intelligent optical control plane
play

Intelligent Optical Control Plane Ying ( Em ily) Hu Bell Labs-NPS, - PowerPoint PPT Presentation

Intelligent Optical Control Plane Ying ( Em ily) Hu Bell Labs-NPS, Advanced Netw ork Modeling&Optim ization Lucent Technologies Holm del, New Jersey, USA 1 Outline Optical Control Plane What is it Operator Drivers for Optical


  1. Intelligent Optical Control Plane Ying ( Em ily) Hu Bell Labs-NPS, Advanced Netw ork Modeling&Optim ization Lucent Technologies Holm del, New Jersey, USA 1

  2. Outline � Optical Control Plane – What is it � Operator Drivers for Optical Control Plane � Requirements of Intelligent Control Plane � Control Plane Applications � Standards for Intelligent Control Plane � Co-Existence with Traditional Networks 2

  3. Optical Control Plane: What Is It? Traditional Transport Network Transport Network w. Intelligent Control Plane NML/EML NML/EML Control Plane Signaling+Routing+Discovery Transport Plane Transport Plane •Configuration management is moved from NMS to control plane •Centralized management (FCAPS) at NMS • Distributed database at NE •Centralized database at NMS • Intelligence at NE •No Intelligence at NE •Inventory and circuit information is self-discovered and flowed- •Inventory and circuit information is provisioned via back to the EML/NML NML •Circuit design occurs via NE independently of the NML •Circuit design is an NML function •Connection setup/teardown – Fast (Automate via the Control •Connection setup/teardown – Slow (Human Operator Plane) via the Management Plane) 3

  4. Operator Drivers For Optical Control Plane � Accuracy Of Network Database And Improved Utilization Of Network Resources � Improved Multi-Vendor Interworking � Change Of Operation Environment From Existing OS – Reduced Network Costs (CapEx) • Better network efficiency than SONET rings >>Meshed topology & Mesh restoration – Reduced Operation Costs (OpEx) • Reduced cost of provisioning • Reduced provisioning time >> Automatic provisioning – New service and revenue opportunities • Broad range of differentiated services >>Flexibility • Bandwidth On Demand >> Fast provisioning • OVPN (Optical Virtual Private Network) 4

  5. What Makes an Intelligent Control Plane? There are two elements to an intelligent control plane: � Intrinsic awareness of network resources – Network topology – Inventory management • Commissioned/de-commissioned port units • Available bandwidth (e.g. time-slots) � Intrinsic service activation – The network knows how to optimally route circuits in the network 5

  6. Control Plane Enabled Intelligent Network Elements Example of an I-NNI How it works: Each Network Element (NE) auto 0 discovers its port-to-port nearest 0 neighbor adjacencies Each NE floods the network domain with Link State 1 1 Advertisements (LSA) containing NE adjacencies Each NE uses LSA’s to build a 2 2 view of the network database (at least once every 30 minutes) The network database is used to 3 3 create “lowest cost” path between endpoints Network Auto Discovery An automatic, self-optimizing network Connection Management Auto-reroute shared mesh restoration Capacity Management 6

  7. Control Plane Applications (1) Accurate resource management in churn situations Applicable to mature NMS NMS NMS networks UNUSED UNUSED 7

  8. Control Plane Applications (2) UNI or E-NNI Rapid Provisioning of TDM slots Data organization uses UNI or E-NNI to commission/de-commission Customer OC-N Premises customer access service E-NNI E-NNI Automated Cross Domain Provisioning E-NNI is used for multi domain (even multi-carrier) connection management 8

  9. Control Plane Applications (3) � Mesh Networks – Operational Simplicity – Bandwidth Savings – Increased Reliability (Restoration On Top Of Protection) – Service Differentiation � Bandwidth On Demand / “Dynamic Optical Networks” – Traffic Patterns, Call Rates, Holding Times? – Subscribers (Number, Distribution)? – Growth Rates? � “Optical Virtual Private Networks” Via OMS Are An Alternative Solution. 9

  10. Standards Three Standard Groups 1. IETF’s GMPLS: Mostly driven by data (protocol view) focused entities. Requires proprietary protocol extensions to support heterogeneous networks 2. ITU-T’s ASON/GMPLS: More Practical Approach for transport networks, especially for multi-vendor/operator domains, heterogeneous networks 3. OIF: Details options in ITU-T Standards. Inter-working testing between largest suppliers & operators (Mostly Our Main Customers) => Go With ITU- -T/OIF Implementation T/OIF Implementation => Go With ITU But: Some Strongly Data Oriented Operators / Organizations But: Some Strongly Data Oriented Operators / Organizations May Not Be In Line With ITU- -T/OIF T/OIF May Not Be In Line With ITU 10

  11. Introduction – Global Standardization Work on the Distributed Control Plane Members: Members: IP Data Networking Telecommunication Community Sectors Driving Driving Force Force IETF OIF ITU http://www.ietf.org http://www.oiforum.com http://www.itu.int Bottom-Up • G-MPLS/ASON Common Ground Approach: • Accelerate uptake of optical MPLS WG CCAMP WG Study Group 13 & 15 Make use of networking existing • Publish Implementation protocols in Agreements Top-Down Approach: response to • Inter-operability Demonstration Produces protocols to meet MPLS MPLS-TE industry well-defined architecture requirements requirements O-UNI G-MPLS ASON SDH/SONET, IP Concepts [REFERENCE [SUITE OF SS7, ATM (P-NNI) PROTOCOLS] ARCHITECTURE] E-NNI ASTN G.8070 Distributed Control Plane Distributed Control Plane Distributed Control Plane ITU International Telecommunication Union IETF Internet Engineering Task Force ASON Automatically Switched Optical Networks MPLS Multi-Protocol Label Switching OIF Optical Internetworking Forum CCAMP Common Control and Management Plane O-UNI Optical User Network Interface MPLS-TE Multi-Protocol Label Switching – Traffic Engineering 11 E-NNI External Network Node Interface G-MPLS Generalized Multi-Protocol Label Switching

  12. Introduction – ASON Signaling Protocols Specifications � User Network Interface (UNI) : Operations between end-user and service provider administrative / control domains � OIF O-UNI: � Addresses the client/user signaling – i.e. The call management portion � Based on GMPLS signaling extensions / modifications to support O-UNI 1.0 / 2.0 � Supports both RSVP-TE and CR-LDP based signaling protocol options � Enhancements in O-UNI 2.0 (e.g. bandwidth modification, support for Ethernet) � Network-to-Network Interface (NNI) : Multi-control domain operation for a single service provider as well as multi-control domain operation among different service providers � OIF E-NNI: � Work is starting for the specification of an implementation agreement for E-NNI signaling specifications (close linkage between ITU-T G.7713.X series expected) � I-NNI : Intra-control domain operation � Network-to-Management Interface (NMI) : Operations between management systems and service provider administrative domains 12

  13. Introduction – The GMPLS/ASON Networking Models IETF G-MPLS User Administrative PEERING MODEL Domain IP Vendor B IP Ethernet Ethernet Tunnel Access Ethernet ATM NMI Across Optical ATM Node Network Operators A Operator A and C are Operator A O-UNI Administrative single domain Administrative networks Domain Domain Vendor A I-NNI (ONNS™) E-NNI I-NNI (ONNS™) E-NNI Core Node Operator B Operator B Operator C Administrative Vendor B Vendor C Operator C Administrative I P IP Administrative Domain Access Access Administrative E t h e r n e t Ethernet Domain Domain I-NNI (PNNI) A T M ATM Node Node Domain I-NNI (PNNI) E-NNI I-NNI (EMS) I-NNI (EMS) Vendor A Distributed Control Plane Distributed Control Plane Distributed Control Plane Core Node Operator B has divided G-MPLS Links network into multiple control domains Distributed Control Plane Distributed Control Plane Distributed Control Plane O-UNI (e.g. by vendor, geographical, political, User Administrative etc. considerations) Domain Vendor C I P I P Access E t h e r n e t Domain 1 E t h e r n e t A T M Domain 1 Node I-NNI E-NNI A T M I-NNI Domain 2 Domain 2 I-NNI ITU-T ASON I-NNI Client OVERLAY / DOMAIN Protocols BASED MODEL 13

  14. The ASON Model Architecture enabling boundaries for policy and information sharing – Separation of call and connection control – Inherent support for multiple address spaces – Inherent support for multiple domains, cross-domain call & connection mgmt. UNI UNI E-NNI E-NNI E-NNI E-NNI UNI UNI Domain 2 Domain 2 Domain 3 Domain 3 Domain 1 Domain 1 User 1 User 1 Connections Connections User 2 User 2 Connections Connections Connections Connections End-to-end call End-to-end call Connection Connection segments segments Example of call with multiple connection segments Example of call with multiple connection segments E-NNI applications examples include: Both the UNI and E-NNI are service demarcation points; i.e., call control is Difference in service realization provided Separate address spaces Independence of survivability Call properties must accompany the (protection/restoration) for each domain connection Trust boundaries 14

Recommend


More recommend