Cross Border CEF NetworksCesnet Experience Stanislav Šíma
Customer Empowered Fibre Networks across borders • Emerged Model for R&E networks (either Regional Optical Networks - RONs or National Research and Education Networks -NRENs) on Layer0: Dark fibre connections of near R&E premises with mesh topology • Emerging Model for R&E networks interconnection on Layer0 is nearly the same: Dark fibre connections of near R&E premises with mesh topology disregarding borders • Two or more physically different fibre routes to one premise is recommended to avoid single point of failure in both above cases • We can see this as dark fibre interconnections of Customer Empowered Fibre networks (CEF Networks) across borders
Day after RONs/NRENsinterconnection • Primary rule for CEF networks fibre topology is KIS (but now it means Keep It Short): fibre is the most expensive part of network (e.g. 70%) • Dark fibre topology is changing very slowly • We can imagine day after mesh interconnection of all RONs/NRENs by short dark fibres across borders: How to implement continental-wide services by means RONs/NRENs facilities and in what cases we need also parallel continental-wide fibre infrastructure? • CEF networks are constructed by many institutions now, so solution is more important
Railways example for Layer0 • Mesh network without parallel lines (with exceptions) • Parallel railway lines are very expensive ! • Reasons for parallel lines: • Transport speed, local traffic • Heavy workload • Different owners, etc. • Parallel lines are rare!
Layer0 optimization of Fibre Networks • Parallel fibres are very expensive and should be avoided (with exceptions) • User (for example university) is paying directly or indirectly for fibre in: campus network, metropolitan network, regional network, continent-wide network, world-wide network • Above networks has some parallel fibres or fibre segments • Forget borders and institutions for a while: Usershave economical interest to ask coordination of design and construction of Layer0 of all above networks (for example sometimes is possible to use OADMs instead of parallel fibres). • What could be reasons for parallel fibre lines? • Special good price conditions for fibre lease/owning (e.g. for short lines) • Special types of fibres • Very expensive equipment for some types of fibre sharing • others? • Organizational work should respect this technical and economical constrains.
Cross Border Fibres in European R&E – first steps • Our contribution to SERENATE study, p.24, April 22., 2003: „Prices of equipment are relatively important for the economics of transmission systems, but the overall costs heavily depend on the network architecture and topology. The basic approach is to connect geographically neighbouring universities by fibres.“ http://www.serenate.org/publications/d9-serenate.pdf • Cross Border Fibre (CBF) Brno-Bratislava connecting NRENs CESNET2 and SANET operated since April 2nd, 2003. • Results presented by Altmannova, L., Sima, S.: Fibres and advanced optical devices for a new networking strategy. TERENA TF-NGN Cambridge, September 16th, 2003. http://www.terena.nl/tech/task-forces/tf-ngn/presentations/tf-ngn12/20030915_SS_Optical.pdf • Proved impacts of CBF was at least: • improved services provided by NRENs • decreased overall costs of connectivity • strengthened collaboration of NRENs
CEF Networks in Central Europe now • CESNET2, Pionier, SANET, SWITCH is operating on fibres • Germany is deploying X-WIN • ACONET prepares fibre operation • ACONET, SANET, CESNET and Pionier are interconnected by short CBFs (for backup we use GEANT and other CBFs are in preparation) • Initial GEANT2 topology uses dark fibres (next page) • Discussion about future steps continues
Layers of networks interconnection • We discussed some topics of Layer0 topology, constrains and networks interconnection. Layer0 topology is limited by necessity of long term stability, changes are usually very expensive • Key to network interconnection is Lighting on Layer1 now • Layer 0 (dark fibre, photonic air links or wireless spectrum) • Layer 1 (primary wavelengths, wavelengths with constrains) • Layer 2 (1 or 10 Gigabit Ethernet lightpaths, SONET circuits) • Layer 2.5 (MPLS tunnels) • Layer 3 (IPv4/v6)
Open approach to Fibre lighting • Lighting on Layer1 enables rich service topology of CEF networks across borders, avoiding parallel fibres, providing primary and flexible wavelengths, etc. • Overcome distances by light only (thousands of kilometers are possible)! • New types of transmitters, receivers, amplifiers, gratings etc. are available, some of them even with MultiSource Agreements (this is very important for designers), and very cost-effective • Some of them have influence now (NIL approach, hut skipping DWDM, etc.). • Open approach to Fibre lighting is needed: no vendor is able to deliver all new types of lighting equipment on competitive price level
CzechLight network and prototypes • We use three stages of interoperability testing on Layer1: lab, testbed, pilot deployment • The same stages we use also for CESNET-made prototypes of lighting equipment build from commercially available top level optical elements • Prototype building and testing is technically progressive and cost-effective and gives excellent opportunity for some research universities and graduated students • Testing on fibre reels in lab is followed by deployment in our wide arrea testbed, i.e. CzechLight experimental & breakable network (see next pages) • Final prototype testing before industry production is pilot lighting of some dark fibre lines in NREN • For example, CESNET-made CzechLight Amplifiers are used now in CzechLight lines Prague – Brno and Prague - Plzen and in CESNET2 line Praha-Hradec Kralove • Many metropolitan CEF Networks has experience with network lighting by own staff • Open approach to fibre lighting is field proved, see presentations on CEF-Networks workshop http://www.ces.net/doc/seminars/20050516/ and SEEFIRE technical workshop http://www.seefire.org/schedule.php?ctn=1&language=en
Previous presentations available on-line I. Development of the CESNET2 optical network. TERENA Networking Conference, Limerick, June 4, 2002. (Nothing-In-Line approach + 189 km NIL operation announced) http://www.terena.nl/conferences/tnc2002/Slides/sl3c2.ppt Long distance fiber connections in NREN, TF-NGN Budapest, October 18, 2002. (10 Gb/s NIL possible up to about 300 km, single fibre bi-directional transmission in production network up to 125 km) http://www.terena.nl/tech/task-forces/tf-ngn/presentations/tf-ngn9/sima.ppt Fibres and advanced optical devices for a new networking strategy. TERENA TF-NGN Cambridge, September 16th, 2003. (Fibre is strategic asset. CEF networks concept. Free space optics for cost-effective first mile. Connect Near over Border by fibre.) http://www.terena.nl/tech/task-forces/tf-ngn/presentations/tf-ngn12/20030915_SS_Optical.pdf
Previous presentations available on-line II. Procurement and lighting of dark fibre. CEF Networks workshop, Prague May 2004. (Double payment of fibre infrastructure? NIL results. CESNET-made optical amplifiers deployment announced.) http://www.ces.net/doc/seminars/20040525/pr/CEF%20Networks3ext.ppt Towards a Nation-wide Fibre Footprint in research and education networking. TERENA Networking Conference, Rhodes, June 9, 2004. (Dark fibres brings progress to REN design and subsequently to WAN design generaly. From naive economy to strategy. National fibre footprint in European NRENs. GLIF and CzechLight.) http://www.terena.nl/conferences/tnc2004/programme/presentations/show.php?pres_id=97
I hope it helps a little… • Acknowledgement to colleagues from ACONET, CESNET, Pionier, SANET and SWITCH • Thanks for your attention