A Comprehensive Approach to Multi-layer Transport Network Automation

WHITE PAPER A Comprehensive Approach to Multi-layer Transport Network Automation Service providers continue to face staggering increases in bandwidth...
Author: Gwenda Phelps
49 downloads 2 Views 684KB Size
WHITE PAPER

A Comprehensive Approach to Multi-layer Transport Network Automation Service providers continue to face staggering increases in bandwidth, driven primarily by video and cloud combined with increasingly competitive broadband and mobile access speeds. This growth requires ongoing CapEx investment in the network at both the IP/MPLS layer (Layer 3/2.5) and the OTN and DWDM transport layer (Layer 1/0) so that service providers can continue to deliver high quality network experiences to their customers. Wall Street, on the other hand, continues to demand strong financial performance with new revenues and flat or declining CapEx-to-revenue and OpEx-to-revenue ratios.

This tension results in a need for service providers to break away from the status quo and evolve their network architectures so that they can scale while simultaneously creating new revenue sources and driving cost efficiencies. In 2013 Infinera unveiled the Intelligent Transport Network, which was built to help service providers with this challenge, enabling them to: •M  assively scale networks based on photonic integration, featuring platforms and an architecture that is designed for 100G and 500G DWDM super-channels today and terabit+ DWDM super-channels in the future. •C  onverge multiple layers of the network, such as DWDM and OTN switching, to further reduce both CapEx and OpEx costs (power, space, fiber patches, maintenance, etc.). •A  utomate the network with GMPLS and SDN to support rapid service delivery, resulting in a more competitive market posture for increased revenues complemented by single screen automation to reduce operational costs. The Intelligent Transport Network architecture allows network scale that not only supplies a more efficient transport layer but also leverages convergence and intelligence to make the IP/MPLS layer and thus the network as a whole more efficient by reducing the number of router resources needed to deliver a certain set of service demands.

EVOLVING TO THE TERABIT ERA 10G ’ 100G

100G ’ 500G ’ Terabit

Vision An Infinite Pool of Intelligent Bandwidth

DIGITAL OPTICAL NETWORK™

INTELLIGENT TRANSPORT NETWORK™ Automation

Open Software Control

Convergence

PIC

Multi-layer Switching DTN

ATN

DTN-X

Photonic Integrated Circuit (PIC) An industry first

Converged OTN Switching and WDM Without compromise

2003

2007

Scalability

Super-channel Transmission

2013

FIGURE 1: ITN Architecture Diagram

Page 2

2020

Intelligent Transport Network Evolution This architecture has seen wide adoption across the globe and is now evolving to extend the highly regarded ease of use of the Intelligent Transport Network into a dynamically reconfigurable flexible grid optical layer. Three key enhancements have been added: • Infinera’s third generation, 500G super-channel line cards have been introduced to support flexible grid operation. The scalable optical layer has now evolved from fixed grid to flexible grid to reflect the latest developments in the ITU-T DWDM standards. • A super-channel colorless, directionless, contentionless (CDC) FlexROADM has been incorporated to create a uniquely powerful multi-layer switching capability as an enhancement to, and working in concert with, Infinera’s existing non-blocking OTN switching capability. • The resulting flexible multi-layer digital and optical network is automated with a unified, carrier-grade control plane that represents the first production implementation of the Spectrum Switched Optical Network (SSON) extensions to GMPLS. This ongoing evolution has always been a part of Infinera’s vision for Intelligent Transport; and it benefits from the ability to work from a clean slate design—delivering a solution without compromises (no retrofits) in terms of scale, converged multi-layer switching, and end-to-end automation. While the flexible grid super-channel layer delivers vital additional capacity, this paper will focus on two specific capabilities of this architecture: converged multi-layer switching and unified automation.

Converged Switching: Phase 1 Chronologically, the first phase of this convergence was the introduction of a high-capacity (multiterabit), non-blocking OTN switch with the ability to switch up to 10 Tb/s in a single bay and 240 Tb/s in a multi-bay configuration. There were two major factors that enabled Infinera to deliver this convergence without compromise: • Large scale photonic integrated circuits (PICs) provide massive amounts of DWDM capacity in a small power and space envelope as well as the headroom to integrate the electronic switching functionality. • A system was designed from the ground up for this convergence (using custom-designed ASICs) versus a retrofit approach, seen with many other solutions on the market. This timing of this first phase was vital because the switch was purpose-built to groom large numbers of lower data rate services, such as 1 GbE and 10GbE, into coherent DWDM superchannels of 500 Gb/s and beyond. A key aspect of network planning is that there is no point in optically switching 100G or 500G super-channels if they are not efficiently filled.

Page 3

This means that each slot must be capable of providing 500 Gb/s access to the switch fabric. In fact, Infinera has designed the system with a backplane capacity that supports over 1 Tb/s per slot to support terabit+ super-channels in the future, potentially delivering long term investment protection. In addition, service providers can use advanced OTN switching functionality to leverage Infinera’s Fast Shared Mesh Protection (FastSMP™) to provide per-service granularity for sub-50ms service protection with excellent levels of bandwidth efficiency. FastSMP, featuring the FastSMP Processor™, is the industry’s only hardware accelerated shared mesh protection solution.

Converged Switching: Phase 2 Once the long-haul, coherent super-channels are efficiently filled via OTN switching and grooming, service providers need a way to manage this bulk capacity at the lowest possible cost, and to be able to shift it around the network as demand dictates. This is where Infinera’s up to nine degree super-channel CDC FlexROADM shines as a complementary tool. The OTN function fills the super-channels, and then, when needed, the ROADM can optically express these filled super-channels to lower network costs or to reroute the super-channels to meet changing network demands. ROADM technology is not new, but until recently ROADMs were based on fixed grid technology that cannot support flexible multi-carrier super-channel capacity. A FlexROADM is designed to operate on the new ITU-T G.694.1 Flexible Grid, using a building block granularity of 12.5GHz – which is ideal for the efficient support of coherent super-channels. Infinera’s super-channel CDC FlexROADM is available in a modular format that can be used to construct any configuration of basic, C, CD, and CDC FlexROADMs with up to nine degrees.

OTN SWITCHING • Point and click ODU0/flex switching granularity

MULTI-LAYER SWITCHING

• Digital grooming maximizes WDM fill->CapEx savings

Comprehensive tool Kit

• Sub-λ switching & FastSMP protection -> OpEx savings

• Fill super-channels with OTN mux/switching • Route optically to reduce regens & λ planning

OPTICAL • Wavelength & super-channel granularity • Optical express of filled super-channels -> CapEx savings

• All remotely configured with point-click GMPLS/SSON control plan • Maximum efficiency AND flexibility to remotely redeploy BW & new services

• Reconfigurable super-channel switching -> OpEx Savings

Figure 2: DTN-X Multi-Layer Switch Diagram

Page 4

Unified Control Plane

A CDC ROADM is designed to dramatically reduce the need for rigorous offline planning because it can be reconfigured at any time without any blocking conditions while allowing wavelength or super-channel switching operations to take place in a highly deterministic fashion. The end result is that service providers can make changes to the fundamental optical routing of traffic in the network remotely, and avoid expensive truck rolls.

The Best of Both Worlds As Figure 2 shows, by combining a high capacity, non-blocking OTN switch with a super-channel CDC FlexROADM, service providers can have the best of both worlds. Table 1 summarizes the individual use cases for one technology or the other, but the fact that both of these functions are co-located in the DTN-X node means that the network designer has the best of both worlds.

Requirement

OTN Switching

Optical Switching

Bandwidth granularity

Point and click provisioning and ODU0/flex switching granularity

Wavelength and superchannel granularity for bulk traffic management

CapEx saving

CapEx saving achieved by sub-super-channel service grooming for efficient fill

CapEx saving achieved by optical express of already efficiently filled superchannels

Restoration strategy

FastSMP sub-50ms protection at service level granularity

Optical restoration (several seconds) at the wavelength, super-channel or fiber level

Table 1: Service Requirements vs. Switching Architecture

A Unified Layer GMPLS/SSON Control Plane So, to realize the full value of the flexible digital and optical switching functions (data plane) it’s essential to develop a unified control plane so that these no-compromise switching tools can help automate day-to-day network operations. Generalized MPLS was designed to be able to control practically any underlying data plane—including packet, frame, TDM and wavelength technologies. Infinera’s existing and highly successful GMPLS control plane has been widely deployed for OTN operation in accordance with the GMPLS Framework. The SSON extensions allow GMPLS to extend its capabilities to a flexible grid super-channel data plane, thus creating an industry-leading unified control plane.

Page 5

DIGITAL DOMAIN

PHOTONIC DOMAIN

ADAPT

Automatic adjustment & monitoring of WDM parameters Ø “Hands-free” link turn-up & optimization Ø

Hello

Hello

AUTO-DISCOVERY Ø Ø

Topology Discovery Inventory and Configuration Management

NEW

GMPLS/SSON Ø Ø

Route Computation Point and Click Provisioning

NEW

Network Failure

RESTORATION Ø Ø Auto-Restored Trafic