SURVIVABLE SUBMARINE OPTICAL NETWORKS

SURVIVABLE SUBMARINE OPTICAL NETWORKS Intelligent Optical Switch Control Plane Enables Resilient Submarine and Terrestrial End-to-End Worldwide Networks Globalization has changed the nature and relevance of communication networks dramatically. This change is evidence that communications systems have become the most significant element of mission-critical infrastructure for consumer, business, and government market segments worldwide. While carrier-class service has become a minimum network requirement for delivering services, its limitations become clear when disrupted by human error, acts of war, terrorism, or natural disasters. Minimizing the risk of such disruptions requires a new approach to building a survivable optical network in both submarine and terrestrial applications and, ultimately, a combined global network. This paper discusses the challenges and solution for building highly resilient subsea networks with high availability using industry standards-based intelligent control plane technology. It explores the migration path toward mesh networking, and examines how networks are evolving toward a global mesh to improve network survivability, increase network automation, and reduce costs. The Challenge of Architecting Resilient Submarine Networks Network globalization has been an ongoing strategic goal for carriers and network operators worldwide. Governmentfunded research and education networks rely on submarine and terrestrial infrastructures for global data interchange during bandwidth-intensive experiments, many of which share supercomputers and cannot tolerate service interruptions. Likewise, large financial organizations Ciena Intelligent-Optical Switch Mesh Networks at a Glance: > Resilient Proven six-9s availability in large terrestrial networks, with the ability to survive multiple simultaneous failures > Proven Both submarine and terrestrial mesh networks in operation today > Robust World s largest single-domain mesh network, with over 800 nodes in current operation > Scalable Switching capacities of 640 Gb/s to 3.6 Tb/s today, scaling to 15 Tb/s in the future > Flexible Seamlessly migrates from ring to mesh or multi-tiered ring-mesh architectures > Intelligent automation Standards-based intelligent control plane discovers new nodes, ports, and circuits, automatically reconfiguring the network upon link failures transacting in world markets depend on global telecommunications infrastructures and require low latency with close to 100 percent uptime, as even a few seconds of network failure can cost millions of dollars. Enterprises with global presence rely on submarine networks to connect their globally distributed offices, R&D centers and manufacturing plants for day-to-day operations. Submarine networks are even needed to transport medical imaging scans to radiologists halfway around the world, as distance radiology is becoming a commonplace method for handling off-hour emergencies. In short, submarine networks play an increasingly critical role to the stability, longevity, and, in some cases, survival of global companies and human beings. High-profile natural disasters challenge the reliability of submarine networks. As an example, the March 2011 earthquake and tsunami in Japan damaged about half of the existing cables running across the Pacific. Additional examples, such as the December 2006 Hengchun earthquake off the coast of Taiwan and the January/ February 2008 cable cuts in the Mediterranean Sea and Persian Gulf (believed to be caused by ship anchors), underscore the need for increased resiliency in global networks. Due to the critical nature of international bandwidth, many operators are now looking toward mesh restoration to guarantee network resiliency instead of relying on pre-existing submarine cable network restoration plans and their associated high costs, limited protection, and slow restoration times. W Whitepaper

Unlike most terrestrial networks, submarine networks are located in geographies subject to a high propensity for cable faults and a lack of route diversity, making architecting carrier-class resiliency in these networks challenging. Traditional transoceanic and terrestrial ring protection schemes typically do not protect against multiple failures, which are more likely to occur during natural disasters when cable routes have limited route diversity. In the case of the March 2011 earthquake, the repair times were quoted to be between a week and several weeks, while in the case of the Hengchun earthquake, seven of the nine cables went out of service for an extended period of time, with the last repair completed 35 days after the earthquake occurred. The lack of route diversity meant the loss of both legs of many of the rings, resulting in traffic loss throughout the region. Traffic was also severely affected over a two-week period in the Middle East and India when ship anchors allegedly cut as many as four cables in the Mediterranean Sea and Persian Gulf. Cable cuts, mainly from ship anchors and fishing trawlers, are common, with approximately 50 undersea cuts every year. For these reasons, submarine networks are migrating towards shared protection paths over multiple geographically separated paths to significantly improve resiliency and reliability. The Solution The mission-critical nature of worldwide data transfer from global customers demands the same resiliency and reliability in both submarine and terrestrial networks. The lack of routes and route diversity provides a unique challenge in architecting resilient submarine networks. As shown in Figure 1, mesh architectures have proven more resilient than traditional ring architectures in terrestrial networks, improving availability to six-9s a factor of 10. 100.0000% 99.9999% 99.9998% 99.9997% 99.9996% 99.9995% 99.9994% 99.9993% 99.9992% 99.9991% 99.9990% Six-9s Measured service availability 2006 2007 2008 2009 2010 Figure 1. Six-9s availability measured in large Ciena terrestrial mesh network. 15 Tb/s to handle increased loads as traffic demands increase. The benefits of an automated control plane include improved network resource utilization, simple provisioning, and real-time inventory management, lowering CAPEX, reducing provisioning and/or change intervals from months to minutes, and allowing carriers to optimize time to revenue. Figure 2 shows Verizon Business transatlantic mesh network, which is based on Ciena s intelligent control plane and optical switching technology. Using traditional ring protection, a double fiber cut on the north and south routes of TAT-14 would put the network out of service. However, because the mesh network employs automated control plane-based restoration, Verizon can share restoration capabilities across all undersea cables it accesses in the Atlantic basin. This capability allows the network to survive multiple simultaneous failures including the one referenced above while simultaneously increasing wavelength utilization compared to traditional ring-based architectures. Global Mesh Networks with Intelligent Control Plane Technology Ciena s CoreDirector FS Multiservice Optical Switch was the first to implement and deploy control plane functions for discovery, routing, and signaling. CoreDirector FS signaling and routing protocols have been deployed by over 35 carriers around the world with over 800 nodes in a single global mesh network. This field-proven control plane technology is also integrated in Ciena s ActivFlex 6500 Series and ActivFlex 5400 Series, offering capacities between 640 Gb/s to 3.6 Tb/s today, and are scalable to Apollo North Apollo South Flag FA1 North TAT-14 North Flag FA1 South TAT-14 South Figure 2. Verizon s transatlantic mesh network. 2

1. Service originally on black path 2. Black path disabled, service restores on green path 3. Green path disabled, service restores on red path Figure 3. Global mesh-based restoration in action. The success of this control plane-based mesh architecture led Verizon Business to deploy a mesh architecture in the company s transpacific expansion, which was put to the test in August 2009, when an earthquake struck the Luzon Strait south of Taiwan. Within milliseconds of this major earthquake, which in total caused 20 faults on 10 cables, all of Verizon s traffic was rerouted. This network architecture proved invaluable once again during the March 2011 9.0 magnitude earthquake in Japan, where any impact to customer traffic was averted using the meshed infrastructure. Ciena s intelligent control plane-based mesh restoration solution is currently deployed in several transatlantic mesh networks as well as in transpacific and regional submarine networks routes. In an effort to improve global resiliency, carriers are combining both terrestrial and submarine mesh domains to create one large, resilient, mesh-based global network. FastMesh Restoration in a Global Network FastMesh, the industry s only 50 ms mesh restoration implementation proven to work in live networks of hundreds of nodes, can be deployed alone or concurrently with traditional SONET/SDH linear and ring protection, to provide the highest levels of service availability. FastMesh continually monitors the network and makes restoration connections on an end-to-end basis. In case of a catastrophic event, this function continues to restore network connections, ensuring the highest priority traffic is carried first despite multiple network failures. The restoration function operates independently of underlying linear and ring protection schemes, offering a multi-tiered protection hierarchy for each end-to-end connection. Figure 3 illustrates how FastMesh routes around a single network failure in a global mesh network. Traditional ring networks are designed to restore service quickly in the event of a single failure, but usually cannot manage a second failure, requiring manual intervention. Mesh-restored networks can handle multiple failures, maximizing resiliency and providing customer satisfaction and continuous revenue generation. This important benefit is leading carriers to migrate their global networks to mesh architectures. As described previously, the Mean Time To Repair (MTTR) for submarine failures can take weeks, leaving the protection leg of a ring unprotected during repair. Unless the network is re-architected during this period, a second failure would result in a service outage. Mesh architectures automatically reconfigure the network to prevent service outages if another failure occurs during repair, without manual intervention in the network. In long-distance submarine or global networks, high network availability is difficult to achieve using rings. Availability is distance-dependent due to factors that include the number of fiber cuts-per-kilometer and the number of amplifier failures. Network availability is reduced further by the long MTTRs characteristic of submarine networks. Multi-degree resiliency of mesh networking is the only way to achieve high availability in large transoceanic and global networks. 3

Migration from Rings to Mesh Submarine networks often do not have the initial connectivity GbE to support a mesh architecture. As connectivity and nodes are added to the network, ring architectures can migrate gracefully OC-3 to mesh restoration and multi-tiered protection, further enhancing network resiliency. This additional mesh connectivity, or parts of it, OC-192 can be terrestrial-based, as in the Verizon Business network shown in Figure 2. The flexibility of FastMesh enables macro-level network protection, granular protection by circuit, and predetermined routes for the most rapid restoration. Customers can purchase varying class-of-service levels congruent with their Service Level Agreement (SLA). Network operators benefit from high levels of protection and resultant high margins. Ciena s Intelligent-Optical Switching portfolio including the ActivFlex 5400 Series, ActivFlex 6500 Series, and the Core Director FS provide multi-tiered protection and link aggregation thus enabling the most flexible and resilient network available today for submarine applications. Link Aggregation Link aggregation allows multiple links between two nodes to be managed as a single link, and supports intelligent network scaling and faster restoration times in addition to simplifying the network. For customers requiring submarine cable diversity without the complexity of managing multiple circuits, link aggregation is ideal. Submarine cable restoration architectural schemes, typically controlled only by the cable operator, may be expensive and limit restoration to only one other cable system. Ciena s FastMesh can aggregate capacity from multiple cable systems into a bandwidth pool and optimize usage and restoration decisions, lowering resource costs, providing greater levels of flexibility, and creating the highest level of network survivability. An example of multiple circuits sharing undersea cable capacity is shown in Figure 4. These unique protection and restoration capabilities are enabled through link aggregation and Ciena s intelligent control plane the most advanced, deployed, and robust control plane available today. Cable System Cable System Cable System #1 #2 #3 Submarine Mesh Network 1GbE STM-1 STM-64 Figure 4. Resiliency-enhancing link aggregation combined on shared paths in a mesh architecture. Enabling Mesh Networking through the Intelligent Optical Control Plane The intelligent control plane enables the creation of a fully automated meshed optical network, the single most important feature of a survivable optical network. An intelligent control plane, in very simple terms, is software that controls all configurable features of a network element and/or an entire network. This capability includes the simple configuration, activation, and deactivation of circuits from individual circuits to a full mesh that make up an entire network. Fully meshed networks enable interconnection of a single node to every other node. Managing such a complex network requires a control plane that, given some parameters, can make decisions on its own. Known as an intelligent control plane, this software-based automation helps improve network survivability, scalability, and cost. For example, when a Ciena intelligent switching network is expanded, the inventory of available circuits and ports is discovered automatically and placed into context within the entire network in which it functions. Once installed, the new capacity is made available for new services, and any new circuits are added automatically to the pool of circuits that may be used for restoration. Ciena s standards-based control plane manages this process in real time, automates circuit provisioning, and eliminates manual inventory management. Software which requires years of field experience to achieve high levels of reliability, resiliency, and efficiency is key to this automation. 4

The International Telecommunications Union (ITU) established the Automatically Switched Optical Network (ASON) standard to guide the development of common optical control planes for intelligent optical networks. ASON requirements and recommendations, per ITU-T G.8080, describe how a suite of control plane protocols should react to service requests, automatically provisioning end-to-end network resources across a multi-technology, multi-vendor optical network. This provisioning allows the network s scope and capabilities to be increased without a corresponding increase in manpower. Ciena s intelligent control plane software is compliant with the ITU-T G.8080 standard, and is field-proven for feature-rich functionality, reliability, and scalability based on more than ten years of live network deployment and operational experience. Multi-vendor Interoperability Ciena s intelligent control plane is a standards-based solution that supports the Optical User-to-Network Interface (O-UNI) and the External Network-to-Network Interface (E-NNI), as defined by the Optical Internetworking Forum (OIF), enabling Ciena s optical switching platforms to extend intelligence and provision services across a multi-vendor network. The OIF has created O-UNI specifications to allow subtended network elements which also support O-UNI to request the setup or tear-down of light paths across the switched network. Upon receiving a request from a subtended network element, the optical control plane will perform the operation automatically. Automation of the request, setup, and tear-down processes reduces operational costs and allows rapid response to service requests and generation of additional revenue. Optical control plane interoperability, based on OIF E-NNI/ITU-T G.7713.2, expands the end-to-end automated provisioning capabilities of the intelligent optical network across disparate optical control domains. This interoperability allows rapid service deployment worldwide across multiple carrier networks and varied vendor equipment. E-NNI is an ASON standard for control plane signaling and routing of Label-Switched Paths (LSPs) between the optical control domains of a carrier s network, or between the network and another carrier s network. customer after deploying an optical switching solution from Ciena. Average service provisioning time decreases significantly as the coverage of the intelligent network increases. Reducing the time required to provision services results in incremental revenue generation. In addition, operators experience a dramatic decrease in errors as service activation eliminates human intervention and thus human error. Automated service activation can reduce time to revenue and service provider operational expenses associated with manual provisioning. This automated process enables a very simple management structure, eliminating the typical problems associated with connectivity between different nations and carriers. Summary Globalization Requires Survivable Networks Networks in need of near-zero downtime include educational entities, large enterprises, multi-national corporations, medical services, and large financial institutions. Operators rely even more on global infrastructures during catastrophic failures, to sustain and grow high-margin revenue streams, such as those associated with global voice and data services. An intelligent control plane expands carrier-class operator networks to help operators thrive, scale new services, lower costs, and most importantly survive single, multiple, or even catastrophic failures. As communication networks continue to grow, businesses that implement survivable optical networks will garner higher margins for services, reduce CAPEX and OPEX, and retain the best customers. Average Provisioning Time* Months Weeks Days Reduced OPEX via Automation The automation of operating functions such as service provisioning, equipment inventory, and topology updates allows carriers to reduce the number of network operation personnel and respond more quickly to requests for new services, moves, teardowns, and modifications. Figure 5 shows the change in service provisioning time for an existing Hours Minutes 100% Legacy Network 100% Intelligent Optical Network Network Evolution *Assumes access is available Figure 5. Provisioning time comparison. 5

Ciena s ActivFlex 5400 series, ActivFlex 6500 Series, and CoreDirector FS possess the technologies necessary for implementing survivable networks. As a result of the company s innovation in intelligent switching, Ciena is the undisputed market leader in mesh network deployments. Ciena s optical switching systems are deployed in some of the world s largest carriers, cable Multiple Service Operators (MSOs), and submarine cable systems, and enable survivable optical networks by offering the industry s most robust control plane technology. Customers are reaping the financial and operational benefits of Ciena-enabled intelligent networking functionality and proven reliability for carrier-class networks. Ciena may from time to time make changes to the products or specifications contained herein without notice. Copyright 2011 Ciena Corporation. All rights reserved. WP023A4 6.2011 Specialists in unlocking network potential to help you change the way you compete. 1201 Winterson Road Linthicum, MD 21090 1.800.207.3714 (US and Canada) 1.410.865.8671 (outside US and Canada) +44.20.7012.5555 (international) www.ciena.com