Migration to Ethernet in the Access networks
June 01, 2010
End-to-end Ethernet provides numerous benefits that make the network more scalable, delivering carrier grade high performance at much lower cost per b...
Embedded devices on the edge are driving network bandwidth demand in the Access networks to double yearly, and the benefits of open standards built on Ethernet technology are fast becoming the way to get there. This discussion on the activities going on in the Access networks reveals the key issues.
Ethernet has become the dominant networking technology in enterprises, homes, and embedded applications due to several factors, including its compliance with open standards, scalable bandwidth, strong interoperability among vendors, and low cost of ownership. Emerging Carrier Ethernet standards enable end-to-end networks to deliver rapidly scalable bandwidth, reliable quality of service, and the high availability associated with carrier grade networks while dramatically reducing carriers’ CAPEX and OPEX.
As broadband connectivity increases its penetration the world over, it is important to provide a rich set of services in a broadband connection. The basis of these services is a next-generation network transport that can support not only Internet Protocol (IP) connectivity, but also Ethernet and legacy connectivity, such as asynchronous transfer mode, frame relay, time division multiplexing, and more.
Network bandwidth is now doubling every year, clearly driven by new services that support on-demand video, social networking, mobile applications, and gaming. With support already in place from service providers and equipment manufacturers around the world, Carrier Ethernet technologies are becoming an established platform for profit acceleration. Furthermore, although Carrier Ethernet started by providing last-mile access to the Internet, it is increasingly being used across the provider network.
Scaling current networks
Most of today’s current infrastructure is based on legacy Synchronous Optical Network (SONET). These networks were specifically designed for telecom and had all the characteristics of resiliency, high availability, and manageability as required by service providers. However, what was not considered when these networks were built is that they would have to run at much higher speeds and would require delivery of many other types of private and public services not known at the time.
At the client end, Ethernet equipment can pump out 100 Mbps. At this speed, SONET networks have a difficult time keeping up in an economical manner from both a CAPEX and OPEX standpoint. A traditional DS1/T1 line runs at 1.544 Mbps; the plesiochronous digital hierarchy equivalent E1 runs at 2.04 Mbps. To match up these speeds, multiple T1s or E1s need to be run, which increases the complexity of provisioning, protection, and management. Figure 1 illustrates the increase in SONET costs versus Ethernet.
Ethernet services such as virtual private LAN and leased lines, which enable global enterprises to connect over large geographical regions, have become popular. Providing this service over legacy SONET interfaces is a challenge and drives up costs in a competitive market. Ethernet services have gained popularity in the enterprise as businesses globalize to increase reach and reduce costs. Many enterprises are now using Virtual Private Line Services (VPLS).
As demand for bandwidth increases, providers are forced to upgrade their networks. One of the greatest challenges the service provider faces is increasing the average revenue per user and delivering services at a lower cost per bit. This is where Ethernet becomes a viable alternative to legacy SONET technology. Ethernet is a well-known, ubiquitous entity, and economies of scale have made it extremely feasible as a replacement.
Over the past five years, various standard bodies have launched efforts to develop and standardize Ethernet into a carrier class technology. Carrier Ethernet has been developed to give Ethernet the carrier class capability that was missing in Ethernet, originally developed for the enterprise. The four significant organizations involved in this effort are the Internet Engineering Task Force (IETF), Metro Ethernet Forum (MEF), Telecommunication Standardization Sector (ITU-T), and IEEE 802.1 working groups.
Technologies designed to provide carrier grade characteristics to Ethernet include:
· Provider bridging (IEEE 802.1ah)
· Transparent Interconnections of Lots of Links (TRILL)
· Shortest path bridging (IEEE 802.1aq)
· ITU-T protecting switching (ITU-T G 8031, G 8032)
· Continuity Fault Management (CFM, IEEE 802.1ag)
· Provider backbone bridging (IEEE 802.1ad)
· Ethernet in the First Mile (EFM, IEEE 802.3ah)
From the Carrier Ethernet perspective, several private line services are offered in the metro and Access networks:
· E-line: A MEF-defined service equivalent to IETF Virtual Private Wire Service (VPWS), or ITU-T Ethernet Private Line/Ethernet Virtual Private Line (EPL/EVPL)
· E-LAN: A MEF-defined service equivalent to IETF VPLS or ITU-T; the User Network Interface (UNI) device enables the service, which is also called a demarcation device (see Figure 2)
· E-TREE: A special point-to-multipoint case
By having carrier grade attributes, Ethernet can now be used from end to end to provide services for customers in a physically separate location. Multi-Protocol Label Switching (MPLS) was first introduced in the packet network core several years ago, and since then, has gained popularity by providing private line services.
Services needed in the Access network
Driven by video on demand, IPTV, social networking, and other capacity-hungry applications, the need for high-bandwidth connections has led to the introduction of DSL, Gigabit-capable Passive Optical Network (GPON), and other Ethernet-based services for broadband connection, and the Access infrastructure to support this is quickly moving to Ethernet. Similarly, 3G smartphones are moving to Long-Term Evolution (LTE), meaning that 4G will drive a similar demand for Access equipment to wireless backhaul networks. Having full Carrier Ethernet capabilities in native Ethernet will enable the delivery of end-to-end Ethernet solutions.
The introduction of Ethernet-based customer premises equipment requires that a matching Ethernet Access Device (EAD) be connected to the Access aggregation network. Thus, the migration from T1/E1 links to pure Ethernet with consumer electronics capabilities is already happening. EFM has been developed to expand the capability of Ethernet into the Access.
Links shown at the left of Figure 3 are Ethernet 10/100 connections that have been aggregated into a GbE link that connects to the EAD. Typically, the link between the aggregation device and the EAD is protected with a redundant link that is monitored using CFM. If one link is lost, the failure is detected and a switch-over to the healthy link is achieved. These standards, along with the provider bridge and provider backbone bridge, provide the necessary framework to enable connection and paths from the Access nodes or UNI through the WAN to the end nodes.
With an abundance of standards-based communication protocols involved in the construction of scalable Carrier Ethernet networks, equipment vendors who wish to reduce risk and speed time to market should consider adopting a comprehensive off-the-shelf software platform. To accelerate the introduction of Carrier Ethernet features, IP Infusion has introduced the ZebOS Network Platform, a comprehensive suite of IPv4 and IPv6 Layer 2/Layer 3 MPLS and Metro Ethernet protocols that supports full interworking between Ethernet and IP/MPLS networks.
ZebOS powers many different types of network elements that span both enterprise networks and provider networks. In addition to being modular, scalable, and portable to various single-core and multicore hardware and software platforms, the carrier grade software architecture provides resiliency, high availability, and high performance.
End-to-end Ethernet provides numerous benefits that make the network more scalable, delivering carrier grade high performance at much lower cost per bit. Its ease of use and deployment provides huge return in terms of CAPEX and OPEX. We believe that the migration of Carrier Grade Ethernet into the Access network holds this promise, and major carriers around the world who have deployed this technology are already seeing the benefits.
Asif Hazarika is senior director of product management for IP Infusion, with more than 20 years of experience in the marketing, design, and development of semiconductor products. Prior to IP Infusion, he worked at Fujitsu Microelectronics America, AMD, IBM, Lucent Microelectronics, and Luminous Networks. He is a member of the IEEE and is participating in the IEEE 802.3/802.1 subcommittees and working groups for congestion management, backplane Ethernet, 10GBASE-T, and others. Asif holds an MS in Electrical and Computer Engineering from Oregon State University.