在這篇論文（使用前饋光交換機和光纖延遲線建構可變式封包輸出緩衝多工器和輸出緩衝交換機）當中，我們探討兩種光佇列仿效的問題：(1)N對1可變式封包輸出緩衝多工器(N-to-1 output-buffered multiplexers with variable length bursts)，以及(2) N對N可變式封包輸出緩衝交換機(N×N output-buffered switches with variable length bursts)。對於這兩種光佇列而言，封包延遲的長度是在封包來時就已知的。因此，利用找到一個延遲路徑產生每個封包的正確延遲即可仿效出上述的佇列。
　　為了虛擬出封包的延遲，我們考慮使用光交換機和光纖延遲線的前饋多重網路架構。對於任意固定的延遲d，在這網路架構中存在一個多重的延遲路徑。如果滿足下列三個限制，我們稱這個延遲路徑是可行的：(1)衝突的限制：在同個時間，不能有超過一個封包在同一個的輸入輸出端同時被安排。(2)因果的限制：沒有封包能夠在其沒到達輸入端時被安排。(3)連續的限制：在同一個大封包裡的封包應該要連續地被遞送到任何一條的光纖延遲線。
　　藉由最差狀況的分析，我們發現了利用前饋網路架構在每個層級所需要的光纖延遲線數目的充分條件來精確仿效出這兩種光佇列。對於N對1的輸出緩衝多工器來說，當每個大封包的長度為一個封包時，充分條件亦為必要條件。
　　利用電腦的模擬，我們也展示出因為統計多樣性增益的關係，在每個層級的光纖延遲線數目能夠再進一步的大幅減少。

In this thesis, we study the problem of exact emulation of two types of optical queues:
(i) N-to-1 output-buffered multiplexers with variable length bursts, and (ii) N × N outputbuffered
switches with variable length bursts. For both queues, the delay of a packet (in
a burst) is known upon its arrival. As such, one can emulate such queues by finding a
delay path that yields the exact delay for each packet. For emulating the delay of a packet
in such queues, in this thesis we consider a multistage feedforward network with optical
crossbar Switches and fiber Delay Lines (SDL). For any fixed delay d, there exist multiple
delay paths in such a network. A delay path is feasible if it satisfies the following three
constraints: (i) conflict constraint: no more than one packet can be scheduled at the same
input/output ports of each crossbar switch at the same time, (ii) causality constraint: no
packet can be scheduled before its arrival, and (iii) strong contiguity constraint: packets in
the same burst should be routed through any fiber delay lines contiguously. By the worst
case analysis, we find sufficient conditions for the numbers of delay lines needed in each
stage of such a feedforward network to achieve exact emulation of both queues. For N-to-1
output-buffered multiplexers, our sufficient conditions are also necessary when each burst
contains exactly one packet. By computer simulation, we also show that the number of
delay lines in each stage can be greatly reduced due to statistical multiplexing gain.

[1] R. L. Cruz and J. T. Tsai, “COD: alternative architectures for high speed packet switching,”
IEEE/ACM Transactions on Networking, vol. 4, pp. 11–20, February 1996.
[2] D. K. Hunter, D. Cotter, R. B. Ahmad, D. Cornwell, T. H. Gilfedder, P. J. Legg and
I. Andonovic, “2 × 2 buffered switch fabrics for traffic routing, merging and shaping in
photonic cell networks,” IEEE Journal of Lightwave Technology, vol. 15, pp. 86–101,
1997.
[3] C.-S. Chang, D.-S. Lee, and C.-K. Tu, “Recursive construction of FIFO optical multiplexers
with switched delay lines,” IEEE Transactions on Information Theory, vol. 50,
pp. 3221-3233, 2004.
[4] C.-S. Chang, D.-S. Lee and C.-K. Tu, ”Using switched delay lines for exact emulation
of FIFO multiplexers with variable length bursts,” IEEE Journal on Selected Areas in
Communications, Vol. 24, No. 4, pp. 108-117, 2006.
[5] C.-C. Chou, C.-S. Chang, D.-S. Lee, and J. Cheng, ”A necessary and sufficient condition
for the construction of 2-to-1 optical FIFO multiplexers by a single crossbar switch and
fiber delay lines,” IEEE Transactions on Information Theory, Vol. 52, pp. 4519-4531,
2006.
[6] C.-S. Chang, Y.-T. Chen, and D.-S. Lee, “Construction of optical FIFO queues,” IEEE
Transactions on Information Theory, Vol. 52, No. 6, pp.2838-2843, 2006.
[7] C.-S. Chang, Y.-T. Chen, J. Cheng, and D.-S. Lee, “Multistage constructions of linear
compressors, non-overtaking delay lines, and flexible delay lines,” Proceedings of IEEE
INFOCOM 2006
[8] A. D. Sarwate and V. Anantharam, “Exact emulation of a priority queue with a switch
and delay lines,” Queueing Systems: Theory and Applications, Vol. 53, pp. 115–125,
July 2006.
[9] H.-C. Chiu, C.-S. Chang, J. Cheng, and D.-S. Lee, “A simple proof for the constructions
of optical priority queues,” submitted to Queueing Systems: Theory and Applications,
2005.
[10] H.-C. Chiu, C.-S. Chang, J. Cheng, and D.-S. Lee, “Using a single switch with O(M)
inputs/outputs for the construction of an optical priority queue with O(M3) buffer,”
accpeted by IEEE INFOCOM minisymposium 2007.
[11] M. J. Karol, “Shared-memory optical packet (ATM) switch,” SPIE vol. 2024: Multigigabit
Fiber Communication Systems(1993), pp. 212–222, October 1993.
[12] M. Yoo, C. Qiao, and S. Dixit, “QoS performance of optical burst switching in IPover-
WDM networks,” IEEE Journal on Selected Areas in Communications, vol. 18,
pp. 2062-2071, October 2000.
[13] E. A. Varvarigos and V. Sharma, “An efficient reservation connection control protocol
for gigabit networks, ” Computer Networks and ISDN Systems ,vol. 30, (no. 12), 13 July
1998, pp. 1135-1156.
[14] C. Fraleigh, S. Moon, B. Lyles, C. Cotton, M. Khan, D. Moll, R. Rockell, T. Seely,
and C. Diot, “Packet-level traffic measurements from the Sprint IP backbone,” IEEE
Network, Vol. 17, pp. 6-16, November-December 2003.
[15] C.-S. Chang, D.-S. Lee and Y.-S. Jou, “Load balanced Birkhoff-von Neumann switches,
part I: one-stage buffering,” Computer Communications, Vol. 25, pp. 611-622, 2002.
[16] F. P. Kelly, “Loss networks,” Ann. Appl. Probab., Vol. 1, pp. 319–378, 1991.
[17] L. Tancevski, S. Yegnanarayanan, G. Castanon, et al. “Optical routing of asynchronous,
variable length packets” Journal on Selected Areas in Communications, Vol. 18, pp. 2084-
2093, 2000.
[18] F. Callegati, “Approximate modeling of optical buffers for variable length packets,”
Photonic Network Communications, Vol. 3, pp. 383-390, 2001.
[19] W. Rogiest, K. Laevens, D. Fiems, and H. Bruneel, “A performance model for an
asynchronous optical buffer,” Performance Evaluation, Vol. 62(1-4), pp. 313-330, 2005.
[20] C. M. Gauger, “Optimized combination of converter pools and FDL buffers for contention
resolution in Optical Burst Switching,” Photonic Network Communications, Vol.
8, No. 2, pp. 139-148, 2004.