`Patent Trial and Appeal Board
`U.S. Patent & Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`
`JDS UNIPHASE CORPORATION
`Petitioner
`
`v.
`
`CAPELLA PHOTONICS, INC.
`Patent Owner
`
`____________________
`
`Case IPR2015-00731
`Patent RE42,368
`____________________
`
`
`PATENT OWNER RESPONSE
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`TABLE OF CONTENTS
`
`INTRODUCTION ........................................................................................... 1
`
`BACKGROUND ............................................................................................. 8
`
`I.
`
`II.
`
`A. Optical circulators limited the scalability of optical switches .............. 8
`
`B.
`
`The ’368 patent captured a novel switch scalable to a large number of
`ports and to a large number of spectral channels ................................10
`
`C.
`
`Claims ..................................................................................................13
`
`III. CLAIMS 1-6, 9-13, AND 15-22 ARE NOT OBVIOUS OVER THE
`COMBINATION OF BOUEVITCH, SPARKS, AND LIN .........................14
`
`A.
`
`Petitioner’s obviousness challenge is based on impermissible
`hindsight ..............................................................................................14
`
`1.
`
`2.
`
`3.
`
`Petitioner’s obviousness challenge relies on various elements
`pulled from distinctly different embodiments of Bouevitch ....15
`
`A POSA would not have used a more complex two-axis mirror
`to achieve the same function as a one-axis mirror ....................18
`
`Replacing the single-axis mirror in Bouevitch with the two-axis
`mirror of Sparks is not a simple substitution ............................20
`
`Bouevitch teaches away from misalignment for power control as
`described in Sparks ..............................................................................23
`
`Bouevitch and Sparks are incompatible technologies .........................27
`
`Bouevitch and Dueck are incompatible technologies .........................29
`
`Bouevitch does not teach or suggest “input port,” “output port,” and
`“one or more [other/drop/add] ports” as recited in independent claims
`1, 15 and 16. ........................................................................................31
`
`1.
`
`Petitioner incorrectly maps the ’368 patent’s claimed “ports” to
`Bouevitch’s “circulator ports” ..................................................32
`
`B.
`
`C.
`
`D.
`
`E.
`
`
`
`- i -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`Proper meaning of the term “port,” as recited in the ’368 patent
`claims ........................................................................................33
`
`The ’368 patent disavows circulator ports ................................34
`
`The claimed preamble requires all elements to be in an “optical
`add drop apparatus,” so a POSA would have not construed the
`claimed ports to include circulator ports ...................................39
`
`The ’368 patent’s provisional application is consistent with the
`construction that the claimed ports are not circulator ports ......39
`
`The meaning of the term “port,” as recited in the claims, was
`understood by a POSA ..............................................................42
`
`Bouevitch at most has two ports as recited in the ’368 patent .43
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`F.
`
`G.
`
`The configuration depicted in Bouevitch’s Figure 11 does not reflect
`light beams into the circulator ports ....................................................44
`
`The applied references do not teach or suggest beam deflecting
`elements that are continuously controllable in two dimensions as
`recited in independent claims 1, 15, 16 and 17 ...................................45
`
`1.
`
`2.
`
`3.
`
`4.
`
`Petitioner misconstrues “continuously controllable” ................46
`
`Petitioner concedes that Bouevitch does not teach or suggest
`beam-deflecting elements that are continuously controllable in
`two dimensions .........................................................................47
`
`Sparks does not meet the claimed beam-deflecting elements that
`are continuously controllable in two dimensions .....................48
`
`Lin’s one-axis mirror does not meet the claimed beam-
`deflecting elements that are continuously controllable in two
`dimensions ................................................................................49
`
`H.
`
`Petitioner fails to provide KSR rationale for combining Sparks and Lin
` .............................................................................................................52
`
`
`
`- ii -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`Petitioner does not explain how or why a POSA would have
`incorporated Sparks’s alleged servo control into Bouevitch ..............53
`
`I.
`
`IV. THE INDUSTRY ADOPTED CAPELLA’S OPTICAL CONFIGURATION
` .......................................................................................................................54
`
`A.
`
`B.
`
`The industry recognized the advantages presented in Capella’s optical
`configuration .......................................................................................55
`
`Experts, including Petitioner’s expert, adopted Capella’s ROADM
`configuration .......................................................................................56
`
`
`
`- iii -
`
`
`
`
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`EXHIBIT LIST
`
`
`Exhibit
`2001
`
`2002
`
`Reference
`Defendant’s Motion to Transfer Venue, Capella Photonics, Inc. v.
`Cisco Systems, Inc., Case Number: 1:14-cv-20529-PAS, Docket No.
`19, April 4, 2014.
`Capella Photonics Launches Dynamically Reconfigurable Wavelength
`Routing Subsystems, Offering Unprecedented Operating Cost Savings
`and Flexibility for Telecom Service Providers, Business Wire (June 2,
`2003, 8:16 AM),
`http://www.businesswire.com/news/home/20030602005554/en/Capell
`a-Photonics-Launches-Dynamically-Reconfigurable-Wavelength-
`Routing.
`2003 WavePath 4500 Product Brief, Capella,
`http://www.capellainc.com/downloads/WavePath%204500%20Produc
`t%20Brief%20030206B.pdf.
`U.S. Provisional Patent Application No. 60/183,155
`Deposition Transcript of Dan M. Marom, Ph.D.
`Benjamin B. Dingel & Achyut Dutta, Photonic Add-Drop Multiplexing
`Perspective for Next Generation Optical Networks, 4532 SPIE 394
`(2001).
`Tze-Wei Yeow, K. L. Eddie Law, & Andrew Goldenberg, MEMS
`Optical Switches, 39 IEEE Comm. I Mag. no. 11, 158 (2001).
`Clifford Holliday, Components for R-OADMs ’05 (B & C Consulting
`Services & IGI Consulting Inc. 2005) (Excerpts).
`Patrick B. Chu et al., MEMS: the Path to Large Optical
`Crossconnects, 40 IEEE Comm. I Mag. no. 3, 80 (2002).
`Clifford Holliday, Switching the Lightwave: OXC’s – The Centerpiece
`of All Optical Network (IGI Consulting Inc. & B & C Consulting
`Services 2001) (Excerpts).
`An Vu Tran et al., Reconfigurable Multichannel Optical Add-Drop
`Multiplexers Incorporating Eight-Port Optical Circulators and Fiber
`Bragg Gratings, 13 Photonics Tech. Letters, IEEE, no. 10, 1100
`(2001).
`Jungho Kim & Byoungho Lee, Bidirectional Wavelength Add-Drop
`Multiplexer Using Multiport Optical Circulators and Fiber Bragg
`Gratings, 12 IEEE Photonics Tech. Letters no. 5, 561 (2000).
`
`2004
`2005
`2006
`
`2007
`
`2008
`
`2009
`
`2010
`
`2011
`
`2012
`
`
`
`- iv -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`Exhibit
`2013
`2014
`2015
`
`2017
`2018
`2019
`
`2020
`
`Reference
`U.S. Patent No. 6,984,917 to Greywall & Marom.
`U.S. Patent No. 6,657,770 to Marom et al.
`Affidavit of Nicholas J. Nowak in Support of Pro Hac Vice Admission
`Under 37 C.F.R. § 42.10(c)
`2016 Max Born & Emil Wolf, Principles of Optics (Cambridge Univ. Press,
`6th Corrected Ed. 1986) (Excerpts).
`U.S. Patent No. 6,543,286 to Garverick et al.
`U.S. Patent No. 5,629,790 to Neukermans et al.
`Fraine, D.S. Simon, O. Minaeva, R. Egorov, and A.V. Sergienko,
`Precise evaluation of polarization mode dispersion by separation of
`even- and odd-order effects in quantum interferometry, Optics Express
`v. 19, no. 21, 22820 (2011).
`Abdul Al-Azzawi, Fiber Optics: Principles and Practices (CRC Press
`2006).
`2021 Metallic Coatings, Eksma Optics, available at
`http://eksmaoptics.com/optical-components/coatings/metallic-
`coatings/.
`Declaration of Dr. Alexander V. Sergienko
`Curriculum Vitae of Dr. Alexander V. Sergienko
`Network Strategy Partners, LLC, The Business Case for ROADM
`Technology (2006), available at
`https://web.archive.org/web/20130605173554/http://www.cisco.com/e
`n/US/prod/collateral/optical/ps5724/ps2006/prod_white_paper0900aec
`d8052b792.pdf.
`Herzel Laor et al., Construction and Performance of a 576x576 Single-
`Stage OXC, in 2 IEEE Lasers and Electro-Optics Society 1999 12th
`Annual Meeting, at 481 (1999).
`2026
`U.S. Patent No. 6,798,941 to Smith et al.
`2027 Ming C. Wu, Olav Solgaard, & Joseph E. Ford, Optical MEMS for
`Lightwave Communication, 24 J. Lightwave Tech. 4433 (2006).
`U.S. Patent No. 6,178,284 to Bergmann & Joseph E. Ford et al.
`U.S. Patent No. 6,178,033 to Joseph E. Ford et al.
`U.S. Patent No. 6,859,573 to Bouevitch et al.
`
`2022
`2023
`2024
`
`2025
`
`2028
`2029
`2030
`
`
`
`- v -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`Exhibit
`2031
`
`2032
`2033
`2034
`
`Reference
`J. E. Ford, Optical MEMS: Legacy of the telecom boom, Solid-State
`Sensor, Actuator and Microsystems Workshop, Hilton Head, SC, Jun.
`6-10 (2004).
`Deposition Transcript of Mr. Sheldon McLaughlin
`U.S. Patent No. 7,676,126 to McLaughlin et al.
`U.S. Patent No. 8,233,794 to Colbourne & McLaughlin et al.
`
`- vi -
`
`
`
`
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`I.
`
`Introduction
`
`Capella’s U.S. Patent No. RE42,368 (“’368 patent”) claims at least two
`
`unique features: (1) multiple fiber collimators corresponding to and providing an
`
`input port for a multi-wavelength optical signal, an output port, and one or more
`
`other ports, and (2) beam-deflecting elements being continuously controllable in
`
`two dimensions. These unique features allow the system to reflect individual
`
`spectral channels from an input fiber collimator to any selected output fiber
`
`collimator. The optical system in the ’368 patent has both of these features, so the
`
`system is reconfigurable, highly scalable, and can eliminate lossy and unwanted
`
`peripheral devices, such as circulators, that were traditionally coupled to optical
`
`switches.
`
`Figures 1A and 1B of the ’368 patent
`
`Input Port and Output Ports 110
`
`
`
`
`
`
`
`Micromirror Array 103
`
`
`
`
`
`
`Optical systems at the time of the ’368 patent were forced
`
`to utilize undesirable optical circulators. A circulator circulates
`
`optical signals from one circulator port or opening to the next.
`
`Referencing the schematic reproduced herein, light can both enter and exit
`
`
`
`- 1 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`circulator ports 1, 2, and 3. Light entering circulator port 1 is emitted from
`
`circulator port 2, light entering circulator port 2 is emitted from circulator port 3,
`
`and light entering circulator port 3 is emitted from circulator port 1. Circulators
`
`were effective to perform the simple task of separating incoming and outgoing
`
`optical signals, but optical systems using circulators were not scalable to a large
`
`number of ports or to a large number of spectral channels because every added
`
`circulator contributed cost, bulk, and insertion loss (i.e., crosstalk between
`
`channels) to the optical system without adding fiber collimators or increasing the
`
`number of spectral channels.
`
`To overcome these limitations, the inventors of the ’368 patent designed an
`
`optical configuration with multiple fiber collimators as ports and an array of two
`
`axis beam-deflecting elements. This configuration differentiated Capella from
`
`competitors because Capella’s system was reconfigurable, scalable to a large
`
`number of ports, and scalable to a large number of spectral channels. (See Ex.
`
`1001, ’368 patent, 5:56-58, FIG. 1A (capable of seamlessly adding a port 110-N to
`
`the array of ports 110); see also Ex. 2002, Business Wire, p. 2 (“The introduction
`
`of dynamic reconfigurability will enable service providers to drastically reduce
`
`operating expenses associated with planning . . . by offering remote and dynamic
`
`reconfigurability.”); Ex. 2008, Holliday R-OADMs, p. 61 (“Capella is the only
`
`company to offer a 10-fiber port solution, i.e., one input, one express output, and 8
`
`
`
`- 2 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`service ports.”); Ex. 2003, WavePath, pp. 1, 4; Ex. 2022, Sergienko Dec., ¶¶ 41-46,
`
`59-76, 207-08) Further, Capella was able to use individual channel micromirrors to
`
`reflect spectral channels to any selected output fiber collimator—something that
`
`cannot be done with circulator-based systems because in circulators, light merely
`
`enters one circulator port and exits the next.
`
`In a failed attempt to piece together Capella’s configuration, Petitioner uses
`
`three references: (1) U.S. Patent No. 6,498,872 to Bouevitch et al. (“Bouevitch”);
`
`(2) U.S. Patent No. 6,625,340 to Sparks et al. (“Sparks”); and (3) U.S. Patent No.
`
`5,661,591 to Lin et al. (“Lin”). Combining these references is problematic for at
`
`least the following reasons.
`
`As a threshold matter, no person having ordinary skill in the art (“POSA”)
`
`would have found it obvious to combine the references because they are
`
`incompatible. Petitioner contends that using Sparks’s two-axis mirror in Bouevitch
`
`would have been a “simple substitution,” “obvious to try,” and “predictable.”
`
`However Petitioner does not explain how a POSA would have combined the
`
`references or how a POSA would have reconciled the technical incompatibilities
`
`between the references. Contrary to Petitioner’s contentions, the combination is
`
`complex, would have required undue experimentation, and is a product of
`
`hindsight bias.
`
`
`
`- 3 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`As the Board can glean from the applied references and the expert
`
`declarations, Petitioner over-simplifies issues and leaps to conclusions on
`
`combinability. Petitioner’s analysis should have reflected the technological
`
`complexities of this case. Instead, Petitioner relies on what amounts to nothing
`
`more than impermissible hindsight to reconstruct Capella’s invention, namely
`
`using the ’368 patent claims and specification as a roadmap to pick and choose
`
`various discreet elements from disparate embodiments and disparate references.
`
`The embodiments and references are not combinable in this manner, so the Board
`
`should uphold the patentability of all instituted claims.
`
`The references are also not combinable because Bouevitch specifically
`
`teaches away from misalignment and angular displacement for power control—the
`
`very method by which Sparks performs power control. Specifically, Bouevitch
`
`teaches that “angular displacement is disadvantageous with respect to coupling the
`
`add/drop and/or input/output beams of light into parallel optical waveguides . . . .”
`
`(Ex. 1003, Bouevitch, 2:1-7 (emphasis added).) Moreover, a core aspect of
`
`Bouevitch’s invention is fixing or compensating for any misalignment to increase
`
`coupling efficiency. (Sergienko Dec., ¶¶ 85, 97, 144-48.) Sparks, in sharp contrast,
`
`requires misalignment to control power at the output port. (Ex. 1004, Sparks, 2:26-
`
`28 (“By controlling the misalignment of the optical beam path through the switch,
`
`the optical signal can be attenuated in a controlled manner.”), 4:48-50 (“The
`
`
`
`- 4 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`present invention utilises [sic] a control system to control the mirrors so as to
`
`deliberately misalign the optical beam path 30 through the switch.”).) This
`
`teaching away by Bouevitch of misalignment and angular displacement for power
`
`control—the very method employed by Sparks—further demonstrates the
`
`incongruity of Petitioner’s obviousness challenge.
`
`Regardless of whether Bouevitch teaches away from misalignment and
`
`angular displacement, Bouevitch and Sparks are incompatible technologies that do
`
`not work together. Bouevitch is designed to modulate the power at its modifying
`
`means—away from and on the opposite end of its system from the input and output
`
`ports. It is in this way that Bouevitch creates an optical configuration that controls
`
`power, while efficiently coupling the output beam to the output port. Indeed,
`
`“efficient coupling” is what Bouevitch calls the spirit and scope of [his alleged]
`
`invention.” (See Bouevitch, 15:19-30; Sergienko Dec., ¶ 148.) Sparks, on the other
`
`hand, is in direct conflict with Bouevitch’s teachings because Sparks performs
`
`attenuation by misaligning the output beam to the output port.
`
`Beside the fact that no POSA would have combined these references, the
`
`Board should also uphold patentability of all instituted claims because the asserted
`
`combination does not disclose each and every claim element, e.g., (1) at least three
`
`fiber collimator ports, and (2) beam-deflecting elements that are continuously
`
`controllable in two dimensions.
`
`
`
`- 5 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`The first reference, Bouevitch, discloses an optical system comprising at
`
`most two collimator ports. Like the prior-art systems described in the ’368 patent,
`
`Bouevitch uses peripheral circulators to add optical signals to an input port and to
`
`drop optical signals from an output port. To compare the ’368 patent to Bouevitch,
`
`the multiple fiber collimators that provide an input port and multiple output ports
`
`in the ’368 patent and the circulators in Bouevitch are reproduced below:
`
`
`
`Figure 1A of the ’368 patent
`
`
`
`
`
`Figure 11 of Bouevitch
`
`
`
`
`Petitioner contends that the circulator ports in Bouevitch read on the claimed
`
`“input port” and “output ports.” But such interpretation is inconsistent with the
`
`’368 patent, the ’368 patent’s earliest provisional application, and the inventors’
`
`underlying motivation to design an optical switch scalable to a large number of
`
`ports. The ’368 patent, its earliest provisional application, and Bouevitch itself
`
`distinguish circulators from the fiber collimators that act as the claimed “ports” and
`
`emphasize that conventional optical systems could not scale to a large number of
`
`ports or to a large number of spectral channels because conventional optical
`
`
`
`- 6 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`systems, like Bouevitch, utilized circulators. The ’368 patent explicitly labels its
`
`ports “collimators” and provides public notice that the present invention is using
`
`multiple fiber collimators to provide the ports. (See, e.g., ’678 patent, 3:54-58,
`
`4:52-63.) So, the ports in the ’368 patent cannot be read to encompass circulator
`
`ports like those used in Bouevitch. Additionally, throughout the ’368 patent, its
`
`provisional application, and in Bouevitch itself, circulators are always shown and
`
`taught as being downstream, spaced from, and coupled to a port. Construing the
`
`claimed ports to read on optical circulator ports is contrary to the ’368 patent and
`
`diminishes the capabilities that the ’368 patent brought to the industry.
`
`The combination of references also does not teach or suggest beam-
`
`deflecting elements that are continuously controllable in two dimensions. For this
`
`claim element, Petitioner uses Sparks and Lin. Sparks and Lin, however, do not
`
`teach beam-deflecting elements that are continuously controllable in two
`
`dimensions. Even more problematic than the shortcomings of Sparks and Lin for
`
`this feature, Petitioner provides no KSR rationale for combining Sparks and Lin or
`
`for combining both references with Bouevitch. Petitioner fails to show in the
`
`Petition that the combination teaches or suggests beam-deflecting elements that are
`
`continuously controllable in two dimensions or that a POSA would have been
`
`motivated to combine the references—deficiencies that cannot be cured. As a
`
`result, the Board should uphold patentability of all instituted claims.
`
`
`
`- 7 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`
`II. Background
`At the time of the effective filing date of the ’368 patent—March 19, 2001—
`
`the number one concern for fiber optic carriers was the ability to provide an optical
`
`switch scalable to a large number of ports and to a large number of spectral
`
`channels. (See Ex. 2007, Yeow, p. 163 (“[O]ptical switches need the capability to
`
`scale in order to manipulate the increased number of wavelengths. MEMS-based
`
`optical switches must incorporate this key feature to gain widespread acceptance of
`
`the carriers.”); Ex. 2009, Chu, p. 81 (“scalability is a paramount concern”);
`
`Sergienko Dec., ¶¶ 47-51, 71.) Providing a scalable switch was an important
`
`concern because the demand for fiber optics communication was increasing, even
`
`as much as 400% per year. (See Ex. 2010, Holliday OXC, p. 12; Sergienko Dec., ¶
`
`47; see also ’368 patent, 1:32-36 (“there is a growing demand for optical
`
`components and subsystems that enable the fiber-optic communications networks
`
`to be increasingly scalable, versatile, robust, and cost-effective”).)
`
`A. Optical circulators limited the scalability of optical switches
`Optical switches were unable to meet increasing demand because optical
`
`switches were typically limited to two ports. (Sergienko Dec., ¶¶ 51-56, 67-70, 76.)
`
`Optical switches required multi-wavelength signals to enter the optical switch on a
`
`single fiber and exit the optical switch on a single fiber (i.e., two ports). (See ’368
`
`patent, 1:59-63; Ex. 1008, ’217 Provisional, p. 2; Sergienko Dec., ¶¶ 48, 51-55, 67-
`
`
`
`- 8 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`68, 76.) An additional means was required to be coupled to the ports to both add
`
`optical signals to an input fiber and to drop optical signals from an output fiber.
`
`(See ’368 patent, 1:59-63; ’217 Provisional, p. 2; Sergienko Dec., ¶¶ 48, 52-55,
`
`69.) Optical circulators and wavelength multiplexers/demultiplexers often served
`
`as the additional means to add and drop optical signals from the two ports. (See
`
`’368 patent, 1:59-2:2; ’217 Provisional, p. 2; Sergienko Dec., ¶¶ 48, 52-55, 69.)
`
`Despite having the capability to add and drop signals, circulators increased the
`
`physical size of a system, contributed to optical loss, and increased costs. (See Ex.
`
`2006, Dingel, p. 401; Sergienko Dec., ¶¶ 51, 55.)
`
`Around the invention date of the ’368 patent, one attempt to meet increasing
`
`demand was to concatenate optical switches (i.e., connecting multiple optical
`
`switches in a chain). (See ’368 patent, 1:59-63; Sergienko Dec., ¶¶ 52-55.)
`
`However, concatenating optical switches also substantially added bulk and cost.
`
`(Sergienko Dec., ¶ 55.) Other inventors were attempting to add circulator ports to
`
`circulators. (See, e.g., Ex. 2011, Tran, p. 1100 (disclosing a circulator with eight
`
`circulator ports); Ex. 2012, Kim, p. 561 (disclosing a circulator with six circulator
`
`ports); Sergienko Dec., ¶ 51.) Systems using complex circulators, however, still
`
`had limited scalability because the circulators only had two output paths from an
`
`optical switch. (Sergienko Dec., ¶ 51.) Ultimately, the industry needed a system
`
`that eliminated optical circulators, while providing an optical switch scalable to
`
`
`
`- 9 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`both a large number of ports and a large number of spectral channels. (See ’368
`
`patent, 1:32-36; Sergienko Dec., ¶¶ 46, 71-76; see also Sergienko Dec., ¶¶ 45-46,
`
`207-08 (describing industry adoption of the configuration for mesh networks).)
`
`B.
`
`The ’368 patent captured a novel switch scalable to a large
`number of ports and to a large number of spectral channels
`
`The inventors of the ’368 patent recognized the limitations of circulator-
`
`based optical switches. (See ’368 patent, 2:49-57, 3:45-46; Sergienko Dec., ¶¶ 51-
`
`55.) To address the problem of limited scalability, the inventors disclosed a
`
`multiple-port optical switch scalable to a large number of spectral channels. (See
`
`’368 patent, 5:51-58 (“[The] underlying architecture is intrinsically scalable to a
`
`large number of channel counts.”); Sergienko Dec., ¶¶ 56, 59-67, 72-75, 169-72.)
`
`Figures 1A/1B (reproduced below) depict the claimed features of the ’368
`
`patent—fiber collimator ports and an array of two axis beam-deflecting elements.1
`
`Figures 1A and 1B of the ’368 patent
`
`
`
`1 Micromirrors are the preferred embodiment for the array of beam-
`
`
`
`deflecting elements.
`
`
`
`- 10 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`The system has, as circled in red, multiple fiber collimators 110-N serving as
`
`the structure for the input and output ports. (See, e.g., ’368 patent, Abstract, FIG.
`
`1A.) The ’368 patent describes a collimator as “typically in the form of a
`
`collimating lens (such as a GRIN lens) and a ferrule-mounted fiber packaged
`
`together in a mechanically rigid stainless steel (or glass) tube.” (Id. at 9:17-20.)
`
`In the system, a multi-wavelength light beam is first sent through the input
`
`fiber collimator port 110. (See id. at 6:64-7:11; Sergienko Dec., ¶ 60.) The input
`
`light beam impinges on the diffraction grating to demultiplex the light beam into
`
`individual wavelengths. (See ’368 patent, 6:64-7:11; Sergienko Dec., ¶ 60.) The
`
`individual wavelengths then diffract off the diffraction grating 101 toward the lens.
`
`(See ’368 patent, 6:64-7:11; Sergienko Dec., ¶ 60.) The lens focuses the individual
`
`wavelengths onto different mirrors among the micromirrors 103. (See ’368 patent,
`
`6:64-7:11; Sergienko Dec., ¶¶ 60, 62.)
`
`A unique feature of the ’368 patent is that each micromirror is individually
`
`and continuously controllable in two dimensions. (See ’368 patent, 8:21-27, 9:8-9;
`
`Sergienko Dec., ¶¶ 63, 66-67.) Since each of the micromirrors 103 is assigned to a
`
`specific spectral channel and is individually and continuously controllable in two
`
`dimensions, the system can dynamically reflect any spectral channel to any output
`
`port. (See ’368 patent, 3:64-4:6, 8:21-27, 9:8-9; Sergienko Dec., ¶¶ 60-63, 66-67.)
`
`
`
`- 11 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`The configuration disclosed in the ’368 patent not only enabled dynamic
`
`switching but also reduced the number of components required to scale the system.
`
`(’368 patent, 5:36-40 (“the underlying OADM architecture thus presented is
`
`intrinsically scalable and [is a system that] can be readily extended.”); Sergienko
`
`Dec., ¶¶ 73-76.)
`
`Neither scaling to a larger number of ports nor scaling to a larger number of
`
`spectral channels required expensive and bulky components, such as circulators,
`
`multiplexers, and demultiplexers. (See Sergienko Dec., ¶¶ 71-76.) For example, the
`
`array of fiber collimators serving as ports 110-N could seamlessly accommodate an
`
`additional input to the system or an additional spectral output from the system—
`
`without the need for circulators. (See ’368 patent, FIGS. 1A, 2A, 2B, 3; Sergienko
`
`Dec., ¶¶ 73-76; see also Business Wire, pp. 2-3 (“Capella’s WavePath product line
`
`[(which uses the technology disclosed in the ’368 patent)] enables system
`
`architects to design optical platforms that offer dynamic and remote
`
`reconfigurability, thus greatly simplifying the engineering and provisioning of
`
`optical networks.”); WavePath, pp. 1, 4 (showing that the WavePath product line is
`
`covered by the ’368 patent).) The system could also scale to additional spectral
`
`channels by adding a micromirror to the micromirrors 103. (Sergienko Dec., ¶ 75.)
`
`The mircomirrors 103 are used not only for switching, but also for power
`
`control. (See ’368 patent, 8:28-36; Sergienko Dec., ¶¶ 77-79.) The system controls
`
`
`
`- 12 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`power by using the micromirrors 103 to alter the coupling efficiency of the spectral
`
`channels into the respective fiber collimator output ports. (See ’368 patent, 8:28-
`
`36; Sergienko Dec., ¶ 77.) The coupling efficiency is defined as the ratio of the
`
`amount of optical power that is coupled into the fiber collimator output port’s fiber
`
`core to the total amount of optical power from the light beam. (’368 patent, 8:31-
`
`36; Sergienko Dec., ¶ 77.)
`
`Embodiments of the ’368 patent also utilize servo control. (’368 patent,
`
`4:47-56, 11:52-57; Sergienko Dec., ¶¶ 78-79.) An embodiment of the servo control
`
`assembly is depicted in Figure 4A of the ’368 patent. (See ’368 patent, FIG. 4A;
`
`Sergienko Dec., ¶ 78-79.)
`
`C. Claims
`The elements of the ’368 patent are incorporated into the claims. The ’368
`
`patent claims recite the following features in illustrative claim 1 (emphasis added):
`
`Claim 1. An optical add-drop apparatus comprising
`an input port for an input multi-wavelength optical signal having
`first spectral channels;
`
`one or more other ports for second spectral channels;
`
`an output port for an output multi-wavelength optical signal;
`
`. . . and
`
`a spatial array of beam-deflecting elements positioned such that each
`element receives a corresponding one of said spectral channels, each
`
`
`
`- 13 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`of said elements being individually and continuously controllable
`in two dimensions to reflect its corresponding spectral channel to a
`selected one of said ports and to control the power of the spectral
`channel reflected to said selected port.
`
`III. Claims 1-6, 9-13, and 15-22 are not obvious over the combination of
`Bouevitch, Sparks, and Lin
`
`A POSA would not have combined Bouevitch, Sparks, and Lin in the
`
`manner presented by Petitioner because the combination could only have been
`
`made through impermissible hindsight, Bouevitch teaches away from Sparks’s
`
`technology, and Bouevitch and Sparks are incompatible technologies. But even if
`
`properly combinable, the combination does not render the claims obvious because
`
`the combination fails to teach or suggest an input port and a plurality of output
`
`ports, as properly construed. The combination also fails to teach or suggest
`
`elements being individually and continuously controllable in two dimensions. For
`
`at least these reasons, the Board should uphold the patentability of the instituted
`
`claims.
`
`A.
`
`Petitioner’s obviousness challenge is based on impermissible
`hindsight
`
`A POSA would not have made Petitioner’s combination of Bouevitch,
`
`Sparks, Lin, and Dueck absent impermissible hindsight. (Sergienko Dec., ¶¶ 121-
`
`40, 217.) Instead, Petitioner started with the ’368 patent and used it as a roadmap
`
`to pick and choose elements from multiple disparate embodiments found primarily
`
`
`
`- 14 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`in Bouevitch and Sparks. Petitioner used these embodiments to construct an optical
`
`system that allegedly maps to the claims of the ’368 patent.
`
`Petitioner repeatedly takes elements from disparate, incompatible
`
`technologies and attempts to piece them together to garner all of the claim
`
`elements. (Id. at 124-30.) Petitioner does so without explaining why or how a
`
`POSA would have been able to reconcile how the disparate technologies could or
`
`would operate together. (Id. at 123.) This is classic hindsight reconstruction of
`
`Capella’s claims. See In re Gorman, 933 F.2d 982, 987 (Fed. Cir. 1991) (“It is
`
`impermissible [to use] applicant’s structure as a template and select[] elements
`
`from references to fill the gaps.”)
`
`1.
`
`Petitioner’s obviousness challenge relies on various
`elements pulled from distinctly different embodiments of
`Bouevitch
`
`As an initial matter, the Board should uphold the patentability of all
`
`instituted claims because Petitioner combines disparate embodiments disclosed in a
`
`single reference without providing KSR rationale. When a petitioner relies on
`
`disparate embodiments, even if those embodiments are disclosed in the same
`
`reference, the petitioner must provide articulated reasoning with rational
`
`underpinning to justify combinability. See Nat’l Envtl. Prod. Ltd v. Dri-Steem
`
`Corp., IPR2014-01503, Decision Denying Institution, pp. 11-12 (Paper 16, Apr. 1,
`
`2015) (citing KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007); Boston
`
`
`
`- 15 -
`
`
`
`Case IPR2015-00731 of
`U.S. Patent No. RE42,368
`Scientific Scimed, Inc. v. Cordis Corp., 554 F.3d 982, 991 (Fed. Cir.