JRA 1 Task 1 Investigation of Emerging Carrier

JRA 1 Task 1 Investigation of Emerging Carrier Class Transport Network Technologies (CCTNT) Victor Olifer (JANET) TNC 2010, Vilnius, 01 -06 -2010 connect • communicate • collaborate

Agenda Introduction to JRA 1 Task 1 CCTNT Introduction. Why are CCTNT necessary? . Requirements. Benefits. CCTNT Descriptions: Ethernet developments NG-OTN. MPLS-TP. PBB-TE. JRA 1 Task 1 Future Plans. connect • communicate • collaborate

Introduction to JRA 1 Task 1 “GN 3 will be revolutionary in terms of the services it provides. Whilst the underlying technology at the lower layers of the network is not going to undergo substantial change, there will be a dramatic change in the services that will be developed and offered to end users” JRA 1 Task 1 will research the exploitation of the hybrid infrastructure by emerging transport technologies such as Carrier Class PBT and MPLS-TP in order to support point-to-point, point-to-multipoint and VPN services. Stage 1 – theoretical investigation of promising technologies: to what extent they can be called carrier-class transport Both emerging and established technologies were in scope: (only new carrierclass features investigated for the latter) The expected results of the work are the: Production of reference papers for GÉANT and NREN’s future transport network technologies. Not about photonic layer – this is JRA 1 Task 2 connect • communicate • collaborate

Carrier Class Transport Network Technologies (CCTNT) What is CCTNT? Requirements: Effective data transmission: to combine flexible multiplexing and provisioning with good performance (latency, bandwidth granularity) for each traffic type Support for standardised services. (e. g. MEF E-LINE & E-LAN) P-OTS readiness Manageability (OAM functionality similar to the traditional SDH/SONET) Simplicity Scalability and versatility. Reliability (Protection & Restoration). Qo. S. Dynamic provisioning (support for Control plane or NMS-based provisioning) Environmental requirements Low cost Benefits: Better and more reliable customer services built upon the transport CAPEX and OPEX reduction: simpler infrastructure, converged Possibility to satisfied the special needs from the research community Additional functionality (e. g. Bo. D ) Higher bandwidth Better possibilities for interoperability and interworking connect • communicate • collaborate

JRA 1 Task 1 Technologies & focus areas Technologies considered as relevant under the scope of JRA 1 Task 1: Next-Generation OTN (NG-OTN) Ethernet (new features) Layer 2 Routing Synchronous Ethernet over Multi-Protocol Label Switching (Eo. MPLS) Multi-Protocol Label Switching Transport Profile (MPLS-TP) Provider Backbone Bridge Traffic Engineering (PBB-TE) GMPLS and focus areas: Scalability Quality of Service (Qo. S) Protection and restoration Operations, Administration and Maintenance (OAM) functionality Multicasting. Control plane protocols (including GMPLS) Multi-domain Standardisation Applications Cost-effectiveness connect • communicate • collaborate

Optical Transport Networks (OTN) IP/MPLS Ethernet SDH/SONET OTN Physical Medium - Fibre A big step from SDH/SONET: Single technology Better scalability and flexibility Transparent for Client Signals (does not transfer network synchronization) Better Forward Error Correction Hierarchical Tandem Connection Monitoring functionality – multidomain support Fast Protection Restoration through GMPLS connect • communicate • collaborate

Next Generation OTN Evolution towards Packet-Optical Transport Sysytem NG-OTN promises a much more flexible mutiplexing hierarchy, designed for data traffic: ODU Flex – Flexible low order container that can be ”right sized”. ODU 0 and OPU 0 to accomodate 1 GE signals. ODU 2 e and OPUe 2 for transport of CBR 10 G 3 for 10 GE. New ODU 3 e and OTU 3 e for transport of 4 x ODU 2 e. ODU 4 and OTU 4 for transport of 100 GE. Enhanced OAM features: OTN Alarms and defects being reviewed by Study Group 15. Control Plane: GMPLS signalling extensions for G. 709 (RFC 4328). Conclusion: • • Carrier-class technology without any doubts Worth to trial NG feaqtures: OAM, dynamic provisioning, P-OTS capabilities connect • communicate • collaborate

Ethernet developments Ethernet is evolving producing: • Some strands that can be treated as separate transport technologies, i. e. PBB-TE or Eo. MPLS • New elements that might be seen as native Ethernet developments: MEF technology-agnostic definitions of Ethernet global services: • E-LINE (EPL & EVPL), E-LAN and E-TREE Ethernet OAM Ethernet Qo. S 40 G/100 G Ethernet connect • communicate • collaborate

Ethernet OAM CFM (802. 1 ag) from IEEE: • Continuity Check Messages (CCM) with end-to-end hierarchy: Ø • service status monitoring Loopback and Linktrace Messages – service troubleshooting Y. 1731 from ITU-T adds Performance Monitoring to CFM : • Frame Loss Messages. • Frame Delay Messages. connect • communicate • collaborate

MPLS-TP (MPLS–Transport Profile) Background & Definition T-MPLS-TP is the result of a joint effort between the ITU-T and the IETF. MPLS-TP is a subset of MPLS with extensions to support the requirements for transport networks. connect • communicate • collaborate

MPLS-TP Transport requirements (I) MPLS-TP OAM Sould be independent on IP forwarding and control plane MPLS-TP provides In-band OAM similar to transport model MPLS-TP generalises the use of Generic Associated Channel (G-ACh) to provide a mechanism to carry management and OAM information (RFC 5586). MPLS-TP defines a set of tools to provide “pro-active” and “on-demand” OAM. On going work in the IETF for definition of these tools. Tools under discussion: • ITU-T Y. 1731 • LSP Ping • BFD • Virtual Circuit Connectivity Verification (VCCV). connect • communicate • collaborate

MPLS-TP Transport requirements (II) Protection for different transport entities: sections, LSPs and PWs < 50 ms switching time. 1+1, 1: 1, n: 1 protection. Protection for uni-directional and bi-directional paths. Linear and ring protection Restoration (Control plane & Management Plane) Manual control. Triggered by operator. Failure triggered actions. OAM signalling. Control plane (GMPLS). connect • communicate • collaborate

MPLS-TP Conclusions & Status Current status: MPLS-TP is currently under development. There are five published RFCs and a lot if Internet Drafts At MPLS World Congress in Paris (February 2010) it was said that the core MPLS-TP standards would be complete by July 2010 Conclusions: MPLS-TP provides packet effiencicy inherited from MPLS, adds transport capabilities and removes some unnessary features Worth to trial and demonstrate connect • communicate • collaborate

Provider Backbone Bridge Traffic Engineering Initially developed by Nortel (in 2006) as Provider Backbone Transport (PBT) – it was Provider Backbone Bridges extension to support: Deterministic paths for point-to-point services (E-LINE) with bandwidth guarantees and Qo. S. Fast path protection switching (1: 1 and m: n) Standardised by the IEEE as PBB TE (802. 1 Qay) in 2009 (E-TREE services were added). Switches off MAC learning and STP but preserves the forwarding table format, population of which might be: Manual NMS-based GMPLS-based connect • communicate • collaborate

PBB TE scalability: Two-tier connection hierarchy Outer transport tunnel: {B-VID, B-MAC DA} as a globally unique transport label мс Customer C мс Customer A B-VID=117 Provider network B-MAC=0 x 35 B-VID=117 B-MAC=0 x 35 мс PE-1 B-VID=117 B-MAC=0 x 35 Customer B мс B-MAC=0 x 35 PE-2 B-VID=117 B-MAC=0 x 35 Customer D мс Inner service connections: - identified by I-SID as a service label: up to 16 millions per tunnel The technique is very similar to MPLS “tunnels+pseudowires” scheme but it uses well-known MAC addresses and VLAN Ids – globally unique labels Edge switches know: Nothing about customer VIDs & MACs for EPL (port-based) service Customer VIDs for EVPL (VLAN-based) service connect • communicate • collaborate

PBB TE features and status Resilience Primary and backup tunnels (1: 1) or groups (n: m); 50 ms. CFM heartbeat messages test tunnels and trigger protection switching. OAM No specific mechanisms; all new Ethernet OAM features can be used; CFM – mandatory for protection switching Control Plane Zero control plane – main option; NMS-based provisioning systems. GMPLS - Internet draft exists, no implementations known. Multi-domain support • Mostly a single-domain technology (access for IP/MPLS) • Can be used in multi-domain environment a cording MEF E-NNI spec Current status Standardised but immature yet (early releases). Eco-system shrunk after early enthusiasm – but there are several major vendors that support it Conclusion: worth to trial connect • communicate • collaborate

JRA 1 T 1 Status: Carrier Class Transport Network Technologies Layer 2 routing MPLS Synchronous Ethernet MPLS-TP PBB-TE NGN OTN Control Plane (GMPLS) Deliverable DJ 1. 1. 1 CCTNT Ethernet over MPLS Carrier Class Transport Network Technologies Next step: Further study and testing - OAM. - Protection & Restoration. - Control Plane (GMPLS). - Cost-effectiveness. - Multi-domain implications. Comprehensive study and demonstration NGN IP NGN OTN Control Plane (GMPLS) connect • communicate • collaborate

JRA 1 Task 1 participants Contributors: Alberto Colmenero – NORDUnet (Task Leader) Rebecca Corn – DANTE Marcin Garstka – PSNC Jac Kloots – SURFNET Victor Olifer – JANET Jan Radil – CESNET Krzysztof Stanecki – PSNC Sue Tyley – DANTE (Technical writer) Please check JRA 1 Task 1 report at: http: //www. geant. net/Media_Centre/Media_Library/Pages/Deliverables. aspx connect • communicate • collaborate