JRA1 T2, Photonic Services What has been done

Transcription

1 JRA1 T2, Photonic Services What has been done SKALAT MAÐR RÚNAR RÍSTA NEMA RÁÐA VEL KUNNI. Joint JRA1 T1 & T2 workshop København, Denmark 2012, November Jan Radil, Josef Vojtěch, Pavel Škoda, Stanislav Šíma CESNET

2 Outline 1. Laboratory and Field Tests of Alcatel-Lucent100G Solution (Based on results DJ1.2.2) 2. Power Consumption (Based on results DJ1.2.2) (SA1, not JRA1) based on my NA4 T1 presentation from Jerusalem 4. NEAT-FT 5. PMD emulator for 10G (100G future) 6. PSs optical reach - in progress 7. New PSs subtasks

3 1. Laboratory and Field Tests of Alcatel-Lucent 100G Solution First verification of 100G coherent technology in CESNET Testing: lab, EF, CESNET2 ALU1830PSS together with Cisco MSTP and CzechLight Effects of filtration, non linearities 600km of different fibres in lab, 1063km in CESNET2 error free operation 100G coherent working with different fibres (G.652/G.655+/G.655-) and with DCFs and FBGs and with 10G NRZ wavelengths Works over single fibre bidirectional transmission too

4 2. Power Consumption,Green networking Analysis of power consumption (access, edges, core) Consumption Indicator Estimations to compare different equipment Comparison of CL, Ciena May be useful for future decisions

5 New types of applications with new requirements Not only fat 10/40/100G pipes but minimal/constant delays, jitter best effort principle not acceptable End-to-end connection between two or more places in network (OOO, OEO in special cases) - AWs Advantages (transparency, latency) and Challenges (reach, interoperability, ITU Black Link) Time/frequency tranfer (NEAT-FT) Real time apps (collaboration, C2C, remote control) Already tested/deployed, comparison of atomic clocks (CZ/AU), ultrastable frequency (FR, DE)

6 Photonic Service End-to-end connection between two or more places in network Described by Photonic-path and allocated bandwidth Photonic-path is a physical route that light travels from the one end point to the other or to multiple other end points respectively Allocated bandwidth is a part of system spectrum that is reserved for user of Photonic Service all along the Photonic-path. Minimal impact of network (no processing) on transmitted data Path is all-optical, no OEO except special cases.

7 Photonics vs Optics The word 'optics 'comes from the ancient Greek word ὀπτική, meaning appearance or look. Rather old word, a book by certain I.Newton called Opticks, 1704 The word 'photonics' is derived from the Greek word 'photos' meaning light Phos Φῶς (genitive: photos Φῶτῶς) means light appeared in the late 1960s to describe a new research field (invention of laser, laser diode, fibre etc).

8 Photonics vs Optics Photonics a new hot word

9 Advantages Transparency to modulation formats Low transmission latency as the shortest photonic path is formed Constant latency (i.e. negligible jitter), because non or only specially tailored electrical processing is present Stable service availability (due allocated bandwidth) with some exception for protection switching Future-proof design thanks to grid-less bandwidth allocation

10 Disadvantages Service reach in general is limited due to missing universal all-optical regeneration, but it can be extended by specialized OOO and/or OEO regenerators suitable just for limited number of applications. Potential waste of bandwidth. All-optical nodes should be grid-less and direction-less. In multi-domain scenario - absence of global management and operation system or communication between separate management systems. Multi-vendor network interoperability with AWs, although first tests were already successful, e.g. concurrent 100G and accurate time transmission and ITU-T also has produced recommendation G Black link.

11 Interactive human collaboration Latency jitter limit: ms (adaptive play-out delay buffer) End-to-end latency: ms Penalty: mild (user disappointment). Climate Refugee Opera?:-) What is minimum latency, considerable bandwidth and reliability? Always isuues... High definition video and Cave-to-cave Latency jitter limit: 20 ms (buffer dependent) End-to-end latency: 150 ms Penalty: mild (user disappointment).

12 Remote instrument control Latency jitter limit: 20 ms End-to-end latency: 100 ms Penalty: depends on application (can be severe in case of telesurgery) Remote control of vehicles Latency jitter limit: 50 ms End-to-end latency: TBD Penalty: not acceptable (vehicle crash).

13 Comparison of atomic clocks (time transfer) Latency jitter limit: 50 ps (short time, typ. over 1000 s) and 1 ns (long time fluctuation, typ. over days) End-to-end latency: should be minimized to the optical signal propagation delay Penalty: mild (experiment failure) - principal (service impossible) Ultra-stable frequency transfer Latency jitter limit*: NA End-to-end latency: should be minimized to the optical signal propagation delay Penalty: mild (experiment failure) - principal (service impossible) *The term jitter is not appropriate here. The phenomenon is rather expressed as a stability that should correspond to the stability of primary frequency standard, e.g in ultimate case of optical clocks.

14 Comparison of atomic clock scales on live network : CESNET (CZ) + ACONET (AT) Transmission of time marks (pulses modulated on optical carrier) Started by loop tests and GPS assisted transmission over standard DWDM systems, in 2010 Comparison of time scales between Czech and Austrian national time and frequency laboratories in Praha and Wien (IPE-BEV) over operational DWDM since Aug RUNNING

15 Comparison of atomic clock scales cont. Photonic path dedicated lambda over operational DWDM network: Mixture of fibre types (G.652/655) Mixture of transmission systems Cisco/Open DWDM Czechlight Mixture of CD compensation types (DCF, FBG) One way distance 550km, including 220km Noting-In-Line, total attenuation 137 db

16 Ultra-stable frequency transfers on live network: RENATER (FR) Transmission of ultra-stable CW optical frequency itself (in region 1550nm) Needs same path for both directions noise correction and propagation delay fluctuation compensation Datacom bidirectional devices must be bypassed (e.g. EDFAs) Source: G. Santarelli at al Transmitting ultra-stable optical signals over public telecommunication networks

17 Ultra-stable frequency transfers on live network: RENATER + LNE- SYRTE (Système de Référence Temps Espace) + LPL (Laboratoire de Physique des Lasers) km DF loop test only LPL-Nogent l Artaud-LPL 300km loop (228km over DWDM system), 100dB attenuation, 4 bidirectional EDFAs LPL-Condé/Reims-LPL 470km loop (398km over DWDM system), 136dB attenuation, 5 bidirectional EDFAs 540km loop (470km over DWDM system), 6 bidirectional EDFAs

18 LPL-Nancy-LPL 1100km with one regenerator station LPL-Strasbourg-LPL1476km with three regenerator stations RENATER: REFIMEVE+ Project: RENATER, LNE-SYRTE and LPL laboratories applied for REFIMEVE for building of national infrastructure on RENATER fiber, able to disseminate ultrastable frequency Planned start in 2012 Interconnections on cross-border fibers would also be studied

19 Ultra-stable frequency transfers: MPQ-PTB germany Max-Planck-Institut für Quantenoptik (MPQ) in Garching and Physikalisch-Technische Bundesanstalt (PTB) in Braunschweig, 2009 dedicated fibre146km Dedicated fibre, 920km, 200 db attenuation, bidirectional transmission and active stabilization 9x low noise bidirectional EDFA and Fibre Brillouin amplification with distributed gain Achieved stability 5 10e 15 in a 1-second integration time, reaching 10e 18 in less than 1000 seconds. Ref: A. Predehl at al A 920-Kilometer Optical Fiber Link for Frequency Metrology at the 19 th Decimal Place, Science 2012

20 bidirectional EDFA and Fibre Brillouin amplification not very common equipment Picture taken from: A theoretical analysis of amplification characteristics of bi-directional erbium-doped fiber amplifiers with single erbium-doped fiber Qinghe Mao, Jingsong Wang, Xiaohan Sun, Mingde Zhang Optics Communications

21 Real-world implementation of Photonic Services optical tunnel going through different networks, equipment (ROADMs), CBFs,... Examples from AT, DE, CZ, FR mentioned it is possible PSs in GEANT, GLIF, NRENs, MANs, LANs,.

22 It is possible, depending on applications of course. All-optical regenerators not available so accurate frequency will be difficult (accurate time OK, real-time OK). But accurate time should be doable (special OEO conversion is not critical). Transatlantic links are all-optical (no OEO), or at least some of them (2005?). All-optical reach up to Photonic applications (time/frequency/remote control) are not usually high speed (no 10G or even 40/100G) so transmission over long distances is not so critical (no problems with dispersions etc). One part of spectra for 10/40/100G and one part for Photonic services/applications. C/L band, Trial now in CESNET (L-EDFAs, L-transceivers splitters, etc.) C/C band for majority players, only C-EDFAs C/C band dense grid 100/50/25GHz

23 4. NEAT-FT The new EU NEAT-FT (FP7 JRP-s11) realized under EMRP (European Metrology Research Programme): Accurate time/frequency comparison and dissemination through optical telecommunication networks fundamental limitations of optical fiber links to the attainable instability and accuracy of optical reference frequencies and timing signals satellite transfer techniques developed over the last 40 years do not reach the performance required for modern clocks optical links will become the standard transmission tool for time & frequency comparisons and mandatory for a redefinition of the SI-Second demonstration of frequency transfer with 17 digits accuracy 10 partners, lead by PTB, DE Implementation on global scale? All-optical/AWs solutions not blocked by equipment.

24 4. NEAT-FT First introductory meeting in Praha (with DANTE, NRENs) And more details available on CEF2012 workshop Few NEAT-FT participants with interesting results NEAT-FT is about time/frequency only but: this si important for telco applications, power networks (electricity), aviation, navigation, financial transactions, astrophysics and other science disciplines, redefinition of SI units (second everywhere), homeland security, military (GN3 NA4 meeting, Jeruzalem, Sep2012),... Interesting reading: Tom O Brian, NIST: Time, Timekeeping and Time Distribution

25 4. NEAT-FT

26 5. PMD emulator for 10G (100G future) Emulation of PMD Done for 10G Prepared for 40G (noncoherent) and 100G coherent Important for PSs too

27 6. PSs all optical reach - in progress Photonic Service all optical end to end connection All optical reach of Photonic Services is important But PSs may differ significantly from standard Ethernet/OTN services Estimation based on OSNR OSNR is the critical factor, PSs are usually low speed and CD/PMD is less critical but CD/PMD must be considered too Other theoretical AND practical verifications of influence of PSs on 1/10/100G production traffic ie mix of PSs, Ethernet, OTN in one fibre

28 7. A new subtask within JRA1 T2 JRA1 T2 subtask E2E photonic services between user premises Start 1/Apr/12, duration 12 months Objectives provide feasibility and demonstration study of e2e photonic services between user premises strengthen research collaboration with vendors experienced in photonic service deployment evaluate feedback from GN3 NRENs concerning innovative transmission services evaluate feedback from research projects interested in photonic services demonstrate e2e photonic services between user premises if feasible in Y4

29 7. A new subtask within JRA1 T2 PSs feasibility provided international connection AT-CZ for atomic clocks comparison This is a demonstration study which can be used for other users Other CZ institutions connected with DFs and PSs available (CEF2012) DE and FR have their own experience (CEF2012) Some vendors contacted and background for PSs discussed Not always easy because nature and background of PSs are not seen as standard, sometimes confusions exist There are some new network designs possible with new coherent equipment, some arguing just false, especially with so called DCM-free designs which may block PSs and other services NRENs contacted and their feedback evaluated Many of them expressed interests in PSs, more details required Feedback from research projects big candidate NEAT-FT Older fibre links from GN3 could be used for PSs, details now available and will be processed

30 7. A new subtask within NA4 T1 JRA1 T2 subtask E2E photonic services between user premises There is another PSs-related subtask, NA4 T1 (dissemination) Photonic services enable advance in research Objectives promote photonic services as a future trend of innovative networking in Europe and beyond support usage of photonic services in advanced research disciplines dissemination of JRA1 T2 results promote effective lighting of acquired dark fibres (also in less developed regions) strengthen liaison with world-wide partners in innovative networking 9 international presentations accepted including US, Japan, Brazil... So hopefully we are going in the right direction!

31 Q&A Thank you for your kind attention! Thanks to other colleagues from different projects.