Designing and Implementing IP/MPLS-Based Ethernet Layer 2 VPN Services

An Advanced Guide for VPLS and VLL

Zhuo Xu

• Publisher: Wiley
• Year: 2010
• Print ISBN: 9780470456569
• eBook ISBN: 9780470730560
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Networking
• Index: Computer Science

A guide to designing and implementing VPLS services over an IP/MPLS switched service provider backbone

Today's communication providers are looking for convenience, simplicity, and flexible bandwidth across wide area networks-but with the quality of service and control that is critical for business networking applications like video, voice and data. Carrier Ethernet VPN services based on VPLS makes this a reality. Virtual Private LAN Service (VPLS) is a pseudowire (PW) based, multipoint-to-multipoint layer 2 Ethernet VPN service provided by services providers By deploying a VPLS service to customers, the operator can focus on providing high throughput, highly available Ethernet bridging services and leave the layer 3 routing decision up to the customer.

Virtual Private LAN Services (VPLS) is quickly becoming the number one choice for many enterprises and service providers to deploy data communication networks. Alcatel-Lucent VPLS solution enables service providers to offer enterprise customers the operational cost benefits of Ethernet with the predictable QoS characteristics of MPLS.

Items Covered:

Building Converged Service Networks with IP/MPLS VPN Technology

IP/MPLS VPN Multi-Service Network Overview

Using MPLS Label Switched Paths as Service Transport Tunnels

Routing Protocol Traffi c Engineering and CSPF

RSVP-TE Protocol

MPLS Resiliency — Secondary LSP

MPLS Resiliency — RSVP-TE LSP Fast Reroute

Label Distribution Protocol

IP/MPLS VPN Service Routing Architecture

Virtual Leased Line Services

Virtual Private LAN Service

Hierarchical VPLS

High Availability in an IP/MPLS VPN Network

VLL Service Resiliency

VPLS Service Resiliency

VPLS BGP Auto-Discovery

PBB-VPLS

OAM in a VPLS Service Network

Part I

IP/MPLS VPN Service Network Overview

Telecommunication operators must constantly evolve their networks to meet the needs of their customers. Building a converged, high performance, highly available, and highly flexible network to provide multiple services in a cost efficient way is the goal for today’s providers. The new generation of IP/MPLS VPN service-oriented networks has become the operators’ best choice to reach this goal.

Chapter 1: Building Converged Service Networks with IP/MPLS VPN Technology

Chapter 2: IP/MPLS VPN Multi-Service Network Overview

Chapter 1

Building Converged Service Networks with IP/MPLS VPN Technology

Multi Protocol Label Switching (MPLS) and Virtual Private Network (VPN) technologies provide features that help service providers meet the evolving needs of their customers. These technologies are essential for building the converged service networks required in today’s market.

Chapter Objectives

• Identify the new trends and demands for a service provider’s backbone network
• Review the evolution of MPLS technology
• Describe the innovation of multi-service VPN

This chapter briefly reviews traditional networks with legacy technologies and their limitations, and shows how the innovations of MPLS and VPN technologies overcome these limitations. It also presents the benefits of using an IP/MPLS VPN service architecture.

1.1 The Increasing Demands on Service Provider Networks

Service provider networks must evolve to keep pace with the changing times. Service providers are often classified by how much of the regional access infrastructure they own, versus how much they must contract from other providers:

• Tier 1 operators — The top one or two providers in a country who typically own the access infrastructure (copper or fiber) within their serving region. Tier 1 service providers are usually the first to establish infrastructures within the region — the incumbent operators.
• Tier 2 or Tier 3 operators — Providers that may either use the Tier 1 operator’s access infrastructure or build its own infrastructure in some service areas. Tier 2 providers use a mix of their own infrastructure and some infrastructure from Tier 1 providers, while Tier 3 providers rely entirely on agreements to use infrastructure from other providers. These providers typically emerge as competitors to the already established Tier 1 providers, and are thus at a disadvantage in competing with the incumbent providers for market control.

Service providers may also be classified according to the types of services they offer to their end-customers:

• Telco — Traditionally offering voice services as well as business services
• Internet Service Provider (ISP) — Offering Internet access for residential and business customers
• VPN Service Provider/Ethernet Service Provider — Offering business VPN services
• Cable Multi-System Operator (MSO) — Offering residential and business services

An operator may offer some or all of these services to their end-customers.

Both residential (consumer) and enterprise (business) customers of service providers constantly demand new services and innovations from their service providers. Traditional Leased Line, Frame-Relay (FR), and Asynchronous Transfer Mode (ATM) based services are characteristic of organizations that manage their own enterprise networks (with their own IT teams), but those enterprises must purchase the connectivity infrastructure (typically point-to-point leased lines or FR/ATM Permanent Virtual Connections) from a service provider. Driven by enterprise business goals and geared toward focusing on core competencies and cost reduction, enterprises have begun looking to service providers for managed connectivity solutions.

Enterprises have also been demanding more in terms of bandwidth speeds for connectivity. The old “80/20 rule” (80% of the traffic stays within the local site, and 20% of the traffic is between remote sites) is no longer valid. Because many enterprises have consolidated their data centers to a few sites, the need for higher-speed remote connectivity has become extremely important to enterprise IT managers. In addition, enterprises are now in the process of implementing bandwidth-intensive applications like video conferencing, web conferencing, and electronic image sharing across a wide area, thus prompting a need for additional bandwidth in their Wide Area Networks (WANs).

Residential services are also evolving from dial-up Internet connectivity to broadband connectivity. Services for residential customers are evolving to include triple- or quad-play services that include voice, Video on Demand (VoD), broadcast television, and Internet access.

Traditionally a service provider has separate networks for offering voice and data services. Within a data network, a traditional service provider would typically have separate networks for offering Leased Line-, FR-, and ATM-based services for business customers and a separate network offering Internet-based services (Internet access and Internet-based secure connectivity) for residential and business customers. In residential areas, TV content for consumers is most often delivered by MSOs, who have their own dedicated infrastructure (mostly cable plants). Enterprises usually use Ethernet switches and IP routers to build their LANs and purchase Leased Line services from operators to connect their remote locations.

Given the ever-changing landscape of customer demands, service provider networks must keep pace by staying competitive while increasing profitability. It is evident that the approach of building separate networks is not cost-effective when a service provider must offer multiple services. The ideal way to approach network design is a solution wherein multiple services can be converged on a single network infrastructure. This is why MPLS as a technology for service provider networks has gained rapid momentum in the marketplace.

The most obvious trend is the fast growth of IP and Ethernet traffic in the network. Because of the boom of the Internet, and the invention of Gigabit Ethernet, IP/Ethernet traffic is now dominant in telecommunication networks. Residential customers require faster Internet access services and better IP service quality to support Voice over IP (VoIP). Enterprise customers are conducting more and more of their business electronically across geographically separated locations. Many bandwidth-intensive and time-sensitive IP-based applications are widely used for business-critical missions. IP data is growing in strategic importance in wireless networks. Mobile users are keen for rich IP-based multimedia services. Service providers also want to deliver television content over IPTV applications, which require a network throughput with very high bandwidth and low latency. It’s clear that building a network optimal for IP/Ethernet traffic delivery is crucial to service providers.

Because enterprises are now starting to use more and more IP/Ethernet-based applications, they require their IT infrastructures to have high throughput, and to be reliable, secure, and cost-efficient. This generates a great demand for the service providers to provide VPN. VPN allows the service provider to deliver services to different customers using the same service delivery backbone network, while isolating each customer using virtual service instances to ensure privacy and security. During the past two decades, there were already many enterprises using the routed RFC2547bis VPN to achieve intranet connectivity. Now, with the fast growth of Ethernet technology, more and more business customers require bridged Layer 2 Ethernet VPN service. Layer 2 VPN gives the customers full control of their routing domains and fewer peering complications with service providers.

Service providers also look for network solutions that consolidate voice, data, and video services into one network infrastructure and allow them to serve residential and business customers from the same network. The network must be cost-efficient and robust. The network must also be capable of providing different Quality of Service (QoS) on the service provided to conform to different Service Level Agreements (SLAs).

With these new trends and demands, service providers intend to transition their networks to IP/MPLS core networks, providing various VPN services to their customers.

1.2 MPLS Overview

Multi Protocol Label Switching is a label-switching mechanism used by MPLS-capable routers or switches to exchange traffic. In the control plane, the MPLS-capable devices assign labels to be used for certain types of traffic and distribute labels through certain label distribution protocols. Each device distributes locally assigned labels to other MPLS devices and receives label distribution information from other devices. Each device builds a Label Information Base (LIB) that stores the label information. In the data plane, each device performs MPLS encapsulation on data traffic before sending it to other MPLS devices. When an MPLS device receives MPLS-encapsulated traffic, the device makes forwarding decisions based on the MPLS label value in the MPLS encapsulation header. In MPLS data encapsulation, the MPLS header (32 bits long, containing a 20-bit numerical value used as the label value) is inserted between the Layer 2 header and the Layer 3 header of the data to be encapsulated. Therefore, MPLS is sometimes referred as a Layer 2.5 protocol, and the MPLS header is sometimes referred to as the shim header.

Before MPLS devices can forward MPLS-encapsulated traffic to each other, MPLS label distribution in the control plane must be completed. When exchanging label information, each MPLS device stores the label, as well as the label mapping information for the type of traffic that uses each label. All traffic that uses the same label is referred to as a Forwarding Equivalent Class (FEC). The label distribution process distributes the FEC–Label mapping information among MPLS devices. Therefore, MPLS devices form a Label Switched Path (LSP) for each FEC. The LSP is an end-to-end connection for traffic belonging to the same FEC to be forwarded. MPLS builds a connection-oriented path in a connectionless network.

MPLS was first introduced to improve Layer 3 routing performance of regular IP routers. For an MPLS-capable router or Layer 3 switch, MPLS label swapping is less expensive than routing IP packets. In a routed IP network, the IP packets are routed from their source to their destination hop-by-hop. When each router routes an IP packet, the router removes the Layer 2 header (usually an Ethernet header), then checks the IP header for the destination IP address. The router then must perform a lookup in its routing table to find the IP address of the next-hop interface and the egress interface’s Layer 2 encapsulation information. After the next-hop lookup is completed, the router rewrites the packet by adding the new Layer 2 encapsulation header to the packet and then forwards the packet to the next-hop interface. This procedure is performed for every IP packet at every hop. With the introduction of MPLS, the routers can build MPLS LSPs for each FEC. All traffic belonging to the same FEC is MPLS-label–switched to its destination rather than routed. When a Label Switched Router (LSR) performs MPLS switching on an MPLS-encapsulated packet, the MPLS label-swapping operation is much simpler. Therefore, the IP destination lookup process is replaced by the relatively cheaper label-swapping process. Using MPLS switching to replace IP routing is sometime referred to as a routing shortcut.

Furthermore, Border Gateway Protocol (BGP) can be removed from the core of the network because the LSR routers in the core of the network do not have to route these packets. As long as the MPLS label distribution process builds the LSP for each router in the core to reach all edge routers that have BGP peerings with routers outside the Autonomous System (AS), traffic across the core network can be MPLS-switched rather than IP-routed. The core router uses the MPLS label to switch the traffic to the correct edge router. BGP full mesh within the AS can be removed. Only the edge routers need to have BGP peering among each other. Using MPLS switching to remove BGP full mesh from the core network to route Internet traffic is sometimes referred to as a BGP shortcut. The label distribution process used by traditional MPLS-capable devices is in most cases the Label Distribution Protocol (LDP).

1.3 The MPLS Value Proposition

MPLS has evolved substantially since its early days of deployment. The reasons for using MPLS in a network have also changed. MPLS is no longer used to provide an IP routing shortcut. The two biggest changes in the MPLS technology are:

• Resource Reservation Protocol (RSVP) is extended to support MPLS label distribution — RSVP-TE. RSVP-TE (the TE stands for traffic engineering) brings many traffic engineering features and resiliency features to MPLS tunneling technology.
• Pseudowire (PW)-based MPLS L2VPN is implemented in many vendors’ MPLS-capable routers and switches.

With these evolutions in MPLS technology, MPLS is now widely deployed in the backbone networks of service providers to provide VPN services to their customers.

The introduction of RSVP-TE into MPLS label distribution gives MPLS outstanding flexibility and reliability that the traditional routed or switched network cannot have:

• MPLS provides traffic engineering capabilities to control the data forwarding...

Table of contents

• Cover
• Table of Contents
• Praise
• Title Page
• Copyright
• Dedication
• About the Author
• Credits
• Acknowledgments
• Foreword
• Introduction
• Part I: IP/MPLS VPN Service Network Overview
• Part II: IP/MPLS VPN Protocol Fundamentals
• Part III: Ethernet VPN Services
• Part IV: Advanced Ethernet VPN Topics
• Appendix A: Spanning Tree Protocol
• Appendix B: RFC and IEEE Standards
• Glossary
• Index