Optical: systems and elements – Deflection using a moving element – Using a periodically moving element
Reexamination Certificate
1998-01-07
2001-06-05
Pascal, Leslie (Department: 2633)
Optical: systems and elements
Deflection using a moving element
Using a periodically moving element
C359S199200
Reexamination Certificate
active
06243178
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to an optical switch network or optical network for use in a network comprising a plurality of computers, an optical exchange network, or a like information switching network.
Comprising optical switches for switching optical signals, an optical network is preferred as a large-capacity network which can not be implemented by purely electric networks. Various optical crossbar exchanges or switches are known which comprise optical switches made of lithium niobate matrix switches or semiconductor optical amplifiers operable as gate switches. Various structures are furthermore known for use as such optical networks.
In the manner which will later be described a little more in detail, a paper was contributed by Yoshiharu Maeno and two others, namely, Yoshihiko Suemura one of the present joint inventors, and another, to the 1996 General Meeting of the Institute of Electronics, Information, and Communication Engineers of Japan, as Paper No. SB-9-5, under the title of “The Possibility of Optical Switching Technology for Parallel Processing Systems” as translated by the contributors. Another paper was contributed by Osamu Ishida and three others to the 1996 Telecommunication Society Meeting of the Institute of Electronics, Information, and Communication Engineers of Japan, as Paper No. B-1072, under the title of “Parallel-Optical-Interconnecting Multiwavelength Star Network (POIMS Net) for High Capacity Switching” as translated by the contributors.
A similar paper was contributed by O. Ishida and two others to the Electronics Letters, the 12th September 1996 Issue, Volume 32, No. 19, pages 1804 to 1805, under the title of “Parallel-optical-interconnecting multiwavelength star network (POIMS Net) for high-capacity switching”.
According to the Maeno et al paper, an optical network is for use between a plurality of transmitting nodes and a plurality of receiving nodes in transmitting a transmission electric signal from one of the transmitting nodes as an optical signal selectively to one of the receiving nodes through an optical crossbar exchange. It will be presumed that such an electric signal is a sequence of packets, each having a packet duration of a certain number of bits, such as four bits. Two of the transmitting nodes may concurrently send respective packet sequences to the optical network.
This Maeno et al optical network comprises memory units, such as FIFO (first-in first-out) units, connected to the transmitting nodes, respectively. Optical transmitters are connected to the memory units, respectively, and to the optical crossbar exchange. Optical receivers are connected to the optical crossbar exchange and respectively to the receiving units. The optical crossbar exchange comprises optical splitters connected respectively through input waveguides to the optical transmitters. Each optical splitter splits the optical signal into a plurality of split signals, equal in number to the receiving nodes. It will be surmised that the transmitting nodes are equal to or less in number than a first predetermined integer N which is not less than two and that the receiving nodes are similarly equal to or not less in number than N.
In the optical crossbar exchange, N amplifier groups are connected respectively to the optical transmitters through the input waveguides. Each amplifier group consists of N semiconductor optical amplifiers. As a consequence, N
2
or square-N semiconductor optical amplifiers are connected to the input waveguides. Each semiconductor optical amplifier serves as the gate switch controlled by a destination of the packet sequence of the split signal to make the split signal pass therethrough to one of N optical combiners that corresponds to the destination. Through N output waveguides, the optical combiners are respectively connected to the optical receivers.
According to each Ishida et al paper referred to above, an optical network is similar to the Maeno et al optical network except for primarily the fact that this Ishida et al optical network has an increased propagation capacity by using a wavelength multiplexed signal in which optical signals of M wavelength, such as first to fourth wavelengths &lgr;
0
, &lgr;
1
, &lgr;
2
, and &lgr;
3
, are multiplexed, where M represents a second predetermined integer which is not less than two. Consequently, M optical transmitters are connected to each memory unit as a transmitter group capable of producing up to M optical signals of different wavelengths respectively in response to up to M packet sequences transmitted from one of the transmitting nodes that is connected to the transmitter group through one of the memory units. Such N transmitter groups can therefore concurrently deliver up to N wavelength multiplexed signals, namely, up to NM individual-wavelength signals, to an input side of the optical crossbar exchange. Connected to an output side of the optical crossbar exchange, are N optical or wavelength demultiplexers, each for demultiplexing such a wavelength multiplexed signal into M individual-wavelength signals for delivery respectively to M optical receivers connected as one of N receiver groups for the receiving nodes, respectively.
It should be noted in connection with such conventional optical networks of Maeno et al and Ishida et al that the memory units are indispensable on supplying the packet sequences to the optical transmitters. If at least two different packet sequences should concurrently be directed from the transmitting nodes to a common receiving node of the receiving nodes without the memory units, a conflict or collision would occur between two single-wavelength signals in either the Maeno et al optical signals or in the wavelength multiplexed signals on one of the output waveguides that is assigned in the optical crossbar exchange to the common receiving node. The memory units are therefore indispensable in temporarily storing such two conflicting packet sequences until an arbitration or compromise is settled therebetween to deliver a privileged one of the conflicting packet sequences to the optical crossbar exchange either through one of the N optical transmitters alone or through such a one of the NM optical transmitters and the optical multiplexer connected to one of the N transmitter groups that includes this one of the NM optical transmitters. The other or others of the conflicting packet sequences are held in the memory units until the privileged packet sequence is delivered through the optical crossbar exchange. Such an optical network is herein called an input buffered switch network.
It is known that throughput of the input buffered switch network can not exceed a theoretical restriction of 58.6%. This restriction is described in a book which is originally written by Martin der Pryker and published by the Prentice-Hall, Inc., and is translated into the Japanese language, pages 178 to 189, by Matusima-Hideki as transliterated according to the ISO Standard No. 3602. It should be pointed out in this connection that, if a leading packet is not privileged in one of the packet sequences that is other than the privileged packet sequence, this nonprivileged packet sequence is held in the memory unit until completion of delivery of at least one simultaneously delivered packet of the privileged packet sequence out of the optical crossbar exchange even though a next following packet of the nonprivileged packet sequence does not conflict with the packets concurrently produced at other transmitting nodes. This gives rise to the theoretical restriction.
It should moreover be noted in the conventional optical networks that the optical crossbar exchange comprises an appreciable number of the semiconductor optical amplifiers as optical gate switches, respectively. The optical crossbar exchange has N input ports connected either to the N optical transmitters or to the N optical multiplexers and N output ports connected either to N optical receivers directly or to the N optical demultiplexers. The number of optical gate switches is equal to a square of N.
Maeno Yoshiharu
Suemura Yoshihiko
NEC Corporation
Ostrolenk Faber Gerb & Soffen, LLP
Pascal Leslie
LandOfFree
Optical crossbar network using wavelength division... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Optical crossbar network using wavelength division..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optical crossbar network using wavelength division... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2484240