R esearch and E xperimental A ssessment of C ontrol plane archi T - - PowerPoint PPT Presentation

▶
r esearch and e xperimental a ssessment of c ontrol plane
SMART_READER_LITE
LIVE PREVIEW

R esearch and E xperimental A ssessment of C ontrol plane archi T - - PowerPoint PPT Presentation

Jul 01, 2023 289 likes •582 views

R esearch and E xperimental A ssessment of C ontrol plane archi T ectures for I n- O peration flexgrid N etwork re-optimization (REACTION) Filippo Cugini, Luis Velasco, Juan Pedro Fernandez-Palacios Copenaghen, November, 2014 Outline Project

slide-1
SLIDE 1

Filippo Cugini, Luis Velasco, Juan Pedro Fernandez-Palacios Copenaghen, November, 2014

Research and Experimental Assessment of Control plane archiTectures for In-Operation flexgrid Network re-optimization (REACTION)

slide-2
SLIDE 2

2

Outline

  • Project overview
  • Implemented project tools and facilities

– OPNET model – Distributed testbed

  • Some project achievements on advanced use cases

– Slice-ability – After failure repair optimization – Multipath restoration and bitrate squeezing

slide-3
SLIDE 3

3

Participants

Participant no. Participant organisation name Participant short name Country 1 (Coordinator) Consorzio Nazionale Interuniversitario per le Telecomunicazioni CNIT Italy 2 Universitat Politècnica de Catalunya, Barcelonatech UPC Spain 3 Telefonica TID Spain

slide-4
SLIDE 4

4

Abstract

  • REACTION targets the design and validation of a flexible
  • ptical network enabling software-controlled super-

channel transmission.

  • The focus is on:
  • 1. Advanced bandwidth variable transponder (BVT)

functionalities supporting multi-carrier transmission, adaptation of transmission parameters (mod. format, spectrum allocation, coding/FEC) and slice-ability

  • 2. Advanced control plane architecture and functionalities

including innovative two-level active stateful PCE including the BGP-LS advertising solution

  • 3. Advanced routing and spectrum assignment (RSA)

algorithms

Activity lead by CNIT Activity lead by TID Activity lead by UPC

slide-5
SLIDE 5
  • 5
  • C programming language
  • OPNET: Event driven

network framework

Tool: OPNET Model

slide-6
SLIDE 6
  • 6

Functionalities:

  • Link state advertisement
  • Spectrum availability adv.
  • Extensions for flex-grid
  • Routing and Spectrum Assignment

OPNET Implementation (1/3): OSPF-TE

slide-7
SLIDE 7
  • 7

Functionalities:

  • Reserve/Release spectrum slots

along the path

  • Extensions for flex-grid

OPNET Implementation (2/3): RSVP-TE

slide-8
SLIDE 8

8

OPNET Implementation (3/3): PCE and PCEP

Finite state machine (FSM) of the root process FSM of the child process: specific for each PCE-PCC session.

slide-9
SLIDE 9

9

Distributed control plane implementation

  • Extended for flexi-grid
slide-10
SLIDE 10

10

Implemented control plane architecture

  • Active stateful front-end PCE, in charge of computing RSA and elastic

provisioning

  • Back-end PCE responsible for performing complex network operation s

(e.g., re-optimizations)

  • To provide the back-end PCE with updated network topology info, we

propose to rely on the North-Bound Distribution of Link-State and TE Information through BGP, known as BGP-LS.

Back-end PCE TED LSP-DB Active Solver PCEP Server PCEP BGP-LS PCEP Server Front-end PCE Provisioning TED LSP-DB Inventory

slide-11
SLIDE 11

11

Testbed resources

OXC1 OXC2 OXC3 OXC4 OXC5 p1 (16Q 16QAM) p2 P3 (QP QPSK)

Tb/s Tx/Rx Tb/s Tx/Rx

PCE TED

CNIT:

  • IP/MPLS network testbed composed of six

IP/MPLS routers (Juniper M7i/M10, Cisco 7206);

  • traffic generator/analyzer,
  • 3 ROADMs and 3 flexi-grid WSS;
  • Optical terabit/s TX with coherent RX
  • Advanced control plane including GMPLS and

PCE as well as SDN controller and agents. UPC: – Back-end PCE The test-bed supports the most advanced RSA algorithms and interoperates with the PCEs of the other partners. Telefonica I+D:

  • Flexgrid testbed with Flexgrid ROADMS from

Cisco

  • control plane emulator supporting multiple

domains and including GMPLS and PCE extensions for Flexgrid.

slide-12
SLIDE 12

12

Slice-ability (1/2)

  • Flexibility to cope with traffic increase

– support connections to different destinations, each served by a sub-set of sub-carriers.

Sliceable functionality applied to a single four- carrier SBVT

37,5GHz

150 GHZ Year 1 Four different destinations 100G 100G 100G 100G 125 GHZ 200G 200G Year 2 Two different destinations 100 GHZ 400G Year 3 Single destination

slide-13
SLIDE 13

13

Slice-ability (2/2)

37,5GHz

150 GHZ Year 1 Four different destinations 100G 100G 100G 100G 125 GHZ 200G 200G Year 2 Two different destinations 100 GHZ 400G Year 3 Single destination

Single path with low spectrum use vs. multi-path with overall larger spectrum?

s d

  • Flexibility in provisioning and recovery
slide-14
SLIDE 14

14

Performance evaluation: Provisioning

Sliced Adaptive No Sliced

slide-15
SLIDE 15

15

Performance evaluation: Recovery

Sliced No Sliced Adaptive

[Ref] M. Dallaglio, A. Giorgetti, N. Sambo, F. Cugini, P. Castoldi, “Impact of slice-ability on dynamic restoration in GMPLS-based Flexible Optical Networks” Optical Fiber Communications Conference (OFC), March 2014

slide-16
SLIDE 16

16

Slice-ability on different SBVT architectures

  • Sliceable BVT architecture can be implemented either with
  • 1. Array of N tunable lasers
  • 2. A single tunable multi-wavelength source (generated from 1 laser)
  • N lasers guarantee full and independent tunability of each sub-carrier

 no constraint s on RSA

  • Conversely, a MW source only supports contiguous frequencies

 RSA constraints

  • A MW source is expected to be to be cheaper (reducing the number of

lasers), with lower footprint and lower power consumption.

  • Moreover a MW source enables better frequency stability among sub-carriers

 less spectrum.

slide-17
SLIDE 17

17

Performance evaluation: Recovery with MW

[Ref] M. Dallaglio, A. Giorgetti, N. Sambo, P. Castoldi, “Impact of SBVTs based on Multi-wavelength Source During Provisioning and Restoration in Elastic Optical Networks”, ECOC Conf, Sept. 2014

slide-18
SLIDE 18

18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1-2 2-3 3-4 4-5 1-6 6-7 7-5 1-8 6-8 7-10 5-10 8-9 9-10

Re-optimization Restoration

  • 8
  • 9
  • 10
  • 2
  • 4

1

  • 6

5

  • 7
  • 3
  • a)
  • 8
  • 9
  • 1
  • 2
  • 4

1

  • 6

5

  • 7
  • 3
  • b)
  • 8
  • 9
  • 1
  • 2
  • 4

1

  • 6

5

  • 7
  • 3
  • c)

4 slices 4 slices 4 slices 8 slices 2 slices 4 slices 8 slices

  • Frequency slice
  • Optical Link

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1-2 2-3 3-4 4-5 1-6 6-7 7-5 1-8 6-8 7-10 5-10 8-9 9-10

  • Frequency slice

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1-2 2-3 3-4 4-5 1-6 6-7 7-5 1-8 6-8 7-10 5-10 8-9 9-10

  • Frequency slice

P1 P2 P2 P3 P4 P5 P5 P1 P3 P2a P4 P5 P5 P1 P2 P3 P4 P5 P5 P2b P2b P2b P2

P1 P2 P1 P1 P2 P2b P2a

After failure repair optimization

slide-19
SLIDE 19

19

Effective re-optimization algorithms

slide-20
SLIDE 20

20

Implementation (1/2)

  • ABNO-driven re-optimization
  • Involves both front-end PCE and Back-end PCE

[Ref] L. Velasco, F. Paolucci, Ll. Gifre, A. Aguado, F. Cugini, P. Castoldi, V. Lopez,, “First experimental demonstration of ABNO-driven in-operation flexgrid network re-optimization”, OFC Conf, March. 2014, post-deadline paper

Telefonica CNIT UPC

slide-21
SLIDE 21

21

Implementation (2/2)

Back-end PCE TED LSP-DB Active Solver PCEP Server PCEP PCEP Server Active Stateful PCE Provisioning TED LSP-DB Telefónica Premises (Madrid, Spain) UPC Premises (Barcelona, Spain) ABNO Controller PCEP CNIT Premises (Pisa, Italy) GCO Controller Controller PCC Conn. Controller Res. Mngr. Controller RSVP-TE 172.16.104.2 172.16.101.3 172.16.50.2 10.0.0.49 PCEP 10.0.0.8 10.0.0.1

n = 1 m = 1

2

[Ref]. L. Gifre, F. Paolucci, L. Velasco, A. Aguado, F. Cugini, P. Castoldi, V. Lopez, “First Experimental Assessment of ABNO-driven In-Operation Flexgrid Network Re-Optimization” Journal of Lightwave Technology (JLT), 2014.

slide-22
SLIDE 22

22

Multipath Restoration and Bitrate Squeezing

slide-23
SLIDE 23

23

SDN-based Implementation

[Ref]. F. Paolucci, A. Castro, F. Cugini, L. Velasco, P. Castoldi “Multipath restoration and bitrate squeezing in SDN-based elastic optical networks [Invited]” Journal of Photonic Network Communications (PNET), Aug. 2014

slide-24
SLIDE 24

24

Multi-domain networks

DataCenter 1 Optical Transport Network DataCenter 2 Multi-domain planning tool ABNO In-operation Planning Tool SDN Controller SDN Controller Optical Transport Network SDN Controller In-operation Planning Tool Broker Abstract Links Inter-domain Links

slide-25
SLIDE 25

25

Example

Domain 1 Domain 2 (ABNO) src tgt Domain 4 Candidate Path 1, 6 , 6 1, 5 , 5, 6 , 6 3, 6 , 6 free s e slices es: 1, 1, 3, 3, 4, 6 , 6 1, 1, 2,6 ,6 1, 1, 5, 6 , 6 1, 1, 3, 6 , 6

slide-26
SLIDE 26

26

Experimental Set-up

147.83.30.189 UC Davis Premises (Davis, California) OF-Controller D2 Broker 169.237.74.168 USTC Premises (Hefei, China) 222.195.92.10 OpenFlow Controller D1 169.237.74.223 OpenFlow UPC Premises (Barcelona, Spain) Controller HTTP REST Server Topology DB Algorithm PLATON OpenFlow OpenFlow HTTP/XML

  • 1
  • 6
  • 8
  • 9
  • 3
  • 1
  • 2
  • 4
  • 8
  • 9
  • 6
  • 7
  • 5
  • 10

[Ref] Ll. Gifre et al, "Experimental Assessment of Broker and Planning Tool Coordination in Multi-domain Environments,“ submitted to OFC, 2015.

slide-27
SLIDE 27

27

[1]. L. Velasco, F. Paolucci, Ll. Gifre, A. Aguado, F. Cugini, P. Castoldi, V. Lopez, “First experimental demonstration of ABNO-driven in-operation flexgrid network re-optimization” OFC Conf., March 2014, Post-deadline Paper [2]. M. Dallaglio, A. Giorgetti, N. Sambo, F. Cugini, P. Castoldi, “Impact of slice-ability on dynamic restoration in GMPLS-based Flexible Optical Networks” OFC Conf., March 2014, Top Scored Paper [3]. Ll. Gifre, A. Castro, M. Ruiz, N. Navarro, L. Velasco, “An in-operation planning tool architecture for flexgrid network re-optimization” ICTON Conf., July 2014 [4]. M. Dallaglio, A. Giorgetti, N. Sambo, P. Castoldi, “Impact of SBVTs based on Multi-wavelength Source During Provisioning and Restoration in Elastic Optical Networks” ECOC Conf., Sept. 2014 [5]. F. Paolucci, A. Castro, F. Cugini, L. Velasco, P. Castoldi “Multipath restoration and bitrate squeezing in SDN-based elastic optical networks [Invited]” Journal of Photonic Network Communications (PNET), August 2014 [6]. L. Gifre, F. Paolucci, L. Velasco, A. Aguado, F. Cugini, P. Castoldi, V. Lopez, “First Experimental Assessment of ABNO-driven In-Operation Flexgrid Network Re-Optimization” Journal of Lightwave Technology (JLT), 2014.

  • One additional JOCN journal under minor revisions, two papers submitted at OFC 2015

REACTION Publications

slide-28
SLIDE 28

28

  • Filippo Cugini

<filippo.cugini@cnit.it>

  • Luis Velasco

<lvelasco@ac.upc.edu>

  • Juan Pedro Fernandez-Palacios Gimenez

<juanpedro.fernandez-palaciosgimenez@telefonica.com> Victor Lopez <victor.lopezalvarez@telefonica.com>

Contacts