Traffic Engineering with MPLS Agenda u Introduction to

  • Slides: 123
Download presentation
Traffic Engineering with MPLS

Traffic Engineering with MPLS

Agenda u Introduction to traffic engineering Brief history v Vocabulary v Requirements for Traffic

Agenda u Introduction to traffic engineering Brief history v Vocabulary v Requirements for Traffic Engineering v Basic Examples v u Signaling LSPs with RSVP signaling protocol v RSVP objects v Extensions to RSVP v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 2

Agenda u Constraint-based traffic engineering Extensions to IS-IS and OSPF v Traffic Engineering Database

Agenda u Constraint-based traffic engineering Extensions to IS-IS and OSPF v Traffic Engineering Database v User defined constraints v Path section using CSPF algorithm v u Traffic protection Secondary LSPs v Hot-standby LSPs v Fast Reroute v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 3

Agenda u Advanced traffic engineering features Circuit cross connect (CCC) v IGP Shortcuts v

Agenda u Advanced traffic engineering features Circuit cross connect (CCC) v IGP Shortcuts v Configuring for transit traffic v Configuring for internal destinations v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 4

Introduction to Traffic Engineering

Introduction to Traffic Engineering

Why Engineer Traffic? What problem are we trying to solve with Traffic Engineering? Traffice

Why Engineer Traffic? What problem are we trying to solve with Traffic Engineering? Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 6

Brief History u Early 1990’s Internet core was connected with T 1 and T

Brief History u Early 1990’s Internet core was connected with T 1 and T 3 links between routers v Only a handful of routers and links to manage and configure v Humans could do the work manually v IGP (Interior Gateway Protocol) Metricbased traffic control was sufficient v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 7

IGP Metric-Based Traffic Engineering u Traffic sent to A or B follows path with

IGP Metric-Based Traffic Engineering u Traffic sent to A or B follows path with lowest metrics 1 1 A 1 C Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. B 2 Slide 8

IGP Metric-Based Traffic Engineering u Drawbacks Redirecting traffic flow to A via C causes

IGP Metric-Based Traffic Engineering u Drawbacks Redirecting traffic flow to A via C causes traffic for B to move also! v Some links become underutilized or overutilized v 1 4 A 1 C Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. B 2 Slide 9

IGP Metric-Based Traffic Engineering u Drawbacks v Only serves to move problem around w

IGP Metric-Based Traffic Engineering u Drawbacks v Only serves to move problem around w Some links underutilized w Some links overutilized v Lacks granularity w All traffic follows the IGP shortest path v Continuously adjusting IGP metrics adds instability to the network Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 10

Discomfort Grows u Mid 1990’s v v v ISPs became uncomfortable with size of

Discomfort Grows u Mid 1990’s v v v ISPs became uncomfortable with size of Internet core Large growth spurt imminent Routers too slow IGP metric engineering too complex IGP routing calculation was topology driven, not traffic driven Router based cores lacked predictability Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 11

Why Traffic Engineering? u There is a need for a more granular and deterministic

Why Traffic Engineering? u There is a need for a more granular and deterministic solution “A major goal of Internet Traffic Engineering is to facilitate efficient and reliable network operations while simultaneously optimizing network resource utilization and performance. ” RFC 2702 Requirements for Traffic Engineering over MPLS Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 12

Overlay Networks are Born ATM switches offered performance and predictable behavior u ISPs created

Overlay Networks are Born ATM switches offered performance and predictable behavior u ISPs created “overlay” networks that presented a virtual topology to the edge routers in their network u Using ATM virtual circuits, the virtual network could be reengineered without changing the physical network u Benefits u Full traffic control v Per-circuit statistics v More balanced flow of traffic across links v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 13

Overlay Networks ATM core ringed by routers u PVCs overlaid onto physical network u

Overlay Networks ATM core ringed by routers u PVCs overlaid onto physical network u A Physical View B Logical View C A C B Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 14

Path Creation u Off-line path calculation tool uses Link utilization v Historic traffic patterns

Path Creation u Off-line path calculation tool uses Link utilization v Historic traffic patterns v u Produces virtual network topology v u Primary and backup PVCs Generates switch and router configurations Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 15

Overlay Network Drawbacks u Growth in full mesh of ATM PVCs stresses everything With

Overlay Network Drawbacks u Growth in full mesh of ATM PVCs stresses everything With 5 routers, adding 1 requires only 10 new PVCs v With 200 routers, adding 1 requires 400 new PVCs v w From 39, 800 to 40, 200 PVCs total Router IGP runs out of steam v Practical limitation of atomically updating configurations in each switch and router v u Not well integrated v Network does not participate in path selection and setup Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 16

Overlay Network Drawbacks u ATM cell overhead Approximately 20% of bandwidth v OC-48 link

Overlay Network Drawbacks u ATM cell overhead Approximately 20% of bandwidth v OC-48 link wastes 498 Mbps in ATM cell overhead v OC-192 link wastes 1. 99 Gbps v u ATM SAR speed v OC-48 SAR w Trailing behind the router curve w Very difficult to build v OC-192 SAR? Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 17

Routers Caught Up u Current generation of routers have High speed, wire-rate interfaces v

Routers Caught Up u Current generation of routers have High speed, wire-rate interfaces v Deterministic performance v Software advances v u Solution Fuse best aspects of ATM PVCs with highperformance routing engines v Use low-overhead circuit mechanism v Automate path selection and configuration v Implement quick failure recovery v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 18

Benefits of MPLS u Low-overhead virtual circuits for IP v Originally designed to make

Benefits of MPLS u Low-overhead virtual circuits for IP v Originally designed to make routers faster w Fixed label lookup faster than longest match used by IP routing v Not true anymore! Value of MPLS is now in traffic engineering u One, integrated network u Same forwarding mechanism can support multiple applications u v Traffic Engineering, VPNs, etc…. Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 19

What are the fundamental requirements? u RFC 2702 v u Requirement for Traffic Engineering

What are the fundamental requirements? u RFC 2702 v u Requirement for Traffic Engineering over MPLS Requirements v v v Control Measure Characterize Integrate routing and switching All at a lower cost Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 20

Fundamental Requirements u Need the ability to: Map traffic to an LSP v Monitor

Fundamental Requirements u Need the ability to: Map traffic to an LSP v Monitor and measure traffic v Specify explicit path of an LSP v w Partial explicit route w Full explicit route v Characterize an LSP w Bandwidth w Priority/ Preemption w Affinity (Link Colors) v Reroute or select an alternate LSP Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 21

MPLS Fundamentals

MPLS Fundamentals

MPLS Header u IP packet is encapsulated in MPLS header and sent down LSP

MPLS Header u IP packet is encapsulated in MPLS header and sent down LSP IP Packet … 32 -bit MPLS Header u IP packet is restored at end of LSP by egress router v TTL is adjusted by default Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 23

MPLS Header Label u TTL Label v u EXP S Used to match packet

MPLS Header Label u TTL Label v u EXP S Used to match packet to LSP Experimental bits v Carries packet queuing priority (Co. S) Stacking bit u Time to live u v Copied from IP TTL Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 24

Router Based Traffic Engineering Standard IGP routing u IP prefixes bound to physical next

Router Based Traffic Engineering Standard IGP routing u IP prefixes bound to physical next hop u v Typically based on IGP calculation 192. 168. 1/24 134. 112/16 New York San Francisco Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 25

Router Based Traffic Engineering u Engineer unidirectional paths through your network without using the

Router Based Traffic Engineering u Engineer unidirectional paths through your network without using the IGP’s shortest path calculation IGP shortest path New York San Francisco JUNOS traffic engineered path Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 26

Router Based Traffic Engineering u IP prefixes can now be bound to LSPs 192.

Router Based Traffic Engineering u IP prefixes can now be bound to LSPs 192. 168. 1/24 New York San Francisco 134. 112/16 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 27

MPLS Labels Assigned manually or by a signaling protocol in each LSR during path

MPLS Labels Assigned manually or by a signaling protocol in each LSR during path setup u Labels change at each segment in path u LSR swaps incoming label with new outgoing label u Labels have “local significance” u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 28

MPLS Forwarding Example An IP packet destined to 134. 112. 1. 5/32 arrives in

MPLS Forwarding Example An IP packet destined to 134. 112. 1. 5/32 arrives in SF u San Francisco has route for 134. 112/16 u v Next hop is the LSP to New York 134. 112/16 IP New York 134. 112. 1. 5 San Francisco 0 1965 1026 Santa Fe Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 29

MPLS Forwarding Example u San Francisco prepends MPLS header onto IP packet and sends

MPLS Forwarding Example u San Francisco prepends MPLS header onto IP packet and sends packet to first transit router in the path 134. 112/16 New York San Francisco 1965 IP Santa Fe Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 30

MPLS Forwarding Example Because the packet arrived at Santa Fe with an MPLS header,

MPLS Forwarding Example Because the packet arrived at Santa Fe with an MPLS header, Santa Fe forwards it using the MPLS forwarding table u MPLS forwarding table derived from mpls. 0 switching table u 134. 112/16 New York San Francisco 1026 IP Santa Fe Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 31

MPLS Forwarding Example Packet arrives from penultimate router with label 0 u Egress router

MPLS Forwarding Example Packet arrives from penultimate router with label 0 u Egress router sees label 0 and strips MPLS header u Egress router performs standard IP forwarding decision u IP 134. 112/16 New York 0 IP San Francisco Santa Fe Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 32

Router Y Example Topology Small. Net Router X IGP Link Metric 10 10 192.

Router Y Example Topology Small. Net Router X IGP Link Metric 10 10 192. 168. 0. 1 Router C 192. 168. 2. 1 10 E- Router B BG P Big. Net Router D 192. 168. 24. 1 30 Router A 192. 168. 16. 1 30 Router E 192. 168. 5. 1 20 20 30 Router F 192. 168. 8. 1 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 20 Router G 192. 168. 12. 1 Slide 33

Router Y Example Topology Small. Net Router X Big. Net 192. 168. 2. 1

Router Y Example Topology Small. Net Router X Big. Net 192. 168. 2. 1 . 1 10 . 0 Router A . 1 10. 0 2. Router C . 24/ 17 . 1 30 . 2 . 0 0 1 /3 . 1 . 2 . 2 Router D 192. 168. 24. 1 0 3 2/ 0 192. 168. 16. 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 16 . 4 /3 0 . 2 . 3 10 30 . 0 5/ 1/ . 1 10 . 8 /3 0. 2 . 1 Router F . 1 10. 0. 13/30 192. 168. 8. 1 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 10 30 Router E 192. 168. 5. 1 Router G 192. 168. 12. 1 Slide 34

Traffic Engineering Signaled LSPs

Traffic Engineering Signaled LSPs

Static vs Signaled LSPs u Static LSPs Are ‘nailed up’ manually v Have manually

Static vs Signaled LSPs u Static LSPs Are ‘nailed up’ manually v Have manually assigned MPLS labels v Needs configuration on each router v Do not re-route when a link fails v u Signaled LSPs Signaled by RSVP v Have dynamically assigned MPLS labels v Configured on ingress router only v Can re-route around failures v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 36

Signaled Label-Switched Paths u Configured at ingress router only RSVP sets up transit and

Signaled Label-Switched Paths u Configured at ingress router only RSVP sets up transit and egress routers automatically v Path through network chosen at each hop using routing table v Intermediate hops can be specified as “transit points” v w Strict—Must use hop, must be directly connected w Loose—Must use hop, but use routing table to find it u Advantages over static paths Performs “keepalive” checking v Supports fail-over to unlimited secondary LSPs v Excellent visibility v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 37

RSVP Path Signaling Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks,

RSVP Path Signaling Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 38

Path Signaling u JUNOS uses RSVP for Traffic Engineering Internet standard for reserving resources

Path Signaling u JUNOS uses RSVP for Traffic Engineering Internet standard for reserving resources v Extended to support v w Explicit path configuration w Path numbering w Route recording v Provides keepalive status w For visibility w For redundancy Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 39

RSVP A generic Qo. S signaling protocol u An Internet control protocol u v

RSVP A generic Qo. S signaling protocol u An Internet control protocol u v Uses IP as its network layer Originally designed for host-to-host u Uses the IGP to determine paths u RSVP is not u A data transport protocol v A routing protocol v u RFC 2205 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 40

Basic RSVP Path Signaling Simplex flows u Ingress router initiates connection u “Soft” state

Basic RSVP Path Signaling Simplex flows u Ingress router initiates connection u “Soft” state u Path and resources are maintained dynamically v Can change during the life of the RSVP session v Path message sent downstream u Resv message sent upstream u Sender PATH RESV Router Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. PATH RESV Router PATH RESV Receiver Slide 41

Other RSVP Message Types u Path. Tear v u Resv. Tear v u Sent

Other RSVP Message Types u Path. Tear v u Resv. Tear v u Sent to ingress router Resv. Err v u Sent to ingress router Path. Err v u Sent to egress router Resv. Conf Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 42

Extended RSVP u Extensions added to support establishment and maintenance of LSPs v Maintained

Extended RSVP u Extensions added to support establishment and maintenance of LSPs v Maintained via “hello” protocol Used now for router-to-router connectivity u Includes the distribution of MPLS labels u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 43

MPLS Extensions to RSVP u Path and Resv message objects v v v u

MPLS Extensions to RSVP u Path and Resv message objects v v v u Explicit Route Object (ERO) Label Request Object Label Object Record Route Object Session Attribute Object Tspec Object For more detail on contents of objects: daft-ietf-mpls-rsvp-lsp-tunnel-04. txt Extensions to RSVP for LSP Tunnels Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 44

Explicit Route Object Used to specify the route RSVP Path messages take for setting

Explicit Route Object Used to specify the route RSVP Path messages take for setting up LSP u Can specify loose or strict routes u Loose routes rely on routing table to find destination v Strict routes specify the directly-connected next router v u A route can have both loose and strict components Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 45

ERO: Strict Route u Next hop must be directly connected to previous hop ERO

ERO: Strict Route u Next hop must be directly connected to previous hop ERO C E B D F Egress LSR B strict; C strict; E strict; D strict; F strict; A Ingress LSR Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Strict Slide 46

ERO: Loose Route u Consult the routing table at each hop to determine the

ERO: Loose Route u Consult the routing table at each hop to determine the best path ERO C E B D F Egress LSR D loose; A Ingress LSR Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Loose Slide 47

ERO: Strict/Loose Path u Strict and loose routes can be mixed ERO C E

ERO: Strict/Loose Path u Strict and loose routes can be mixed ERO C E F B D Strict Egress LSR C strict; D loose; F strict; A Ingress LSR Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Loose Slide 48

Router Y Partial Explicit Route Small. Net u “Loose” hop to Router G u

Router Y Partial Explicit Route Small. Net u “Loose” hop to Router G u Follow the IGP shortest path to G first Router X 192. 168. 2. 1 . 1 10 . 0 Router A . 1 10. 0 2. Router C . 24/ 17 . 1 30 . 2 . 0 0 1 /3 . 1 . 2 . 2 Router D 192. 168. 24. 1 0 3 2/ 0 192. 168. 16. 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 16 . 4 /3 0 . 2 . 3 10 30 . 0 5/ 1/ . 1 10 . 8 /3 0. 2 . 1 Router G Router F 192. 168. 8. 1 . 2 10 30 Router E 192. 168. 5. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 49

Router Y Full (Strict) Explicit Route A – F– G – E – C

Router Y Full (Strict) Explicit Route A – F– G – E – C – D Follow the Explicit Router X. 2 192. 168. 2. 1 . 1 10 . 0 Router A . 1 10. 0 2. Router C . 24/ 17 . 1 30 . 2 . 0 0 1 /3 . 1 . 2 . 2 Router D 192. 168. 24. 1 0 3 2/ 0 192. 168. 16. 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 16 . 4 /3 0 u u Small. Net . 3 10 30 . 0 5/ 1/ . 1 10 . 8 /3 0. 2 . 1 Router G Router F 192. 168. 8. 1 . 2 10 30 Router E 192. 168. 5. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 50

Hop-by-Hop ERO Processing u If Destination Address of RSVP message belongs to your router

Hop-by-Hop ERO Processing u If Destination Address of RSVP message belongs to your router You are the egress router v End ERO processing v Send RESV message along reverse path to ingress v u Otherwise, examine next object in ERO Consult routing table v Determine physical next hop v u If ERO object is strict v u Verify next router is directly connected Forward to physical next hop Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 51

Label Objects u Label Request Object Added to PATH message at ingress LSR v

Label Objects u Label Request Object Added to PATH message at ingress LSR v Requests that each LSR provide label to upstream LSR v u Label Object Carried in RESV messages along return path upstream v Provides label to upstream LSR v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 52

Record Route Object— PATH Message Added to PATH message by ingress LSR u Adds

Record Route Object— PATH Message Added to PATH message by ingress LSR u Adds outgoing IP address of each hop in the path u v u In downstream direction Loop detection mechanism Sends “Routing problem, loop detected” Path. Err message v Drops PATH message v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 53

Record Route Object — RESV Message Added to RESV message by egress LSR u

Record Route Object — RESV Message Added to RESV message by egress LSR u Adds outgoing IP address of each hop in the path u v u In upstream direction Loop detection mechanism Sends “Routing problem, loop detected” Resv. Err message v Drops RESV message v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 54

Session Attribute Object Added to PATH message by ingress router u Controls LSP u

Session Attribute Object Added to PATH message by ingress router u Controls LSP u Priority v Preemption v Fast-reroute v u Identifies session v ASCII character string for LSP name Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 55

Tspec Object u Contains link management configuration Requested bandwidth v Minimum and maximum LSP

Tspec Object u Contains link management configuration Requested bandwidth v Minimum and maximum LSP packet size v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 56

Path Signaling Example u Signaling protocol sets up path from San Francisco to New

Path Signaling Example u Signaling protocol sets up path from San Francisco to New York, reserving bandwidth along the way Seattle San Francisco (Ingress) PATH New York (Egress) PAT H PATH Miami Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 57

Path Signaling Example u Once path is established, signaling protocol assigns label numbers in

Path Signaling Example u Once path is established, signaling protocol assigns label numbers in reverse order from New York to San Francisco Seattle San Francisco (Ingress) 196 RES V 3 5 1026 RESV Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. RESV New York (Egress) Miami Slide 58

Adjacency Maintenance— Hello Message u New RSVP extension Hello message v Hello Request v

Adjacency Maintenance— Hello Message u New RSVP extension Hello message v Hello Request v Hello Acknowledge v u Rapid node to node failure detection Asynchronous updates v 3 second default update timer v 12 second default dead timer v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 59

Path Maintenance— Refresh Messages Maintains reservation of each LSP u Sent every 30 seconds

Path Maintenance— Refresh Messages Maintains reservation of each LSP u Sent every 30 seconds by default u Consists of PATH and RESV messages u Node to node, not end to end u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 60

RSVP Message Aggregation Bundles up to 30 RSVP messages within single PDU u Controls

RSVP Message Aggregation Bundles up to 30 RSVP messages within single PDU u Controls u Flooding of Path. Tear or Path. Err messages v Periodic refresh messages (PATH and RESV) v Enhances protocol efficiency and reliability u Disabled by default u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 61

Traffic Engineering Constrained Routing

Traffic Engineering Constrained Routing

Signaled vs Constrained LSPs u Common Features Signaled by RSVP v MPLS labels automatically

Signaled vs Constrained LSPs u Common Features Signaled by RSVP v MPLS labels automatically assigned v Configured on ingress router only v u Signaled LSPs CSPF not used v User configured ERO handed to RSVP for signaling v RSVP consults routing table to make next hop decision v u Constrained LSPs CSPF used v Full path computed by CSPF at ingress router v Complete ERO handed to RSVP for signaling v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 63

Constrained Shortest Path First Algorithm Modified “shortest path first” algorithm u Finds shortest path

Constrained Shortest Path First Algorithm Modified “shortest path first” algorithm u Finds shortest path based on IGP metric while satisfying additional constraints u Integrates TED (Traffic Engineering Database) u IGP topology information v Available bandwidth v Link color v u Modified by administrative constraints Maximum hop count v Bandwidth v Strict or loose routing v Administrative groups v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 64

Computing the ERO u Ingress LSR passes user defined restrictions to CSPF Strict and

Computing the ERO u Ingress LSR passes user defined restrictions to CSPF Strict and loose hops v Bandwidth constraints v Admin Groups v u CSPF algorithm Factors in user defined restrictions v Runs computation against the TED v Determines the shortest path v u CSPF hands full ERO to RSVP for signaling Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 65

Traffic Engineering Database Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks,

Traffic Engineering Database Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 66

Traffic Engineering Database CSPF uses TED to calculate explicit paths across the physical topology

Traffic Engineering Database CSPF uses TED to calculate explicit paths across the physical topology u Similar to IGP link-state database u Relies on extensions to IGP u Network link attributes v Topology information v u Separate from IGP database Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 67

TE Extensions to ISIS/OSPF u Describes traffic engineering topology v Traffic engineering database (TED)

TE Extensions to ISIS/OSPF u Describes traffic engineering topology v Traffic engineering database (TED) w Bandwidth w Administrative groups v Does not necessarily match regular routed topology w Subset of IGP domain ISIS Extensions v IP reachability TLV v IS reachability TLV v u OSPF Extension v Type 10 Opaque LSA Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 68

ISIS TE Extensions u IP Reachability TLV IP prefixes that are reachable v IP

ISIS TE Extensions u IP Reachability TLV IP prefixes that are reachable v IP link default metric v w Extended to 32 bits (wide metrics) v Up/down bit w Avoids loops in L 1/L 2 route leaking Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 69

ISIS TE Extensions u IS Reachability TLV IS neighbors that are reachable v ID

ISIS TE Extensions u IS Reachability TLV IS neighbors that are reachable v ID of adjacent router v w IP addresses of interface (/32 prefix length) v Sub-TLVs describe the TE topology Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 70

ISIS IS Reachability TLV u Sub-TLVs contain v v v v Local interface IP

ISIS IS Reachability TLV u Sub-TLVs contain v v v v Local interface IP address Remote interface IP address Maximum link bandwidth Maximum reservable link bandwidth Reservable link bandwidth Traffic engineering metric Administrative group Reserved TLVs for future expansion Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 71

OSPF TE Extensions u Opaque LSA v v v u Original Router LSA not

OSPF TE Extensions u Opaque LSA v v v u Original Router LSA not extensible Type 10 LSA Area flooding scope Standard LSA header (20 bytes) TE capabilities Traffic Engineering LSA v Work in progress Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 72

Configuring Constraints— LSP 1 with 40 Mbps Small. Net Follows the IGP shortest path

Configuring Constraints— LSP 1 with 40 Mbps Small. Net Follows the IGP shortest path to D since sufficient bandwidth available 0 /3 Router C 192. 168. 2. 1 . 1 10 . 0 Router A . 1 16 . 4. 1 10. 0 2. 10. 0. 1/30 . 24/ 17 . 2 Router B 192. 168. 0. 1 10 . 2 30 . 2 . 0 0 1 /3 . 2 Router D 192. 168. 24. 1 0 3 2/ 0 192. 168. 16. 1 6/3 . 2 . 1 1: LSP ps Mb 40 0 Router X . 1 . 2 u Router Y . 3 10 30 . 0 5/ 1/ . 1 10 . 8 /3 0. 2 . 1 Router G Router F 192. 168. 8. 1 . 2 10 30 Router E 192. 168. 5. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 73

Configuring Constraints— LSP 2 with 70 Mbps Small. Net Insufficient bandwidth available on IGP

Configuring Constraints— LSP 2 with 70 Mbps Small. Net Insufficient bandwidth available on IGP shortest path 0 /3. 4 16 2. . . 0 0 1 /3 . 2 . 1 30 Router E 70 LSP M 2: bp s 192. 168. 5. 1 10 . 8 /3 0. 2 . 1 Router G Router F 192. 168. 8. 1 Router D 192. 168. 24. 1 0 . 2 3 2/ 0 5/ 30 1/ . 0 192. 168. 16. 1 . 24/ . 3 . 0 10 Router A . 1 192. 168. 2. 1 . 1 10 Router C 17 . 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B 10 . 2 . 1 6/3 . 2 1: LSP ps Mb 40 0 Router X . 2 u Router Y . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 74

Affinity (Link Colors) u Ability to assign a color to each link Gold v

Affinity (Link Colors) u Ability to assign a color to each link Gold v Silver v Bronze v Up to 32 colors available u Can define an affinity relationship u Include v Exclude v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 75

Configuring Constraints— LSP 3 with 50 Mbps u Router Y Small. Net Exlcude all

Configuring Constraints— LSP 3 with 50 Mbps u Router Y Small. Net Exlcude all Bronze links Router X 5/ 30 /3. 4 16 2. 17 192. 168. 24. 1 0 3 2/ . . 0 0 1 /3 0 Router E 70 LSP M 2: bp s 192. 168. 5. 1 10 . 8 /3 0. 2 . 1 Bronze Router F 192. 168. 8. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 30 . 2 . 1 ze . 0 1/ . 0 Router D . 3 . 1 . 2 . 0 30 on 10 . 24/ 192. 168. 2. 1 10 . 2 10. 0 . 1 Br 192. 168. 16. 1 Router C . 2 10 Router A . 1 . 1 s Mbp e 0 : 2 ronz 3 P LS ude B l Exc . 1 10. 0. 1/30 Bronze 192. 168. 0. 1 10 . 2 Router B . 1 6/3 . 2 1: LSP ps Mb 40 0 0 . 2 Router G. 2 192. 168. 12. 1 Slide 76

Preemption Defines relative importance of LSPs on same ingress router u CSPF uses priority

Preemption Defines relative importance of LSPs on same ingress router u CSPF uses priority to optimize paths u Higher priority LSPs u Are established first v Offer more optimal path selection v May tear down lower priority LSPs when rerouting v u Default configuration makes all LSPs equal Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 77

Preemption u Controlled by two settings v Setup priority and hold (reservation) priority w

Preemption u Controlled by two settings v Setup priority and hold (reservation) priority w New LSP compares its setup priority with hold priority of existing LSP w If setup priority is less than hold priority, existing LSP is rerouted to make room Priorities from 0 (strong) through 7 (weak) v Defaults v w Setup priority is 7 (do not preempt) w Reservation priority is 0 (do not allow preemption) u Use with caution v No large scale experience with this feature Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 78

LSP Reoptimization u Reroutes LSPs that would benefit from improvements in the network v

LSP Reoptimization u Reroutes LSPs that would benefit from improvements in the network v u Special rules apply Disabled by default in JUNOS Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 79

LSP Reoptimization Rules u Reoptimize if new path can be found that meets all

LSP Reoptimization Rules u Reoptimize if new path can be found that meets all of the following Has lower IGP metric v Has fewer hops v Does not cause preemption v Reduces congestion by 10% v w Compares aggregate available bandwidth of new and old path u Intentionally conservative rules, use with care Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 80

LSP Load Balancing u Two categories v Selecting path for each LSP w Multiple

LSP Load Balancing u Two categories v Selecting path for each LSP w Multiple equal cost IP paths to egress are available w Random w Least-fill w Most-fill v Balance traffic over multiple LSP w Multiple equal cost LSPs to egress are available w BGP can load balance prefixes over 8 LSPs Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 81

LSP Load Balancing u Selecting path for each LSP v Random is default w

LSP Load Balancing u Selecting path for each LSP v Random is default w Distributes LSPs randomly over available equal cost paths v Least-fill w Distributes LSPs over available equal cost paths based on available link bandwidth v Most-fill w LSPs fill one link first, then next Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 82

Selecting paths for each LSP Router Y Small. Net u Most fill, Least fill,

Selecting paths for each LSP Router Y Small. Net u Most fill, Least fill, Random u Configure 12 LSPs, each with 10 Mbps Router X 10 . 0 Router A 192. 168. 2. 1 . 0 /3 . 2 20 Router D 192. 168. 24. 1 3 2/ . 1 . 2 . 2 30 . . 0 0 192. 168. 16. 1 . 24/ 0 30 30 10. 0 2. . 1 17 Router C . 1 20 . 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 16 . 4 /3 0 . 2 30 10 30 . 8 /3 0. 2 . 1 20 Router F 192. 168. 8. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 1/ 20 . 3 5/ 20 . 1 10 . 2 10 30 Router E 192. 168. 5. 1 Router G. 2 192. 168. 12. 1 Slide 83

Load Balancing u Balancing traffic over multiple LSPs Up to 16 equal cost paths

Load Balancing u Balancing traffic over multiple LSPs Up to 16 equal cost paths for BGP v JUNOS default is per-prefix v Per-packet (per-flow) knob available v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 84

Balancing traffic over equal cost IGP paths Small. Net Without LSPs configured, prefixes are

Balancing traffic over equal cost IGP paths Small. Net Without LSPs configured, prefixes are distributed over equal cost IGP paths Router X. 2 10 . 0 Router A 192. 168. 2. 1 . 0 /3 . 2 20 Router D 192. 168. 24. 1 3 2/ . 1 . 2 . 2 30 . . 0 0 192. 168. 16. 1 . 24/ 0 30 30 10. 0 2. . 1 17 Router C . 1 20 . 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 16 . 4 /3 0 u Router Y 30 10 30 . 8 /3 0. 2 . 1 20 Router F 192. 168. 8. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 1/ 20 . 3 5/ 20 . 1 10 . 2 10 30 Router E 192. 168. 5. 1 Router G. 2 192. 168. 12. 1 Slide 85

Balancing traffic over equal cost LSPs Router Y Small. Net u Same behavior, now

Balancing traffic over equal cost LSPs Router Y Small. Net u Same behavior, now over LSPs u Prefixes distributed over multiple LSPs Router X . 0 20 10 . 3 20 30 10 5/ . 1 . 8 /3 0. 2 . 1 20 Router F 192. 168. 8. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 . 1 2. . 2 . 1 192. 168. 5. 1 . 0 Router D 192. 168. 24. 1 3 2/ Router E 10 . 2 20 . . 0 0 1 . 2 30 0 30 30 /3 . 24/ 192. 168. 2. 1 0 192. 168. 16. 1 10. 0 17 . 1 10 Router A Router C 20 . 1 . 1 192. 168. 0. 1 10 10. 0. 1/30 30 0 6/3 . 2 Router B 1/ . 2 16 . 4 /3 0 . 2 Router G. 2 192. 168. 12. 1 Slide 86

Advanced Traffic Engineering Features

Advanced Traffic Engineering Features

Traffic Protection

Traffic Protection

Traffic Protection u Primary LSP Retry timer v Retry limit v u Secondary LSPs

Traffic Protection u Primary LSP Retry timer v Retry limit v u Secondary LSPs v Standby option Fast Reroute u Adaptive mode u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 89

Primary LSP u Optional v u If configured, becomes preferred path for LSP If

Primary LSP u Optional v u If configured, becomes preferred path for LSP If no primary configured v LSR makes all decisions to reach egress Zero or one primary path u Revertive capability u v Revertive behavior can be modified Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 90

Primary LSP u Revertive Capability v Retry timer w Time between attempts to bring

Primary LSP u Revertive Capability v Retry timer w Time between attempts to bring up failed primary path w Default is 30 seconds w Primary must be stable two times (2 x) retry timer before reverts back v Retry limit w Number of attempts to bring up failed primary path w Default is 0 (unlimited retries) w If limit reached, human intervention then required Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 91

Secondary LSP Optional u Zero or more secondary paths u All secondary paths are

Secondary LSP Optional u Zero or more secondary paths u All secondary paths are equal u v u Selection based on listed order of configuration Standby knob Maintains secondary path in ‘up’ condition v Eliminates call-setup delay of secondary LSP v Additional state information must be maintained v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 92

Secondary Paths— LSP 1, exclude Bronze u Router Y Small. Net Secondary – avoid

Secondary Paths— LSP 1, exclude Bronze u Router Y Small. Net Secondary – avoid primary if possible Router X . 2 10 . 0 Se 20 . 1 5/ 30 c 0 ond M a bp ry s : . 2 192. 168. 8. 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 20 16 2. . 1 . 0 . 8 /3 Bronze. 1 17 30 Router E 10 . 1 192. 168. 24. 1 20 0. 2 Router F . 4 /3. . 0 0 1 Router D 0 . 1 . 2 3 2/ 192. 168. 5. 1 30 30 . 0 192. 168. 16. 1 . 24/ 10 Router A 10 . 1 30 s Mbp e. 0. 0 0 /3 : 2 ronz 1 0 P LS ude B 30 l Exc G . 1 Bronze Go ld Router C 192. 168. 2. 1 10 10. 0 1/ . 1 old . 1 . 3 Router B 192. 168. 0. 1 10. 0. 1/30 Br G onz ol e d . 2 0 10. . 2 . 1 /30. 16 20 . 2 10 0 . 2 Router G. 2 192. 168. 12. 1 Slide 93

Adaptive Mode u Applies to LSP rerouting v Primary & secondary sharing links v

Adaptive Mode u Applies to LSP rerouting v Primary & secondary sharing links v Avoids double counting u SE Reservation style u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 94

Shared Links B E Shared link Ingress LSR A C F D C Egress

Shared Links B E Shared link Ingress LSR A C F D C Egress LSR E Session 1 Session 2 u FF reservation style: Each session has its own identity v Each session has its own bandwidth reservation v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. u SE Reservation style: Each session has its own identity v Sessions share a single bandwidth reservation v Slide 95

Secondary Paths— LSP 1, exclude Bronze Small. Net Secondary – in Standby mode, 20

Secondary Paths— LSP 1, exclude Bronze Small. Net Secondary – in Standby mode, 20 M exclude Gold s Mbp e. 0. 0 0 /3 : 2 ronz 1 0 P LS ude B 30 l Exc S . 2 10 . 0 20 . 1 5/ 30 ec Ex 20 ond cl M ar ud b y e ps : Go ld 0 Router E 10 . 8 /3 Bronze 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. 20 . 4 16 2. 17. 1 192. 168. 24. 1 20 0. 2 . 1 Router D 30 . 1 Router F 192. 168. 8. 1 . . 0 0 1 192. 168. 5. 1 . 2 3 2/ . 2 30 30 . 3 192. 168. 16. 1 0 . 24/ . 0 Router A 10 . 1 10 G . 1 Bronze Go ld Router C 192. 168. 2. 1 10 10. 0 30 . 1 old . 1 1/ Router B 192. 168. 0. 1 10. 0. 1/30 B ro G nz ol e d 0 10. . 2 . 2 /30. 16 . 2 . 1 10 Router X /3 20 . 2 u Router Y Router G. 2 192. 168. 12. 1 Slide 96

Fast Reroute Configured on ingress router only u Detours around node or link failure

Fast Reroute Configured on ingress router only u Detours around node or link failure u v ~100 s of ms reroute time Detour paths immediately available u Crank-back to node, not ingress router u Uses TED to calculate detour u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 97

Fast Reroute Short term solution to reduce packet loss u If node or link

Fast Reroute Short term solution to reduce packet loss u If node or link fails, upstream node u Immediately detours v Signals failure to ingress LSR v u Only ingress LSR knows policy constraints v Ingress computes alternate route w Based on configured secondary paths v Initiates long term reroute solution Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 98

Fast Reroute Example u Primary LSP from A to E F E A D

Fast Reroute Example u Primary LSP from A to E F E A D B Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. C Slide 99

Fast Reroute Example u Enable fast reroute on ingress A creates detour around B

Fast Reroute Example u Enable fast reroute on ingress A creates detour around B v B creates detour around C v C creates detour around D v F E A D B Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. C Slide 100

Fast Reroute Example Short Term Solution u B to C link fails B immediately

Fast Reroute Example Short Term Solution u B to C link fails B immediately detours around C v B signals to A that failure occurred v F E A D B Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. C Slide 101

Fast Reroute Example – Long Term Solution u A calculates and signals new primary

Fast Reroute Example – Long Term Solution u A calculates and signals new primary path F E A D B Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. C Slide 102

LSP Rerouting u Initiated by ingress LSR v u Exception is fast reroute Conditions

LSP Rerouting u Initiated by ingress LSR v u Exception is fast reroute Conditions that trigger reroute More optimal route becomes available v Failure of a resource along the LSP path v Preemption occurs v Manual configuration change v u Make before break (if adaptive) Establish new LSP with SE style v Transfer traffic to new LSP v Tear down old LSP v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 103

Advanced Route Resolution

Advanced Route Resolution

Mapping Transit Traffic u Mapping transit destinations JUNOS default mode v Only BGP prefixes

Mapping Transit Traffic u Mapping transit destinations JUNOS default mode v Only BGP prefixes are bound to LSPs v Only BGP can use LSPs for its recursive route calculations v Only BGP prefixes that have the LSP destination address as the BGP next-hop are resolvable through the LSP v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 105

Route Resolution– Transit Traffic Example Router Y Small. Net I-BGP 13 E 4. BG

Route Resolution– Transit Traffic Example Router Y Small. Net I-BGP 13 E 4. BG 11 P 2/ 16 Router X Router C 192. 168. 2. 1 . 1 10 . 0 Router A . 1 . 24/ . 0 30 . 2 . . 0 0 1 /3 30 Router E. 3 10 30 . 0 5/ 1/ . 1 10 . 0 192. 168. 5. 1 . 8 /3 0. 2 . 1 0 /3 Router G Router F 192. 168. 8. 1 Configure a “next hop self” policy on Router D . 2 10 . 4 . 1 . 2 . 2 Router D 192. 168. 24. 1 0 3 2/ 0 192. 168. 16. 1 16 10. 0 2. . 1 17 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 . 1 134. 112/16 . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 106

What if BGP next hop does not align with LSP endpoint? Router Y Small.

What if BGP next hop does not align with LSP endpoint? Router Y Small. Net I-BGP 13 E 4. BG 11 P 2/ 16 Router X . 1 10 . 0 T 192. 168. 16. 1 192. 168. 2. 1 . 24/ . 0 30 . 2 . . 0 0 1 0 30 Router E 10 . 3 5/ 30 . 0 . 1 1/ . 1 . 8 0 /3 /3 0. 2 . 1 Router G Router F 192. 168. 8. 1 IGP Passive interface 10 . 0 192. 168. 5. 1 . 2 10 . 4 . 1 . 2 . 2 Router D 192. 168. 24. 1 0 3 2/ /3 16 10. 0 2. Router C 17 . 1 ic raff Router A 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 1 0 6/3 . 2 . 1 134. 112/16 . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 107

Traffic Engineering Shortcuts u Configure TE Shortcuts on ingress router Good for BGP nexthops

Traffic Engineering Shortcuts u Configure TE Shortcuts on ingress router Good for BGP nexthops that are not resolvable directly through an LSP v If LSP exists that gets you closer to BGP nexthop v Installs prefixes that are downstream from egress router into ingress router’s inet. 3 route table v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 108

BGP next hops beyond the egress router can use the LSP! Router Y Small.

BGP next hops beyond the egress router can use the LSP! Router Y Small. Net I-BGP 13 E 4. BG 11 P 2/ 16 Router X . 0 0 1 /3 . 2 af f 0 /3. 4 16 . 1 30 Router E 192. 168. 5. 1 . 3 10 30 . 0 5/ 1/ . 1 . 8 /3 0. 2 . 1 BGP Next hop is down stream from LSP endpoint Router G Router F 192. 168. 8. 1 Router D 192. 168. 24. 1 0 ic . 0 . 2 10 10 30 . 2 Tr . 24/ 3 2/ 0 . 2 10. 0 . 1 10 Router A . 1 192. 168. 2. 1 . 1 192. 168. 16. 1 Router C 2. . 1 192. 168. 0. 1 10 10. 0. 1/30 . 2 0 6/3 . 2 Router B 17 . 2 . 1 134. 112/16 . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 109

TE Shortcuts By itself, still only usable by BGP u Installs additional prefixes in

TE Shortcuts By itself, still only usable by BGP u Installs additional prefixes in ingress router’s inet. 3 table u Only BGP can use routes in inet. 3 for BGP recursive lookups u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 110

But, cannot use the LSP for traffic destined to web servers Router Y Small.

But, cannot use the LSP for traffic destined to web servers Router Y Small. Net I-BGP 13 E 4. BG 11 P 2/ 16 Router X 0 6/3 . 0 W 192. 168. 2. 1 . 0 . 2 30 30 Router E 192. 168. 5. 1 ra 1/ t. T 10 ff ic . 0 . 8 /3 0. 2 . 1 0 /3. 4 16 10. 57. 16/24 Webserver Farm part of IGP domain Router G Router F 192. 168. 8. 1 2. . 1 . 3 5/ . . 0 0 1 /3 Router D 192. 168. 24. 1 . 0 . 1 . 2 10 . 0 30 0 Tr 10 . 24/ 3 2/ 0 an si 10. 0 . 2 . 1 . 2 192. 168. 16. 1 Router C. 1 10 ra eb T Router A . 1 ffic . 1 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B 17 . 2 . 1 134. 112/16 . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 111

BGP-IGP knob u Traffic-engineering bgp-igp knob Forces all MPLS prefixes into main routing table

BGP-IGP knob u Traffic-engineering bgp-igp knob Forces all MPLS prefixes into main routing table (inet. 0) v All destinations can now use all LSPs v w IGP and BGP prefixes Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 112

Now all traffic destined to egress Small. Net router and beyond use LSP Router

Now all traffic destined to egress Small. Net router and beyond use LSP Router Y I-BGP 13 E 4. BG 11 P 2/ 16 Router X . 0 30 Router E 192. 168. 5. 1 ff ic 10 . 3 1/ . 1 . 0 30 . 8 /3 0. 2 . 1 0 /3. 4 10. 57. 16/24 Webserver Farm part of IGP domain Router G Router F 192. 168. 8. 1 16 . 1 . 2 Router D 192. 168. 24. 1 10 5/ . 2 . . 0 0 1 /3 l. T ra . 1 30 0 Al. 0 . 24/ 3 2/ 0 192. 168. 16. 1 10 10. 0 . 1 10 Router A . 1 192. 168. 2. 1 . 1 . 2 Router C 2. . 1 17 10. 0. 1/30 192. 168. 0. 1 10 . 2 Router B . 2 0 6/3 . 2 . 1 134. 112/16 . 1 10. 0. 13/30 Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. . 2 192. 168. 12. 1 Slide 113

TTL Decrement u Default is to decrement TTL on all LSR hops Loop prevention

TTL Decrement u Default is to decrement TTL on all LSR hops Loop prevention v Topology discovery via traceroute v u Disable TTL decrement inside LSP No topology discovery v TTL decrement at egress router only v [edit protocols mpls label-switched-path lsp-path-name] user@host# set no-decrement-ttl Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 114

Circuit Cross Connect

Circuit Cross Connect

Circuit Cross-Connect (CCC) Transparent connection between two Layer 2 circuits u Supports u v

Circuit Cross-Connect (CCC) Transparent connection between two Layer 2 circuits u Supports u v u PPP, Cisco HDLC, Frame Relay, ATM, MPLS Router looks only as far as Layer 2 circuit ID Any protocol can be carried in packet payload v Only “like” interfaces can be connected (for example, Frame Relay to Frame Relay, or ATM to ATM) v u Three types of cross-connects Layer 2 switching v MPLS tunneling v Stitching MPLS LSPs v Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 116

CCC Layer 2 Switching A u u u DLCI 600 M 40 DLCI 601

CCC Layer 2 Switching A u u u DLCI 600 M 40 DLCI 601 B A and B have Frame Relay connections to M 40, carrying any type of traffic M 40 behaves as switch Layer 2 packets forwarded transparently from A to B without regard to content; only DLCI is changed CCC supports switching between PPP, Cisco HDLC, Frame Relay PVCs, or ATM PVCs ATM AAL 5 packets are reassembled before sending Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 117

CCC Layer 2 Switching A DLCI 600 M 40 so-5/1/0. 600 DLCI 601 B

CCC Layer 2 Switching A DLCI 600 M 40 so-5/1/0. 600 DLCI 601 B so-2/2/1. 601 [edit protocols] user@host# show connections { interface-switch connection-name { interface so-5/1/0. 600; interface so-2/2/1. 601; } } Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 118

CCC MPLS Interface Tunneling ATM access network A ATM VC 514 IP backbone M

CCC MPLS Interface Tunneling ATM access network A ATM VC 514 IP backbone M 40 MPLS LSP ATM access network M 20 ATM VC 590 B Transports packets from one interface through an MPLS LSP to a remote interface u Bridges Layer 2 packets from end-to-end u Supports tunneling between “like” ATM, Frame Relay, PPP, and Cisco HDLC connections u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 119

CCC MPLS Interface Tunneling ATM access network A ATM VC 514 IP backbone M

CCC MPLS Interface Tunneling ATM access network A ATM VC 514 IP backbone M 40 MPLS LSP 1 ATM access network M 20 ATM VC 590 B MPLS LSP 2 at-7/1/1. 514 [edit protocols] user@M 40# show connections { remote-interface-switch m 40 -to-m 20 interface at-7/1/1. 514; transmit-lsp 1; receive-lsp 2; } Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. at-3/0/1. 590 [edit protocols] user@M 20# show connections { remote-interface-switch m 20 -to-m 40 interface at-3/0/1. 590; transmit-lsp 2; receive-lsp 1; } Slide 120

CCC LSP Stitching LSR TE domain 2 LSR TE domain 1 LSR LSR TE

CCC LSP Stitching LSR TE domain 2 LSR TE domain 1 LSR LSR TE domain 3 LSP stitching LSR Large networks can be separated into several traffic engineering domains (supports IS-IS area partitioning) u CCC allows establishment of LSP across domains by “stitching” together LSPs from separate domains u Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 121

CCC LSP Stitching LSR-E TE domain 1 LSR-B TE domain 2 LSR-C LSR-D LSP

CCC LSP Stitching LSR-E TE domain 1 LSR-B TE domain 2 LSR-C LSR-D LSP stitching LSR-A [edit protocols] user@LSR-B# show connections { lsp-switch LSR-A_to_LSR-E { transmit-lsp 2; receive-lsp 1; } lsp-switch LSR-E_to_LSR-A { receive-lsp 3; transmit-lsp 4; } Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 122

www. juniper. net Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks,

www. juniper. net Traffice Engineering with MPLS–APRICOT 2000– 10/27/2020 Copyright © 2000, Juniper Networks, Inc. Slide 123