Optical Futures: high-performance networking for research and education and optical technologies Heather Boyles Director, International Relations Internet2 heather@internet2.edu 18 March 2016 18 February 2003 Hong Kong Internet2 Mission and Goals Develop and deploy advanced network applications and technologies, accelerating the creation of tomorrow’s Internet. • Enable new generation of applications • Create leading edge R&E network capability • Transfer technology and experience to the global production Internet 2 Internet2 Areas of Work Advanced Applications Middleware Network Engineering • End to End Performance Advanced Network Infrastructure Partnerships and Outreach 3 Internet2 Network Architecture GigaPoP One GigaPoP Two Internet2 Backbone Networks GigaPoP Three GigaPoP (n) 4 Internet2 Network Architecture Internet2 Backbone Network(s) University A GigaPoP One Regional Network University B University C Commercial Internet Connections 5 Abilene Network Core Map, January 2003 6 Abilene Network Logical Map 7 What is optical networking? Utilizing optical fiber to carry light as the lowest level data transport medium Good source for basics of optical networking: http://www.sura.org/opcook 8 How is optical networking being discussed in research networking community? Currently, two main threads of activity: 1) Obtaining and controlling fiber assets in order to build networks (at campus, metropolitan, regional, national levels) 2) Utilizing that control of optically-based transport layer to provision new types of services – “lambdas**” or “lightpaths” ….and sometimes simulating these without underlying control **”a pipe where you can inspect packets as they enter and when they exit, but principally not when in transit” (C. deLaat et al.) 9 How have we provisioned networks in the past? To date, primarily by buying services from telecommunications companies • ATM service, SDH/SONET service, GigabitEthernet Services, wavelength services Example: Abilene backbone network • Qwest provides 10Gbps wavelength between core nodes • Abilene owns, controls routers 10 Providing network infrastructure by acquiring fiber assets Unique window in time for fiber assets • Cause: fiber glut, bankruptcies and telcos in distress • Within a year, opportunity on national scale closes? • Hedge against a regression to ‘bad old days’ of monopolies Technically, getting fiber means controlling the network down to layer 1 (0?) • Would allow deployment of different wavelengths for differentiated networks (high perf advanced services, network research, more general EDU access) • Path to doing optical switching when it makes sense 11 Unique optical requirements in Higher Education Community 10-Gbps: 10 Gigabit Ethernet preferred over OC-192c SONET HPC could need 40-Gbps λ’s prior to the carriers Integrated view of network management • Transport & IP engineering/operational approaches are not intrinsically different • SNMP preferable for network polling HEC can provide experimental environment for development of ‘rational’, customer focused optical switching • Switching tightly integrated with optical transport • Capacity for IP backbone expansion and p2p λ’s 12 Metro/regional implementations lead (for technical and economic reasons) Distance scale (km) Examples Equipment UW(SEA), Metro < 60 State/ Regional < 500 Dark fiber & end terminals USC/ISI(LA) I-WIRE (IL), Add OO CENIC ONI, amplifiers I-LIGHT (IN) PLR, Extended Regional/ National > 500 TeraGrid Add OEO Abilene regenerators National Light Rail & O&M $’s Variety of campus, metropolitan, regional optical networks in US CALREN2 – California IWIRE – Illinois ILIGHT – Indiana SURA Optical Cookbook examples • Harvard Joint Trench Project • Southern Crossroads (Atlanta-area dark fiber buildout) 14 ORION Network – Overview a Canadian (Ontario) example Note.-This overview includes ORION PoP’s, associated RANs, and other ORANO members. Other members could be 50 – 60 R&E sites to connect to the nodes sureNet LaurentianU OMAN CambrianC CBoreal LakeheadU CarletonU UOttawa NorthernC AlgonquinC LaCiteC ConfederationC NipissingU LaurentianU SaultC CanadoreC LambtonC GeorgianC ConestogaC UWaterloo WLU UGuelph SSFlemingC CentennialC CGrandLacs GBrownC HumberC SenecaC TrentU LoyalistC Queen'sU RMC SLawrenceC DurhamC SheridanC LARG*net UWO FanshaweC McMasterU OCAD RyersonPU UToronto YorkU MohawkC BrockU City / Town with University only City / Town with College only WEDnet City with CA*net 3 GigaPoP UWindsor St.ClairC LEGEND City / Town with College and University College [25] University[19] Potential ORION backbone NiagaraC Source: Randy Neals, ORANO 15 Regional optical fanout In the next generation architecture, regional & state based optical networking projects are critical Three-level hierarchy remains vital • National backbone, GigaPoPs (ARNs), Campuses Close collaboration with the Quilt GigaPoPs • Regional Optical Networking project 16 Toward a National Optical Networking Facility Research and education community investment in national-scale fiber assets • Discussions among a number of partners in US ongoing – “National Light Rail” – being led by members of Internet2 community – CENIC, the Pacific NorthWest Gigapop and other partners – SURA – USAWaves project 17 CA*net 4 Architecture source: Bill St. Arnaud Edmonton Saskatoon Winnipeg Vancouver Calgary Kamloops Regina Halifax Thunder Bay St. John's Victoria Quebec City Sudbury Seattle Charlottetown Montreal Ottawa Minneapolis CA*net 4 Node Fredericton Toronto Buffalo London Possible Future Breakout Possible Future link or Option Halifax Kingston Hamilton Boston Albany Windsor Chicago New York CA*net 4 OC192 18 Light Path Scenarios Workstation to Workstation Wavelength CWDM GigaPOP to GigaPOP Wavelength Regina BCnet Vancouver Campus OBGP switch Winnipeg St. John’s RISQ Halifax Calgary Montreal Seattle Toronto Lambda* or Lightpath Networks Current CA*net “customer-empowered networking” - prototypes a day when multiple wavelengths available to a site/desktop Why? • Router limitations - cost • A few very bandwidth needy applications (e.g. between radio astronomy sites) for which dedicated circuits make more (technical, economic) sense • User control – configurability An area needing more investigation, cost modeling, prototyping, etc. 20 Summary There are many networks out there pursuing the dark fiber opportunity • Scale is important • Regulatory factors • Take advantage of others’ experiences! The “lightpath” model will be exciting area of investigation • What are the right models? “Empowered customer” pays traditional circuit-switched network-type fees? 21 www.internet2.edu 22