|
Supporting Ethernet Transport Services
Introduction Over the last few years Ethernet technology, and services based on Ethernet technology, have emerged as an indispensable component of the broadband networking and telecommunications industry, for either network operators or service providers. As an ex-ample, worldwide enterprise customer demand for Ethernet Services by itself is forecast to hit the $30B US mark by year 2012 [1]. Use of Ethernet technology in the feeder net-works that support residential applications, such as “triple or quadruple play” voice, data and video services, is equally on the rise. As the synergies between packet-aware trans-port and service oriented equipment continue to be exploited in the path toward transport convergence, [2] Ethernet technology is well positioned to play a critical part in the evolution towards a converged Optical /Packet Transport infrastructure.
A large amount of work has been undertaken by a number of standardization organi-zations to create an architectural framework to foster the commercial development of the carrier Ethernet market. Ethernet, as defined by the IEEE 802, has a vertically integrated architectural model with physical [3], data link (MAC) [4], and MAC bridging layers [5][6][7] modeled after the OSI reference model. This layered architecture model allows Ethernet technology to be re-used in a variety of data networking and transport roles in support of carrier centric services, as illustrated in Figure 1: - as a service interface used to enable business or residential access to L3+ data services - as a connectivity service used to provide a high bandwidth alternative to Private Lines, Frame Relay, ATM connectivity services for Enterprise site interconnection - as a network infrastructure service used as one of the building blocks of a network operator, or service provider, transport infrastructure (similar in scope to WDM, SDH/SONET or OTN transport services) In most carrier applications “electronic-centric” (TDM or Packet) approaches tend to be preferred in scenarios with high subscriber density; broad range of interconnect granularity (in terms of bandwidth demand) and fluid service demand. Conversely, "photonic centric" approaches tend to be preferred in scenarios with coarser interconnect granularity and more predictable service demand. The Metro Ethernet Forum (MEF) has created an Ethernet services architecture framework [8][9] intended to further enable the definition of Ethernet retail and wholesale connectivity services [10][11] based on native Ethernet physical layer and MAC frame format for its service interface specifications. Well defined service interfaces have been identified to provide i) a User-Network Interface (UNI) between a subscriber and its ser-vice provider, ii) an External Network-to-Network Interface (E-NNI) between Ethernet ser-vice providers, and iii) a Application Service Interfaces to access, say, IP/MPLS services. In this paper we discuss various leading approaches in deploying Ethernet Services and highlight distinctive capabilities in network elements (NE) required to deploy such services. We also contrast the leading candidate technologies in terms of support of customer requirements with respect to ability to handle anticipated traffic demand, network reach/coverage, subscriber density, and end-user service mix, among other considerations. In particular we note that a variety of optical and packet technologies are required to provide Ethernet services that expand beyond a metropolitan area domain. Currently, native all-Ethernet networking solution need be complemented with other transport technologies such as MPLS [12][13] and its transport network optimized versions, Transport MPLS (T-MPLS) [14] / MPLS Transport Profile (MPLS-TP) [15] in order to support scalability requirements for national and international services. Enabling Ethernet Transport Services In a telecommunications environment base Ethernet technology can be used by itself or in conjunction with other transport network centric technologies such as SDH/SONET [16], DWDM/OTH [18] or (T)-MPLS/MPLS TP, to deliver a full complement of carrier-grade forwarding, Operations, Administration & Maintenance (OA&M), survivability and management features for Ethernet oriented services. These so-called Ethernet Transport Services, as a counterpart to traditional Enterprise grade Ethernet, address network operator and service provider needs for carrier-grade packet transport solutions based on Ethernet technology. Below we explore further each application space. Ethernet as Dedicated Network Infrastructure Network elements in this application space deliver “virtual fibre” based connections re-quired to implement an Ethernet-oriented transport network infrastructure. The functional blocks used to implement these services are modeled and managed after ITU-T require-ments for optical transport network equipment [17][19]. The use of IEEE 802.3 physical layer interfaces to interconnect network equipment also falls as a usage case in this ap-plication space. A “virtual fibre” Ethernet Network Service is intended to emulate a physical medium or “wire”. It is typically realized as a point-to-point constant-bit rate (CBR) transport capability that extends the reach of a defined Ethernet PHY (IEEE 802.3) [20][21]. Ethernet frames are mapped into the (circuit/photonic) optical channels via GFP [22] and the resulting connection is complemented with mechanisms from the underlying optical layer network for switching, OA&M and trail protection, including physical layer extensions for link fault detection, performance management and client signal fail (when the emulated physical layer lacks such features). Products that address this application space fall within what is referred to as Multi-Service Transport Platforms (MSTP)/Multi-Service Provisioning Platforms (MSPP) market segment and they include network elements based on: a) Ethernet-over-WDM (EoW)
Figure 2 illustrates the high-level functional components in EoW and EoS based NEs addressing the “virtual fibre” application space. Here, ODU/OTU refers to the Optical Data/Transport Units of the OTN hierarchy.
Figure 2: High-Level NE Model for Ethernet Transport over WDM and TDM Ethernet as Switched Network Infrastructure Network elements in this application space deliver “virtual circuit” based connections re-quired to implement an Ethernet-oriented transport network infrastructure. The functional blocks used to implement these systems use basic EoW/EoS capabilities to implement its network links, while the Ethernet switching components are modeled after IEEE Ethernet Bridging [5] complemented with ITU-T requirements for Ethernet-enabled transport network equipment [21]. A “virtual link” Ethernet Transport Services is intended to emulate a “shared” or “fraction-al” link with bursty transmission capabilities. It is typically realized as a point-to-point or point-to-multipoint variable-bit rate (VBR) packet-oriented transport capability that pro-vides differentiable level of services independent of the access media speed. Hence, basic Ethernet transmission capabilities are complemented with Ethernet networking me-chanisms. NEs targeting this application space are expected to be managed consistent with evolving architectural and operational models for Ethernet-oriented transport network [21]. Products that address this application space fall within what is referred to as the NG Multi-Service Provisioning Platforms (NG MSPP)/ Optical Packet Transport System (OPTS) market segments and they include network elements based on: a) Ethernet-over-Fibre (EoF) Figure 3 high-level functional components in hybrid WDM/TDM/Packet NEs addressing this “virtual link” application space. Ethernet Connectivity Services Network elements in this application space reuse “virtual fibre” and “virtual link” intercon-nect capabilities in order to provide end-to-end connections for Enterprise data network-ing applications. Ethernet connectivity services are based on Metro Ethernet Forum’s E-Line and E-LAN service definitions [10], use the IEEE 802.1 MAC frame format as the basis for the services and require support of a number of customer oriented Ethernet Service Attributes [11]. MEF’s Ethernet connectivity services framework enable for a variety of enterprise data networking services based on two fundamental connectivity constructs:
Network elements that address this application space fall within what is designed as either Carrier Ethernet Switch/Routers (CESR) for the datacom centric market or Optical Packet Transport Systems (OPTS) market segment with integrated Packet, TDM and WDM fabrics for the telecomm centric market. Addressing Ethernet as a Service Interface Network elements in this application space address a wide class of interconnect services that use both Ethernet and other high-end technologies to support value-added Residen-tial, Enterprise or Carrier applications. They typically involve, as a minimum, the use of an MEF specified service towards the customer and the extension of this “service” interface with higher-layer IP functions typically delivered as part of an IP/MPLS service edge device (see Figure 1). A combination or optical, Ethernet or MPLS transport technologies may be used for the purposes of extending the reach of this service interface. For instance,
Network elements that address this application space includes products in the CESR or OPTS market segment complemented with Layer 3 (IP/MPLS) functionality intended to facilitate interworking with the service –aware, carrier grade, edge routers/switches. Key Components of the Ethernet Transport Solutions This section discusses the scope of Ethernet Transport capabilities under various network implementation approaches and highlights distinctive features of relevance to the usage scenarios. In particular, the minimal set of native Ethernet Transport capabilities required and the suitable complements at either Layer 1 (optical or photonic) or Layer 2 based on T-MPLS/TMPLS-TP are identified. Ethernet-over-Fibre Approach Ethernet-over-Fibre refers here to the use of native Ethernet networking capabilities, as specified by IEEE 802.1, as the packet networking layer. The foundation of EoF networking is the Ethernet MAC frame, including Virtual LAN (IEEE 802.1q) for service instance demarcation at the UNI and Provider Bridging (IEEE 802.1ad) for switching/networking. Native Ethernet networking is highly optimized for Enterprise applications, and hence, it is the base specification for the Ethernet services by the MEF [10]. OAM & Resiliency Key EoF OA&M components covering connectivity verification, defect indication, alarm suppression and performance monitoring for link, connection and subscriber maintenance entities (MEs) have been jointly specified by ITU-T SG13 and IEEE 802.1 WGs under IEEE 802.1ag/ITU-T Y.1731 specifications. These OA&M facilities may be optionally complemented with two additional facilities to for remote management of Ethernet CPEs: IEEE 802.3ah also referred to as “(client) link OAM”. Native Ethernet survivability mechanisms are primary based on shared link protection techniques such a Link Aggregation (IEEE 802.3ad) or dynamic restoration procedures via xSTP (IEEE 802.1d/s). Carrier-grade protection facilities for native Ethernet are a recent standardization activity. Shared sub-network (for either link or node failure) and ring protection mechanisms are being developed in ITU-T under recommendations G.8031/G.8032. Solution Scope As a carrier technology native Ethernet is used widely used in access and aggregation networks for business and residential services, primarily in densely populated metropolitan areas with abundant fiber availability. For regional, national or international services, carrier Ethernet technology need be further complemented with a) IEEE 802.1ah (Provider Backbone Bridging - PBB) which enables better MAC address and service instance scalability, as well as b) other connection oriented packet transport capabilities for traffic engineering and QoS management, such as MPLS (EoM). PBB Traffic Engineering (PBB-TE) is the evolving IEEE specification (IEEE P802.1Qay) that addresses connection-oriented point-to-point transport and traffic engineering for native Ethernet. As an example, an emerging approach support massive Ethernet service deployment is to combine the so called MAC-in-MAC functionality from PBB (at the edge of a metror regional network domain) with the TE capabilities of T-MPLS/MPLS-TP based metro/regional core transport network. Here, Provider Bridging technology is used as a the internal interface between native Ethernet-centric access/metro networks, while Provider Backbone Bridge technology is used as part of the adaptation process into an MPLS Virtual Private LAN Service (VPLS) [24][25]. This approach facilitates isolation of the customer MAC address space at the network interconnect points. This combination of PBB and VPLS is especially attractive for consolidation of pre-existing metro networks and interoperability with “legacy” nodes. Figure 4 illustrates a typical deployment framework for Ethernet Transport Services using EoF.
Ethernet-over-Optical Approach Ethernet-over-Optical is used here to refer to hybrid packet/photonic or packet/circuit solutions based on combination of native Ethernet capabilities with either WDM/OTN (EoW) or SONET/SDH (EoS) technology. This approach emerged from the need to ad-dress high-bandwidth content distribution applications (1GbE level and above) such as IP TV, broadband video or corporate data center interconnect while providing compatibility with existing SONEt/SDH and WDM networks. EoW can be deployed in either Coarse or Dense Wavelength Division Multiplexer (CWDM/DWDM) flavors depending on fiber reach, traffic volume and facility cost. Similarly, EoS comes with a large variety of speeds, i.e., from 1Mbps (E1/DS1) to 40Gbps (OC768/STM192), and reach options, i.e., from 10m to over 100km. EoW and EoS are anchored on a set of standards purposely developed for transport network operators. These capabilities include: - Generic Framing Procedure (GFP/ITU-T G.7041) for mapping Ethernet MAC frames PDH or SDH virtual containers. - Dynamic bandwidth allocation, ITU-T G.7042/G.7043 (SDH/PDH) - Virtual Concatenation of transport links, ITU-T G.707/G.8040 (SDH/PDH) Hence, sub-lambda multiplexing can be supported either at the packet level, circuit or WDM level. OAM & Resiliency EoS and EoW are purposely designed to reuse transport network facilities for OA&M, survivability and network operations, making it an ideal vehicle for initial introduction of carrier-grade Ethernet services. These capabilities include native continuity check, con-nectivity verification and resiliency mechanisms at either the optical [16] or the photonic [18] layers. In addition, GFP client-defect management facilities [22] help address gaps in client signal defect propagation across the service provider network. Linear and/or ring protection options at the photonic and optical level are also available to complement EoF currently more limited protection options. However, these capabilities are only applicable to link protection scenarios, not path/segment protections scenarios. For that level of protection, or for automated restoration, EoS and EoW need be complemented with additional sub-lambda grooming capabilities, such as from EoM. Solution Scope At the application level, EoS/EoW based Ethernet Private Lines are an examples of premium Ethernet WAN transport services offering very low latency, high resiliency, low jitter, and high QoS typically associated with optical transport approaches. Incumbent carriers are representative providers of EoS/EoW based Ethernet services, including all RBOCs and PTTs, with services available in leading metropolitan centers. Carriers typi-cally deploy EoS using legacy SONET/SDH equipment or NG MSPPs/MSTPs which are also capable of handling point-to-point and multipoint Ethernet connectivity services. When this approach is used, Service Providers also adapt their service management platforms to manage Ethernet based links and associated facility identifiers (e.g., VLAN tags)
Figure 5 illustrates a typical deployment framework for Ethernet Transport Services using EoW when complemented with native Ethernet bridging in the access portion of a network and MPLS in the metro/regional core. Figure 6 illustrates a equivalent deploy-ment framework for Ethernet Transport Services using EoS. The choice of packet, circuit or photonic based sub-lambda grooming is dependent on volume of local vs. transit traffic and transport facility cost.
Ethernet-over-MPLS (EoM) Transport Multi-Protocol Label Switching (T-MPLS) [14] and MPLS Transport Profile (MPLS-TP) [15] are emerging approaches to deliver carrier grade forwarding and OAM&P capabilities, including traffic engineering, for packet-oriented services, such as Ethernet Transport Services. T-MPLS/MPLS-TP is envisioned as a “profile” of the pre-existing MPLS tools most applicable to a transport network operator environment. It is consistent with IETF’s MPLS Architecture and Packet formatting (encapsulation) procedures [12][13]. Yet, T-MPLS/MPLS-TP is also focused on packet transport applications that adhere to ITU-T layer network architecture principles. That mandates a number a network operator design principles, including: 1. Support for transparent transport of a variety of client traffic types, 2. Support of minimal set of protection switching and operation & maintenance (OAM) functions optimized for long-lived connections, and 3. Ability to operate independently of its clients and its associated management and control networks (i.e., MCN and SCN), MPLS offers point-to-point and a point-to-multipoint service constructs that can be used to deliver both E-LINE, E-LAN and E-TREE type services. It also offers native bridge Ethernet service emulation via Virtual Private LAN Service (VPLS). OAM & Resiliency EoM offers native continuity check, connectivity verification, performance monitoring and resiliency mechanisms based on BFD [26] and soon to be complemented with of G.8114-like [15] OA&M capabilities. Service protection options are designed to meet 50 ms or less and very low latency [15]. Among the OAM & protection capabilities that EoM brings to bear for a complete transport solutions are linear access link protection based on MPLS Fast Reroute [27] and G.8131 (1:1 or 1+1) [33], plus dual-homing with linear access protection (1:1) based on Dual Node Interconnect and dual-homing with ring access protection based G.8132 shared-protection ring (SPRing) planned under the MPLS TP framework. Solution Scope EoM has been the leading approach for carrier Ethernet implementations in access and aggregation networks, whether business or residential services, where service scaling and traffic management has been a key concern. Key component of the solution is the emulation of Ethernet bridging via VPLS/Hierarchical VPLS [24][25], combined with diff-serv traffic engineering capabilities for QoS management [28] and comprehensive industry support for pipe or hose based bandwidth profile models [29][30]. Figure 7 illustrates the end-to-end EoM framework for Ethernet Transport Services. Footnote: Some aspects of MPLS Fast Reroute operations will likely require further enhancements particularly for environments where control plane protocols are not suppored.
As noted before, for regional, national or international services it can be useful to util-ize VPLS emulation capabilities with IEEE 802.1ah MAC-in-MAC encapsulation where MAC address table scalability is an issue. It is also plausible to envision scenarios, either for network transition or other engineering tradeoffs, where Provider Bridging technology is used as the internal interface between native Ethernet-centric access/metro networks, while EoM is used in the metro/regional portion of the network. Figure 8 illustrates a typical deployment framework for Ethernet Transport Services using a combined EoF/EoM approach.
Closing Remarks Transport networks are in the process of undertaking a substantial transformation to better cater to the dominance of packet oriented applications as their main traffic source. This network transformations is likely to lead to the integration of WDM, TDM and packet transport technologies into network element equipment allowing efficient and cost effective solutions addressing customers and service providers need in the highly competitively telecommunications market. Packet transport will in all likelihood merge the best aspects of Ethernet and MPLS based transport capabilities to deliver such converged transport network expectations. Optical/Packet transport solutions that flexibly deliver Ethernet-oriented services with the resiliency and robustness of optical transport equipment and the traffic engineered capacities of T-MPLS/MPLS-TP are uniquely positioned to satisfy the reliability, resiliency, scalability, performance and manageability expectations associated with carrier-grade transport environments. References [1] Business Ethernet Services: Worldwide Market Update (MEF). Vertical System Group. Jan 28, 2008. [2] Market Alert: 4Q07 and Global 2007 Optical Networking. Ovum. March 4, 2008. [3] IEEE Std 802.3 – 2005, Information technology – Telecommunications and information exchange between systems – Local and metropolitan area networks – Specific requirements – Part 3: Carrier sense multiple access with collision detection (CSMA/CD) access method and physical layer specifications, 9 December 2005. [4] IEEE Std 802.1D – 2004, IEEE Standard for Local and Metropolitan Area Networks: Media Access Control (MAC) Bridges, June 2004. [5] IEEE Std 802.1Q – 2003, IEEE Standards for Local and metropolitan area networks— Virtual Bridged Local Area Networks, 7 May 2003. [6] IEEE Std 802.1ad – 2005, IEEE Standard for Local and Metropolitan Area Networks— Virtual Bridged Local Area Networks — Amendment 4: Provider Bridges, May 2006. [7] IEEE Project 802.1ah - Provider Backbone Bridging. See http://www.ieee802.org/1/pages/802.1ah.html [8] MEF4: Metro Ethernet Network Architecture Framework - Part 1: Generic Framework. May 2004. [9] MEF12: Metro Ethernet Network Architecture Framework: Part 2: Ethernet Services Layer [10] MEF6.1: Ethernet Services Definitions - Phase 1. June 2004. [11] MEF10.1: Ethernet Services Attributes - Phase 2. November 2006. [12] IETF RFC 3031, Multi-protocol label switching architecture. January 2001. [13] IETF RFC 3032, MPLS label stack encoding. January 2001. [14] ITU-T Recommendation G.8110, Architecture of Transport MPLS Layer Network. February 2006. [15] M. Bocci, M. Lasserre,S. Bryant, “A Framework for MPLS in Transport Networks “, IETF Internet Draft, Jul.2008, < http://tools.ietf.org/html/draft-blb-mpls-tp-framework-00>. [16] ITU-T Recommendation G.707 Network node interface for the synchronous digital hierarchy (SDH). [17] ITU-T Recommendation G.783. Characteristics of synchronous digital hierarchy (SDH) equipment functional blocks. [18] ITU-T Recommendation G.709. Interfaces for the Optical Transport Network (OTN). [19] ITU-T Recommendation G.798. Characteristics of optical transport network hierarchy equipment functional blocks. [20] ITU-T Recommendation G.8011. Ethernet over Transport - Ethernet services framework. [21] ITU-T Recommendation G.8021 Characteristics of Ethernet transport network equipment functional blocks. [22] ITU-T Recommendation G.7041. Generic Framing Procedure (GFP). [23] ITU-T Recommendation G.7042. LCAS. [24] IETF RFC 4761. Virtual Private LAN Service (VPLS) Using BGP for Auto-Discovery and Signaling. January 2007. [25] IETF RFC 4762. Virtual Private LAN Service (VPLS) Using Label Distribution Protocol (LDP) Signaling. January 2007. [26] IETF RFC4090. Fast Reroute Extensions to RSVP-TE for LSP Tunnels. May 2005, <http://tools.ietf.org/html/rfc4090>. [27] R. Aggarwal, K. Kompela, T. Nadeau, G. Swallow, “BFD for MPLS LSPs,” IETF Internet Draft, Jun. 2008, <http://tools.ietf.org/html/draft-ietf-bfd-mpls-07.txt>. [28] IETF RFC 2475. An Architecture for Differentiated Service. December 1998 [29] N.G. Duffield, P. Goyal, A. Greenberg, P. Mishra, K.K. Ramakrishnan, and J.E.V.D. Merwe, "Resource management with hoses: Point-to-cloud services for virtual private networks," IEEE/ACM Trans. Netw., vol.10, no.5, pp.679–692, 2002. [30] M. Kodialam and T.V. Lakshman, “Dynamic routing of bandwidth guaranteed multicasts with failure backup,” IEEE International Conference on Network Protocols, 2002, pp. 259– 268, Nov. 2002. [31] ITU-T Recommendation G.8113 (2006), Requirements for Operation & maintenance functionality in T-MPLS networks. [32] ITU-T Recommendation G.8110.1 (2006), Architecture of Transport MPLS Layer Network. [33] ITU-T Recommendation G.8131 (2007), Linear protection switching for transport MPLS (T-MPLS) networks. [34] ITU-T Draft Recommendation G.8114 (2007) Operation & maintenance mechanism for TMPLS layer networks. Trackback(0)
Comments
(2)
ITU-T G.8031, discusses providing linear protection mechanism, exactly like SONET/SDH linear protection mechanism. It also says 1+1 where head end bridges and tail end makes the selection. In case of 1+1 as defined in the G.8031 the traffic will be bridged at the head end and selected at the tail end, but the inherent capability of the Ethernet does not have any selection mechanism. Rather, this can be done by blocking at the head end of the transmission.
Votes: +0
for e.g. Eth#A(VLAN 10) is bridging to Eth#W(VLAN 10) and Eth#P(VLAN 10) on the head end will also receive from both Working and protect as the FDB is set as below for VLAN 10 to send data on both #W and #P. So the packets in 1+1 will be duplicated. Can someone clarify in lay man terms how this (selection at tail end) can be acheived for a VLAN10 Eth#W Eth#P FDB VLAN10 VLAN10 --- o-----------------------------------o VLAN MAC Port / 10 XXX Eth#W Eth#A o o Eth#A 10 XXX ETH#A VLAN10 / VLAN10 10 XXX Eth#B o-----------------------------------o Eth#P Eth#P VLAN10 VLAN10 report abuse
vote down
vote up
You must be logged in to a comment. Please register if you do not have an account yet.
|
| Last Updated on Tuesday, 17 March 2009 09:06 |
TrackBack URI for this entry
Subscribe to this comment's feed
May I ask a question? Regarding the use of EoM to deliver E-TREE, it is common to read that this is viable. However, to my understanding, the actual creation of E-TREE using MPLS applications is poorly supported via much of the HW available today.
Do you have any examples of existing E-TREE deployment instances supporting unicast flows?
Many thanks,
Dana