-function Ethernet switches, FPGAs and Network Processors are either too expensive, or too limited in terms of performance and scalability of flexibility. Programmable Ethernet switches give NEVs an answer to carrier demands to reduce platform cost while increasing the capabilities of the platforms and supporting ongoing feature development for revenue-generating services.About the Author
Thomas Eklund, VP of Marketing and Business Development Mr Eklund brings over 13 years of experience in Ethernet Switching, Mobile Internet and IPv6. He is currently a member of the IPv6 Forum Technical Directorate and co-author of several IETF drafts. Eklund is responsible for Xelerated's business development and for securing the company's first tier-one customers. Prior to founding Xelerated, he held key positions as Systems Engineer at SwitchCore and Research Engineer at Ericsson Research. Mr Eklund holds a Master's Degree in Computer Science from the University of Stockholm, Sweden.
About Xelerated
Xelerated is the global leader in field-proven next-generation ASSP-based Carrier Ethernet chipsets. The Xelerated Dataflow Architecture delivers wirespeed deterministic performance, enabling network equipment vendors to build fully-programmable carrier-class systems that reduce R&D costs, minimize risk and accelerate time-to-market. With its highly flexible network processor platform and linearly scalable products and solutions, Xelerated is an integral part of Tier One carrier networks around the world, and a vital component of any Metro, Access, or High-end Enterprise market. The company has offices in Santa Clara, Stockholm, Tel Aviv and Beijing.
Send us your response to this article.
Learn How to Get Your Column Published on this Site
