Mapping VPN Traffic onto LSPs Using Route Policies

Similar documents
Network Configuration Example

DD2491 p MPLS/BGP VPNs. Olof Hagsand KTH CSC

Demonstrating the high performance and feature richness of the compact MX Series

RFC 2547bis: BGP/MPLS VPN Fundamentals

Basic Configuration Examples for BGP

Using OSPF in an MPLS VPN Environment

Monitoring and Troubleshooting BGP Neighbor Sessions

MPLS VPN over mgre. Finding Feature Information. Prerequisites for MPLS VPN over mgre

CLOS IP FABRICS WITH QFX5100 SWITCHES

MPLS VPN Route Target Rewrite

Network Configuration Example

MONITORING NETWORK TRAFFIC USING sflow TECHNOLOGY ON EX SERIES ETHERNET SWITCHES

How To Set Up Bgg On A Network With A Network On A Pb Or Pb On A Pc Or Ipa On A Bg On Pc Or Pv On A Ipa (Netb) On A Router On A 2

Monitoring Network Traffic Using sflow Technology on EX Series Ethernet Switches

Network Configuration Example

BGP Best Path Selection Algorithm

How To Make A Network Secure

Network Configuration Example

Understanding Virtual Router and Virtual Systems

MPLS Layer 2 VPNs Functional and Performance Testing Sample Test Plans

MPLS-based Layer 3 VPNs

Internetworking II: VPNs, MPLS, and Traffic Engineering

HP Networking BGP and MPLS technology training

Cisco Configuring Basic MPLS Using OSPF

IP/MPLS-Based VPNs Layer-3 vs. Layer-2

MPLS/BGP Network Simulation Techniques for Business Enterprise Networks

Introducing Basic MPLS Concepts

IPv6 over IPv4/MPLS Networks: The 6PE approach

Table of Contents. Cisco Configuring a Basic MPLS VPN

MPLS-based Virtual Private Network (MPLS VPN) The VPN usually belongs to one company and has several sites interconnected across the common service

Introduction to MPLS-based VPNs

Configuring a Load-Balancing Scheme

SRX High Availability Design Guide

Juniper Exam JN0-343 Juniper Networks Certified Internet Specialist (JNCIS-ENT) Version: 10.1 [ Total Questions: 498 ]

Examination. IP routning på Internet och andra sammansatta nät, DD2491 IP routing in the Internet and other complex networks, DD2491

MikroTik RouterOS Introduction to MPLS. Prague MUM Czech Republic 2009

Quidway MPLS VPN Solution for Financial Networks

PRASAD ATHUKURI Sreekavitha engineering info technology,kammam

How Routers Forward Packets

Department of Communications and Networking. S /3133 Networking Technology, Laboratory course A/B

DD2491 p BGP-MPLS VPNs. Olof Hagsand KTH/CSC

Fast Re-Route in IP/MPLS networks using Ericsson s IP Operating System

APNIC elearning: Introduction to MPLS

JNCIA-Junos Study Guide Part 2

APPLICATION NOTE 211 MPLS BASICS AND TESTING NEEDS. Label Switching vs. Traditional Routing

Enterprise Network Simulation Using MPLS- BGP

Network Configuration Example

MPLS Architecture for evaluating end-to-end delivery

JNCIE Juniper Networks Certified Internet Expert

Tackling the Challenges of MPLS VPN Testing. Todd Law Product Manager Advanced Networks Division

OVERLAYING VIRTUALIZED LAYER 2 NETWORKS OVER LAYER 3 NETWORKS

Understanding Route Redistribution & Filtering

Configuring MPLS Hub-and-Spoke Layer 3 VPNs

ENSURING RAPID RESTORATION IN JUNOS OS-BASED NETWORKS

WAN Topologies MPLS. 2006, Cisco Systems, Inc. All rights reserved. Presentation_ID.scr Cisco Systems, Inc. All rights reserved.

Junos MPLS and VPNs (JMV)

Kingston University London

Analyzing Capabilities of Commercial and Open-Source Routers to Implement Atomic BGP

MPLS VPN Services. PW, VPLS and BGP MPLS/IP VPNs

APNIC elearning: BGP Attributes

Introduction Inter-AS L3VPN

Junos OS. MPLS Configuration Guide for Security Devices. Release Published: Copyright 2012, Juniper Networks, Inc.

For internal circulation of BSNLonly

Traffic Engineering Management Concepts

MPLS Virtual Private Networks

Network Configuration Example

MPLS Implementation MPLS VPN

How To Understand Bg

MPLS is the enabling technology for the New Broadband (IP) Public Network

IPv6 over MPLS VPN. Contents. Prerequisites. Document ID: Requirements

BGP FORGOTTEN BUT USEFUL FEATURES. Piotr Wojciechowski (CCIE #25543)

IP interconnect interface for SIP/SIP-I

Computer Network Architectures and Multimedia. Guy Leduc. Chapter 2 MPLS networks. Chapter 2: MPLS

Configuring a Basic MPLS VPN

IMPLEMENTING CISCO MPLS V3.0 (MPLS)

MPLS L2VPN (VLL) Technology White Paper

WHITEPAPER. Bringing MPLS to Data Center Fabrics with Labeled BGP

MPLS Basics. For details about MPLS architecture, refer to RFC 3031 Multiprotocol Label Switching Architecture.

S ITGuru Exercise (3: Building the MPLS BGP VPN) Spring 2006

Network Configuration Example

Tutorial: Options for Blackhole and Discard Routing. Joseph M. Soricelli Wayne Gustavus NANOG 32, Reston, Virginia

Technology Overview. Lawful Intercept Using Flow-Tap. Published: Copyright 2014, Juniper Networks, Inc.

Exterior Gateway Protocols (BGP)

- Multiprotocol Label Switching -

Course Description. Students Will Learn

Fault Management In MPLS Networks

Expert Reference Series of White Papers. An Overview of MPLS VPNs: Overlay; Layer 3; and PseudoWire

DDoS Mitigation Techniques

Juniper / Cisco Interoperability Tests. August 2014

Implementing Firewalls inside the Core Data Center Network

Inter-domain Routing Basics. Border Gateway Protocol. Inter-domain Routing Basics. Inter-domain Routing Basics. Exterior routing protocols created to:

Implementing MPLS VPNs over IP Tunnels

netkit lab MPLS VPNs with overlapping address spaces 1.0 S.Filippi, L.Ricci, F.Antonini Version Author(s)

basic BGP in Huawei CLI

Implementing Cisco MPLS

MPLS Layer 3 and Layer 2 VPNs over an IP only Core. Rahul Aggarwal Juniper Networks. rahul@juniper.net

MPLS in Private Networks Is It a Good Idea?

Transcription:

Mapping VPN Traffic onto LSPs Using Route Policies Test Case February 2017 Version 1.0

Juniper Networks, Inc. 1133 Innovation Way Sunnyvale, California 94089 USA 408-745-2000 www.juniper.net Juniper Networks assumes no responsibility for any inaccuracies in this document. Juniper Networks reserves the right to change, modify, transfer, or otherwise revise this publication without notice. The information in this document is current as of the date on the title page. Copyright 2017, Juniper Networks, Inc. All rights reserved. ii Copyright 2017, Juniper Networks, Inc.

Contents Overview... 1 Test Network... 2 Test Scenarios for Placing VPN Traffic onto LSPs... 4 Default Forwarding Behavior... 4 Initial Configuration... 4 Verification... 5 Scenario 1 Mapping VPNs to Specific LSPs... 7 Configuration... 7 Verification... 8 Scenario 2 Mapping Specific Destinations to Specific LSPs... 9 Configuration... 9 Verification... 10 Additional Tests... 12 Non-matching Traffic... 12 Forwarding Behavior When LSP is Down... 12 The Strict Option... 13 Appendix A PE 11 Configuration... 14 Interfaces Hierarchy... 14 Protocols OSPF Hierarchy... 15 Protocols BGP Hierarchy... 15 Protocols RSVP Hierarchy... 16 Routing Instances Hierarchy... 16 Copyright 2017, Juniper Networks, Inc. iii

iv Copyright 2017, Juniper Networks, Inc.

Overview Many enterprise customers are migrating towards using Multiprotocol Label Switching (MPLS) as a core technology for interconnecting data centers. MPLS provides traffic engineering capabilities; the ability to establish logical paths [Label Switched Paths (LSPs)] through the network based on different constraints such as bandwidth, link color, and priority while protecting traffic in the event of link or node failures using fast re-route mechanisms. The enterprise s IT department can also use the core to offer MPLS based Layer 2 and Layer 3 VPN services to their internal customers (to different company departments) and external partners. Based on business requirements, you can map traffic flows onto the specific LSPs based on various criteria such as VPN, destination IP address, or class of service (CoS). To explain this functionality, a simplified MPLS lab network was constructed, and different scenarios were configured and tested. After examining the default forwarding behavior of VPN traffic by a provider edge (PE) router, route policies were configured to map specific VPN traffic to specific LSPs, and then modified to map traffic to specific destinations within a VPN to specific LSPs. When the incoming traffic did not match any terms in the route policy, and when the referenced LSP was inactive, additional tests were performed to determine the forwarding behavior. This document describes the configuration and verifies the functionality of using route policies to place VPN traffic onto specific LSPs. Copyright 2017, Juniper Networks, Inc. 1

Test Network Figure 1 shows two data centers (DC1 and DC2) interconnected by a pair of MX Series routers forming an IP/MPLS core. This core network offers the following services: Two IP VPNs, Blue and Red, each with two sites One Layer 2 VPLS VPN, Grey, with two sites A single OSPF area is configured to establish connectivity between all core devices and interfaces. Multiprotocol IBGP is configured to exchange Network Layer Reachability Information (NLRI) for the various VPNs within the core with routers 12 and 21 acting as route reflector (RR) servers. Provider edge (PE) routers 11 and 21 have a Layer 3 EBGP connection to customer edge (CE) routers 13 and 23 for the IP VPNs, and Layer 2 connectivity to CEs 14 and 24 for the VPLS service. Figure 1 Test Network Architecture RSVP is used in the core to establish the following LSPs from Router 11 to 21: 1. from-11-to-21-fast-1 2. from-11-to-21-fast-2 3. from-11-to-21-slow-3 2 Copyright 2017, Juniper Networks, Inc.

The fast and slow represent the relative speed of the links. In this network, the inter-data center WAN link speeds between routers 11 and 21 are faster than the connections between routers 12 and 22 (see Figure 2). Figure 2 Inter Data Center Link Speeds Other LSPs are configured on PE router 21 to enable forwarding of VPN traffic between PEs. However, the focus of the testing is on PE 11 with traffic flows between CE 13 to 23, and CE 14 to 24 that utilize the three LSPs previously listed. All devices were running Junos OS Release 10.4R3, and all core interfaces were on line cards containing the Trio chipset. Copyright 2017, Juniper Networks, Inc. 3

Test Scenarios for Placing VPN Traffic onto LSPs Default Forwarding Behavior Initial Configuration This initial configuration focuses on the portion most relevant to the tests. For additional configuration details for PE 11, see Appendix A PE 11 Configuration. Use the protocols mpls hierarchy to define the LSPs from PE 11 to PE 21, and their paths: label-switched-path from-11-to-21-fast-1 { from 10.0.0.11; to 10.0.0.21; primary path-11-to-21-1; label-switched-path from-11-to-21-fast-2 { from 10.0.0.11; to 10.0.0.21; primary path-11-to-21-2; label-switched-path from-11-to-21-slow-3 { from 10.0.0.11; to 10.0.0.21; primary path-11-to-21-3; path path-11-to-21-1 { 10.12.1.21; path path-11-to-21-2 { 10.12.2.21; path path-11-to-21-3 { 10.0.0.12; Use the policy-options hierarchy to create a policy to load-balance traffic, and use the routing-options hierarchy to apply the policy: Note: This routing policy is evaluated when routes are exported from the routing table into the forwarding table. policy-statement load-balance-per-packet { term 1 { load-balance per-packet; routing-options { router-id 10.0.0.11; autonomous-system 64512; forwarding-table { export load-balance-per-packet; 4 Copyright 2017, Juniper Networks, Inc.

Verification On PE 11, viewing the route in the IP-VPN-Blue table to destination 10.2.231.0/24 on CE 23 shows that there are three equal cost paths as the next-hop towards the destination. user@pe11-re0> show route 10.2.231.0/24 table IP-VPN-Blue IP-VPN-Blue.inet.0: 15017 destinations, 25027 routes (15017 active, 0 holddown, 0 hidden) Restart Complete + = Active Route, - = Last Active, * = Both 10.2.231.0/24 *[BGP/170] 04:00:53, localpref 100, from 10.0.0.21 AS path: 65521 I to 10.12.1.21 via xe-2/1/0.0, label-switched-path from-11-to-21-fast-1 to 10.12.2.21 via xe-2/2/0.0, label-switched-path from-11-to-21-fast-2 > to 10.1.1.12 via xe-2/0/0.0, label-switched-path from-11-to-21-slow-3 [BGP/170] 04:00:53, localpref 100, from 10.0.0.12 AS path: 65521 I to 10.12.1.21 via xe-2/1/0.0, label-switched-path from-11-to-21-fast-1 to 10.12.2.21 via xe-2/2/0.0, label-switched-path from-11-to-21-fast-2 > to 10.1.1.12 via xe-2/0/0.0, label-switched-path from-11-to-21-slow-3 Although the output shows that only one of the LSPs as selected (indicated with the > symbol), the forwarding table shows that all three paths are active: user@pe11-re0> show route forwarding-table destination 10.2.231.0/24 table IP-VPN-Blue Routing table: IP-VPN-Blue.inet Internet: Destination Type RtRef Next hop Type Index NhRef Netif 10.2.231.0/24 user 0 indr 1048588 3103 ulst 1048598 3 10.12.1.21 Push 16 801 1 xe-2/1/0.0 10.12.2.21 Push 16 880 1 xe-2/2/0.0 10.1.1.12 Push 16, Push 300160(top) 881 1 xe-2/0/0.0 The equal cost LSPs are also present in the inet.3 table. The following output shows the paths used to reach the egress PE s IP address, for this example, PE 21. There are three LSPs with a metric of 1, and as a result, all of them are used to forward traffic. user@pe11-re0> show route 10.0.0.21 table inet.3 extensive inet.3: 4 destinations, 4 routes (4 active, 0 holddown, 0 hidden) Restart Complete 10.0.0.21/32 (1 entry, 1 announced) State: <FlashAll> *RSVP Preference: 7/1 Next hop type: Router Next-hop reference count: 13 Next hop: 10.12.1.21 via xe-2/1/0.0 weight 0x1 Label-switched-path from-11-to-21-fast-1 Next hop: 10.12.2.21 via xe-2/2/0.0 weight 0x1 Label-switched-path from-11-to-21-fast-2 Next hop: 10.1.1.12 via xe-2/0/0.0 weight 0x1, selected Label-switched-path from-11-to-21-slow-3 Label operation: Push 300112 Label TTL action: prop-ttl State: <Active Int> Local AS: 64512 Copyright 2017, Juniper Networks, Inc. 5

Age: 17:19:35 Metric: 1 Task: RSVP Announcement bits (2): 1-Resolve tree 1 2-Resolve tree 2 AS path: I Using the tester (shown in Figure 3), ten streams were transmitted on the Blue VPN from CE 13 to destination subnet 10.2.231.0/24 on CE 23 with a total rate of 10000 packets/seconds (pps). Monitoring the interface statistics showed that the flows are load-balanced among the three interfaces corresponding to the three LSPs. Figure 3 Tester (Connection Manager) Additionally, the MPLS LSP statistics show the active load-balancing behavior: {master user@pe11-re0> clear mpls lsp statistics {master user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21-* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 7953 72833574 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 3977 36421366 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 7955 72851890 from-11-to-21-slow-3 Total 3 displayed, Up 3, Down 0 6 Copyright 2017, Juniper Networks, Inc.

Scenario 1 Mapping VPNs to Specific LSPs Configuration In this example, a routing policy is created with traffic flows mapped from DC1 to DC2: Layer 2 VPN Grey traffic maps to LSP from-11-to-21-fast-2 Layer 3 VPN Blue traffic maps to LSPs from-11-to-21-fast-1 and from-11-to-21-fast-2 Layer 3 VPN Red traffic maps to LSP from-11-to-21-slow-3 Each VPN is configured using the same VRF target parameter at all locations (see Table 1). This means that PE 21 advertised NLRI information with equivalent targets to PE 11. For additional configuration for each routing instance on PE 11, see Appendix A PE 11 Configuration. Table 1 VPN and VRF Route Target Information VPN Grey Blue Red VRF Route Target target:64512:10 target:64512:1 target:64512:2 When creating the route policy, this VRF target information is used as match criteria to uniquely identify the VPN to which the incoming traffic belongs. The action of the route policy is to map the traffic onto a specific LSP(s). Note: For the policy action to take effect, the specified next-hop LSP must be an active and selected path to the destination PE. Use the policy-options hierarchy to configure the route policy: policy-statement MapVpnToLsp { term 1 { from community vpls-grey; install-nexthop lsp from-11-to-21-fast-2; term 2 { from community vpn-blue; install-nexthop lsp [ from-11-to-21-fast-1 from-11-to-21-fast-2 ]; load-balance per-packet; term 3 { from community vpn-red; install-nexthop lsp from-11-to-21-slow-3; Copyright 2017, Juniper Networks, Inc. 7

community vpls-grey members target:64512:10; community vpn-blue members target:64512:1; community vpn-red members target:64512:2; Next, under the routing-options hierarchy, apply this policy to the forwarding-table. This new policy is applied before the existing load-balancing policy so that it is evaluated first. You can use the insert command to modify the order, if necessary. user@pe11-re0# show routing-options forwarding-table { export [ MapVpnToLsp load-balance-per-packet ]; Verification First, ten streams of Grey VPLS traffic were transmitted from CE 14 to CE 24. The MPLS LSP statistics show that all traffic maps to the correct LSP from-11-to-21-fast-2 : user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 18941 173187412 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-slow-3 Total 3 displayed, Up 3, Down 0 Next, ten streams of Blue VPN traffic from CE 13 to CE 23 were transmitted. The MPLS LSP statistics show that traffic maps correctly to the two fast LSPs: user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 15908 145685464 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 23860 218509880 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-slow-3 Finally, ten streams of the Red VPN traffic from CE 13 to CE23 were transmitted. The MPLS LSP statistics show that traffic flows map to the slow LSP, as expected: user@pe11-re0> show mpls lsp ingress statistics name from-11-to-21-* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 59652 546293016 from-11-to-21-slow-3 Total 3 displayed, Up 3, Down 0 8 Copyright 2017, Juniper Networks, Inc.

Scenario 2 Mapping Specific Destinations to Specific LSPs Configuration In this example, PE 21 is configured to advertise, through MP-IBGP in the core, the IP VPN Blue subnets 10.2.231.0/24 and 10.231.0.0/16 with unique VRF targets. The corresponding Blue VRF on PE 11 is configured to accept network prefixes with both of these VRF targets. On PE 21, configure the BlueRouteCommAnnounce policy using the policy-options hierarchy: policy-statement BlueRouteCommAnnounce { term 1 { from { route-filter 10.2.231.0/24 exact; community set vpn-blue; term 2 { from { route-filter 10.231.0.0/16 orlonger; community set vpn-blue2; community vpn-blue members target:64512:1; community vpn-blue2 members target:64512:101; On PE 21, the policy is applied as a VRF export policy on the Blue VRF (under the routing-instances hierarchy): IP-VPN-Blue { instance-type vrf; interface xe-2/0/0.230; interface lo0.1; route-distinguisher 10.0.0.21:1; vrf-export BlueRouteCommAnnounce; vrf-target target:64512:1; vrf-table-label; etc. On PE 11, there are two policies configured under the policy-options hierarchy. The MapVpnToLsp policy (used in the previous configuration example) is modified so that the traffic destined to 10.231/16 is mapped to the slow LSP. The policy remains applied under the routing-options forwarding-table export configuration (not shown below). policy-statement MapVpnToLsp { term 1 { from community vpls-grey; install-nexthop lsp from-11-to-21-fast-2; Copyright 2017, Juniper Networks, Inc. 9

term 2 { from community vpn-blue; install-nexthop lsp [ from-11-to-21-fast-1 from-11-to-21-fast-2 ]; load-balance per-packet; term 3 { from community [ vpn-blue2 vpn-red ]; Map Blue2 destined traffic to slow LSP install-nexthop lsp from-11-to-21-slow-3; policy-statement BlueRouteCommAccept { term 1 { from community [ vpn-blue vpn-blue2 ]; then community vpls-grey members target:64512:10; community vpn-blue members target:64512:1; community vpn-blue2 members target:64512:101; community vpn-red members target:64512:2; The second policy, BlueRouteCommAccept (shown in the above configuration), accepts routes with both communities into the Blue VRF. This policy is applied as a VRF target import policy on the Blue IP VPN under the routing-instances hierarchy: IP-VPN-Blue { instance-type vrf; interface xe-2/0/1.130; interface lo0.1; route-distinguisher 10.0.0.11:1; vrf-import BlueRouteCommAccept; etc Verification First, confirm that each of the routes is received and placed into the Blue VRF with the appropriate community string: user@pe11-re0> show route 10.2.231.0/24 table IP-VPN-Blue detail match "10.2.231 Comm" 10.2.231.0/24 (2 entries, 1 announced) Communities: target:64512:1 Communities: target:64512:1 {master user@pe11-re0> show route 10.231.0.0/16 table IP-VPN-Blue detail match "10.231.0 Comm" 10.231.0.0/16 (2 entries, 1 announced) Communities: target:64512:101 Communities: target:64512:101 10 Copyright 2017, Juniper Networks, Inc.

The configured LSP next-hops are reflected in the routing table for the unique destinations: user@pe11-re0> show route 10.2.231.0/24 table IP-VPN-Blue IP-VPN-Blue.inet.0: 10017 destinations, 20027 routes (10017 active, 0 holddown, 0 hidden) Restart Complete + = Active Route, - = Last Active, * = Both 10.2.231.0/24 *[BGP/170] 05:07:33, localpref 100, from 10.0.0.21 AS path: 65521 I > to 10.12.1.21 via xe-2/1/0.0, label-switched-path from-11-to-21-fast-1 to 10.12.2.21 via xe-2/2/0.0, label-switched-path from-11-to-21-fast-2 [BGP/170] 05:07:33, localpref 100, from 10.0.0.12 AS path: 65521 I > to 10.12.1.21 via xe-2/1/0.0, label-switched-path from-11-to-21-fast-1 to 10.12.2.21 via xe-2/2/0.0, label-switched-path from-11-to-21-fast-2 {master user@pe11-re0> show route 10.231.0.0/16 table IP-VPN-Blue IP-VPN-Blue.inet.0: 10017 destinations, 20027 routes (10017 active, 0 holddown, 0 hidden) Restart Complete + = Active Route, - = Last Active, * = Both 10.231.0.0/16 *[BGP/170] 00:29:36, MED 2, localpref 100, from 10.0.0.21 AS path: 65521 I to 10.1.1.12 via xe-2/0/0.0, label-switched-path from-11-to-21-slow-3 [BGP/170] 00:29:36, MED 2, localpref 100, from 10.0.0.12 AS path: 65521 I to 10.1.1.12 via xe-2/0/0.0, label-switched-path from-11-to-21-slow-3 The tester was configured to send ten traffic streams from CE 13 to ten unique hosts on the 10.2.231.0/24 subnet connected to CE 23. The MPLS LSP statistics confirm that traffic traverses the two fast LSPs: user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 15907 145676306 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 23861 218519038 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-slow-3 Total 3 displayed, Up 3, Down 0 The tester was reconfigured to send traffic from CE 13 to destination 10.231.0.0/16 on CE 23. The MPLS LSP statistics confirm that this traffic traverses the slow LSP: user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21* Ingress LSP: 9 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-2 10.0.0.21 10.0.0.11 Up 158509 1451625422 from-11-to-21-slow-3 Total 3 displayed, Up 3, Down 0 Copyright 2017, Juniper Networks, Inc. 11

Additional Tests Non-matching Traffic To test the behavior of incoming traffic with a destination that does not match any of the terms in the MapVpnToLsp policy, Term 3 of the policy was temporarily deactivated. user@pe11-re0# deactivate policy-options policy-statement MapVpnToLsp term 3 user@pe11-re0# commit and-quit The routing table for destination 10.231/16 (corresponding to the vpn-blue2 community in the deactivated term), shows that all three LSPs are now available for forwarding to that destination: user@pe11-re0> show route 10.231.0.0 IP-VPN-Blue.inet.0: 15017 destinations, 25027 routes (15017 active, 0 holddown, 0 hidden) Restart Complete + = Active Route, - = Last Active, * = Both 10.231.0.0/16 *[BGP/170] 01:03:45, MED 2, localpref 100, from 10.0.0.21 AS path: 65521 I > to 10.12.1.21 via xe-2/1/0.0, label-switched-path from-11-to-21-fast-1 to 10.12.2.21 via xe-2/2/0.0, label-switched-path from-11-to-21-fast-2 to 10.1.1.12 via xe-2/0/0.0, label-switched-path from-11-to-21-slow-3 etc When resending the streams in the Red VPN, and streams with destination 10.231/16 in the Blue VPN, traffic was loadbalanced among the three LSPs, and ultimately delivered to the destination. This is essentially the default forwarding behavior described in the first section. After testing this behavior, term 3 in the MapVpnToLsp policy was re-activated. user@pe11-re0# activate policy-options policy-statement MapVpnToLsp term 3 user@pe11-re0# commit and-quit Forwarding Behavior When LSP is Down To test the behavior when the LSP, specified as a next-hop in the policy, is down, deactivate the LSP from-11-to-21- slow-3. user@pe11-re0# deactivate protocols mpls lsp from-11-to-21-slow-3 user@pe11-re0# commit and-quit The traffic streams corresponding to the Red VPN, and to destination 10.231/16 in the Blue VPN (corresponding to community vpn-blue2 as configured in the policy), were transmitted. Traffic to a given destination subnet took a single path (as shown below). Traffic was not load-balanced as before, and delivered to the destination ultimately. However, traffic to different destinations was load-balanced. user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21-* Ingress LSP: 8 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 129680 1187609440 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 0 0 from-11-to-21-fast-2 Total 2 displayed, Up 2, Down 0 12 Copyright 2017, Juniper Networks, Inc.

The reason for this is that the accept action in term 3 of the policy is a terminating action. As a result, no loadbalancing action is applied to the traffic because the second load-balance-per-packet policy is not invoked. To force traffic to a specific destination subnet to be load-balanced, the load-balance per-packet action was added to term 3 of the policy. Any traffic matching term 3 was confirmed as being load-balanced. user@pe11-re0# set policy-options policy-statement MapVpnToLsp term 3 then load-balance per-packet user@pe11-re0# commit and-quit user@pe11-re0> show mpls lsp statistics ingress name from-11-to-21-* Ingress LSP: 8 sessions To From State Packets Bytes LSPname 10.0.0.21 10.0.0.11 Up 133731 1224708498 from-11-to-21-fast-1 10.0.0.21 10.0.0.11 Up 133732 1224717656 from-11-to-21-fast-2 Total 2 displayed, Up 2, Down 0 You can achieve the same result by deleting the accept action in term 3. Traffic matches term 3 and the second policy, load-balance-per-packet, is applied to the traffic and load-balanced among the two active LSPs. The test for this configuration was successful. The Strict Option In some cases, it may be desirable not to forward traffic when the next-hop LSP specified in the policy is down. You can implement this behavior by including the strict keyword when configuring the next-hop LSP. This keyword was added to term 3 of the route policy: policy-statement MapVpnToLsp { term 1 { from community vpls-grey; install-nexthop lsp from-11-to-21-fast-2; term 2 { from community vpn-blue; install-nexthop lsp [ from-11-to-21-fast-1 from-11-to-21-fast-2 ]; load-balance per-packet; term 3 { from community [ vpn-blue2 vpn-red ]; install-nexthop strict lsp from-11-to-21-slow-3; include strict keyword The traffic streams corresponding to the Red VPN and to destinations 10.231/16 in the Blue VPN (corresponding to community vpn-blue2 as configured in the policy) were dropped on ingress PE 11, since they matched term 3 of the route policy. Copyright 2017, Juniper Networks, Inc. 13

Appendix A PE 11 Configuration This appendix includes CLI configuration code examples used to configure the PE 11 device. Interfaces Hierarchy CLI configuration for the interfaces hierarchy: ge-1/0/0 { flexible-vlan-tagging; encapsulation flexible-ethernet-services; unit 10 { encapsulation vlan-vpls; vlan-id 10; xe-2/0/0 { description Core-to-12; unit 0 { family inet { address 10.1.1.11/24; family mpls; xe-2/0/1 { flexible-vlan-tagging; encapsulation flexible-ethernet-services; unit 130 { vlan-id 130; family inet { address 10.1.30.11/24; unit 150 { vlan-id 150; family inet { address 10.1.50.11/24; xe-2/1/0 { description Core-to-21; unit 0 { family inet { address 10.12.1.11/24; family mpls; xe-2/2/0 { description Core-to-21; unit 0 { family inet { address 10.12.2.11/24; family mpls; 14 Copyright 2017, Juniper Networks, Inc.

xe-2/3/0 { description Core-to-12; mtu 9192; unit 0 { family inet { address 10.1.2.11/24; family mpls; lo0 { unit 0 { family inet { address 10.0.0.11/32; Protocols OSPF Hierarchy CLI configuration for the protocols ospf hierarchy: protocols { ospf { traffic-engineering; area 0.0.0.0 { interface all; interface fxp.0 { disable Protocols BGP Hierarchy CLI configuration for the protocols bgp hierarchy: protocols { bgp { log-updown; local-as 64512; group Internal { type internal; local-address 10.0.0.11; family inet-vpn { any; family l2vpn { signaling; peer-as 64512; neighbor 10.0.0.21; neighbor 10.0.0.12; Copyright 2017, Juniper Networks, Inc. 15

Protocols RSVP Hierarchy CLI configuration for the protocols rsvp hierarchy: protocols { rsvp { interface all; interface fxp0.0 { disable; Routing Instances Hierarchy CLI configuration for the routing-instances hierarchy where two IP VPN instances and the VPLS instance are defined: IP-VPN-Blue { instance-type vrf; interface xe-2/0/1.130; route-distinguisher 10.0.0.11:1; vrf-target target:64512:1; vrf-table-label; protocols { bgp { group EbgpTo13 { type external; hold-time 20; peer-as 65511; local-as 64512; neighbor 10.1.30.13; IP-VPN-Red { instance-type vrf; interface xe-2/0/1.150; route-distinguisher 10.0.0.11:2; vrf-target target:64512:2; vrf-table-label; protocols { bgp { group EbgpTo13 { type external; hold-time 20; peer-as 65511; local-as 64512; neighbor 10.1.50.13; VPLS-Grey { instance-type vpls; vlan-id none; interface ge-1/0/0.10; route-distinguisher 10.0.0.11:10; vrf-target target:64512:10; 16 Copyright 2017, Juniper Networks, Inc.

protocols { vpls { site-range 100; no-tunnel-services; Copyright 2017, Juniper Networks, Inc. 17