Title: Wavelength Swapping Using Tunable Laser for Fractional Lambda Switching
1Wavelength Swapping Using Tunable
LaserforFractional Lambda Switching
IEEE 14th LANMAN Workshop
- Viet-Thang Nguyen, Renato Lo Cigno, Yoram Ofek
- University of Trento Italy
- Mario Baldi
- Politecnico di Torino - Italy
2Outline
- Overview
- F?S principles
- Tunable laser
- F?S designs using tunable laser
- Fixed Connection-F?S architecture
- Wavelength Router-F?S architecture
- Broadcast and Select-F?S architecture
- Scheduling feasibility
- Comparisons between designs
- Discussion
3F?S principles
- IP switching
- Header processing
- Buffering and unpredictable jitter and delay
- Time-driven switching of IP packets
- No header processing
- Time label switching using UTC
- constant jitter and no loss
- Related works
- Optical Burst Switching
- Optical Packet Switching
- Both require header processing
- Both may have packet loss due to congestion
4F?S principles TDS in optical layer
- Common Time References
- For freq. and phase synchronization
- UTC (GPS) or the likes
Link channel 10Gbps, 1000 TFs per time-cycle, 80
time-cycle per super cycle, then 1TF capacity -
per time-cycle 10Mbps - per super cycle
125Kbps
5F?S principles TDS in optical layer
- Forwarding schemes based on the coordination of
UTC signal - Immediate Forwarding (IF) scheme
- Non-IF scheme (little buffering)
6Tunable lasers and ? Swapping
- Tunable laser
- Receive on one predefined wavelength
- Transmit light signal on a range of wavelengths
Source M. Kauer et.al. in the Photonics
Technology Letters, Vol. 15, No. 3, Mar. 2003
- Wavelength swapping
- As label switching, e.g. change the color of TFs
7FC-F?S architecture (fabric-less)
- Facts
- Fixed connection network of p-2-p links from TLs
to out-port MUXs - TLs are controlled to tune every TF such that no
conflict is allowed - The color of a TF after being switched defines
the route of that TF
The Pros - Low HW cost - Low control overhead
and Cons - Rigid routing - Low scheduling
flexibility
8WR-F?S architecture
- Facts
- The same design for OBS first introduced in JSAC
Sep 03. - Dif. in-ports use dif. sets of channels to reach
the same out-port. - Connection ratio r C/N is the size
- The Pros
- Simple design (still only TLs are active devices)
- No internal conflict (nature of static WRs)
- and Cons
- Scheduling flexibility is limited by the factor r
9BS-F?S architecture (Clos equivalency)
- Facts
- 1 TL 1 BSS per input channel.
- BSS
- 1-to-N star coupler
- N simple ON/OFF elements
- Strictly non blocking space design (wavelength
time) - If NC, the HW complexity (i.e. of switching
elements) is Clos equivalency
- and Cons
- OO elements are controlled and coordinated using
UTC signal ? higher OH
10Scheduling Unavailability (Blocking?)
Exact Results (using combinatorial analysis)
UTC
BS non blocking design
Outlet channel
Inlet channel
1
k
2
1
k
2
Inlet
Outlet
1
2
3
4
5
6
7
8
9
10
11
12
11Scheduling feasibility
- For comparison, we define
Scheduling feasibility definition a function of
forwarding method (IF or NIF), k, h, C and N to
measure the max. availability of schedules along
the route path for 1-TF fractional lambda pipe,
given zero load
- Scheduling feasibility
- Computation based on hop-based combinatorial
analysis - Relation with blocking performance?!
- higher the scheduling feasibility is, the less
probable that a TF is blocked at scheduling time.
12Comparisons
13Discussion
- Three novel switch designs that are based on the
use of tunable lasers, in which two are
interesting - FC design is fabric-less
- BS design is strictly non-blocking with Clos
equivalency if of channels/port of ports
(CN) - Analytical result 'scheduling feasibility' that
measures the total number of possible different
schedules for each switch design, given zero
load. - But not blocking performance!
- Scheduling unavailability measures blocking
probability while scheduling TFs - Exp. growth of scheduling feasibility implies
that optimum scheduling to fulfill the long-route
F?P setup is not trivial ?heuristic algorithms.
14- More info
- IP-FLOW project (EU funds)
- http//dit.unitn.it/ip-flow/index.html
- Yoram Ofek ofek_at_dit.unitn.it (project
coordinator) - or nguyen_at_dit.unitn.it (for this work)
- Thank you!
- QA
15References
- 1 Rajiv Ramaswami, Kumar N. Sivarajan, Optical
networks a practical perspective, Morgan
Kaufmann Publishers, 2nd edition, 2001. - 2 C.Qiao, M.Yoo, Optical burst switching (OBS)
a new paradigm for an optical internet, Journal
of High Speed Networks, vol. 8, no. 1, Jan 1999,
pp 6984. - 3 Y. Xiong, M. Vandenhoute, H. C. Cankaya,
Control architecture in optical burst switched
WDM networks, IEEE Journal on Selected Areas of
Communication, vol. 18, no. 10, Oct 2000, pp
18381851. - 4 K. Dolzer, C. Gauger, J.Spaeth, S. Bodamer,
Evaluation of reservation mechanisms for optical
burst switching, International Journal of
Electronics and Communications, vol. 55, no. 1,
2001, pp 1826. - 5 Shun Yao, S.J.B. Yoo, B. Mukherjee, S. Dixit,
All-optical packet switching for metropolitan
area networks opportunities and challenges,
IEEE Communications Magazine, vol. 39, issue 3,
Mar 2001, pp142148. - 6 M. Baldi, Y. Ofek, Fractional lambda
switching, Proc. of ICC 2002, - New York, vol.5, pp 2692 2696. 7 M. Baldi, Y.
Ofek, "Realizing dynamic optical networking,"
Optical Networks Magazine, vol. 4, no. 5, Sep/Oct
2003, pp 100-111. - 8 A. Pattavina, M. Bonomi, Y. Ofek,
Performance evaluation of time driven switching
for flexible bandwidth provisioning in WDM
networks, Proc. of Globecom 2004, Dallas, Texas,
vol. 3, pp 1930-1935. - 9 M. Kauer, M. Girault, J. Leuthold, J.
Honthaas, O. Pellegri, C. Goullancourt, M.
Zirngibl, 16-channel digitally tunable
external-cavity laser with nanosecond switching
time., IEEE Photonics Technology Letters, vol.
15, no. 3, Mar. 2003, pp 371-373. - 10 Jeyashankher Ramamirtham, Jonathan Turner,
Joel Friedman, Design of wavelength converting
switches for optical burst switching, IEEE
Journal On Selected Areas In Communications, vol.
21, no. 7, Sep 2003, pp 1122-1132.