`
`Doug G. Muehlhauser (Reg. No. 42,018)
`William H. Shreve (Reg. No. 35,678)
`Payson LeMeilleur (pro hac vice)
`Mark Lezama (pro hac vice)
`Kendall M. Loebbaka (pro hac vice)
`KNOBBE, MARTENS, OLSON & BEAR, LLP
`2040 Main Street, Fourteenth Floor
`Irvine, CA 92614
`Tel.:
`(949) 760-0404
`Fax:
`(949) 760-9502
`E-mail: BoxNomadix@knobbe.com
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`GUEST TEK INTERACTIVE ENTERTAINMENT LTD.,
`Petitioner,
`
`v.
`
`NOMADIX, INC.,
`Patent Owner.
`
`
`
`
`
`
`
`Case IPR2019-00211
`Patent 7,953,857
`
`
`
`
`
`
`
`
`PATENT OWNER’S SUR-REPLY TO REPLY IN SUPPORT OF
`PETITION FOR INTER PARTES REVIEW
`
`
`
`
`
`
`
`TABLE OF CONTENTS
`
`Page No.
`
`I.
`
`INTRODUCTION .................................................................................... 1
`
`II. CLAIM CONSTRUCTION ..................................................................... 2
`
`III. THE BOARD SHOULD CONFIRM THE PATENTABILITY OF
`THE ’857 PATENT .................................................................................. 7
`
`A.
`
`Petitioner fails to carry its burden for Grounds 1 and 3 ................. 7
`
`1.
`
`2.
`
`3.
`
`Bonomi does not calculate a delay period ........................... 7
`
`A POSITA would not be motivated to combine
`Bonomi and Borella ........................................................... 10
`
`A POSITA would not be motivated to combine
`Bonomi and Teraslinna ...................................................... 14
`
`B.
`
`Petitioner fails to carry its burden for Ground 3 .......................... 16
`
`1.
`
`2.
`
`3.
`
`4.
`
`Petitioner did not meet its burden to show Rupp is
`prior art ............................................................................... 16
`
`Rupp does not disclose a user-selected network
`communication bandwidth or data transmission
`parameter associated with the user within the
`constraints of claims 1 and 9 .............................................. 17
`
`One of ordinary skill in the art would neither be
`motivated to combine Chandran and Rupp, nor have
`a reasonable expectation of succeeding in combining
`the two to arrive at the claimed inventions ........................ 19
`
`Petitioner fails to carry its burden of demonstrating
`Rupp is analogous art ......................................................... 23
`
`
`
`-i-
`
`
`
`TABLE OF AUTHORITIES
`
`Page No(s).
`
`Apple Inc. v. Samsung Elecs. Co.,
`839 F.3d 1034 (Fed. Cir. 2016) ............................................................. 15, 16
`
`Bausch & Lomb, Inc. v. Barnes-Hind/Hydrocurve, Inc.,
`796 F.2d 443 (Fed. Cir. 1986) ..................................................................... 13
`
`In re Clay,
`966 F.2d 656 (Fed. Cir. 1992) ............................................................... 23, 24
`
`In re Gordon,
`733 F.2d 900 (Fed. Cir. 1984) ..................................................................... 14
`
`in Eli Lilly and Co. v. Actavis Elizabeth LLC,
`731 F. Supp. 2d 348 (D.N.J. 2010), aff’d in relevant part,
`435 Fed. App’x 917 (Fed. Cir. 2011) .......................................................... 13
`
`InTouch Techs., Inc. v. VGO Commc’ns, Inc.,
`751 F.3d 1327 (Fed. Cir. 2014) ................................................................... 13
`
`In re Klein,
`647 F.3d 1343 (Fed. Cir. 2011) ................................................................... 24
`
`McGinley v. Franklin Sports, Inc.,
`262 F.3d 1339 (Fed. Cir. 2001) ................................................................... 14
`
`Merck Sharp & Dohme B.V. v. Warner Chilcott Co., LLC,
`711 Fed. App’x 633 (Fed. Cir. Oct. 19, 2017) ............................................ 13
`
`Microsoft Corp. v. Proxyconn, Inc.,
`789 F.3d 1292 (Fed. Cir. 2015), overruled in part on other grounds
`by Aqua Prod., Inc. v. Matal, 872 F.3d 1290 (Fed. Cir. 2017) ..................... 3
`
`In re NTP, Inc.,
`654 F.3d 1279 (Fed. Cir. 2011) ................................................................... 13
`
`-ii-
`
`
`
`TABLE OF AUTHORITIES
`(Cont’d)
`
`Page No(s).
`
`PPC Broadband, Inc. v. Corning Optical Commc’ns RF, LLC,
`815 F.3d 747 (Fed. Cir. 2016) ................................................................... 2, 3
`
`Power Integrations, Inc. v. Fairchild Semiconductor Int’l, Inc.,
`711 F.3d 1348 (Fed. Cir. 2013) ..................................................................... 3
`
`Wang Labs., Inc. v. Toshiba Corp.,
`993 F.2d 858 (Fed. Cir. 1993) ............................................................... 23, 24
`
`In re Wesslau,
`353 F.2d 238 (C.C.P.A. 1965) ..................................................................... 12
`
`
`
`
`
`-iii-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`Exhibit No. Description
`
`EXHIBIT LIST
`
`(Withdrawn) Declaration of Dean Sirovica, Ph.D.
`
`Declaration of Payson LeMeilleur in Support of Motion to Appear Pro
`Hac Vice on Behalf of Patent Owner Nomadix
`
`Declaration of Kendall M. Loebbaka in Support of Motion to Appear
`Pro Hac Vice on Behalf of Patent Owner Nomadix
`
`Declaration of Stuart G. Stubblebine, Ph.D.
`
`Deposition Transcript of Dr. Peter Dordal, dated September 13, 2019
`
`Errata and Signature Pages – Dordal Transcript
`
`Declaration of Mark Lezama in Support of Motion to Appear Pro Hac
`Vice on Behalf of Patent Owner Nomadix
`
`2001
`
`2002
`
`2003
`
`2004
`
`2005
`
`2006
`
`2007
`
`
`
`Exhibit List Page 1
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`I. INTRODUCTION
`
`Petitioner offers a claim construction for “calculating a delay period” divorced
`
`from the specification. Petitioner contends the term includes calculating a specific
`
`timeslot or other interval for delaying packets. Rather than start with the
`
`specification, Petitioner begins its claim construction with extrinsic evidence—a
`
`dictionary definition of “period.” Petitioner then selectively cites the specification,
`
`ignoring aspects of the specification that are counter to its construction. But the ’857
`
`patent specification discloses calculating a delay period as calculating a length of
`
`time. In particular, when discussing traffic shaping, the ’857 patent specification
`
`discloses calculating a delay period and then determining where to place the packet
`
`in a queue.
`
`With respect to Grounds 1 and 3, Bonomi fails to disclose “calculat[ing] a
`
`delay period.” Instead, Bonomi calculates discrete points in time and then
`
`determines which bin in a queue to place a packet. None of these calculations is a
`
`delay period. For at least this reason, Grounds 1 and 3 fail. Grounds 1 and 3 also fail
`
`because a POSITA would not be motivated to combine Bonomi and Borella, or
`
`Teraslinna and Bonomi.
`
`Ground 2 fails because (1) Petitioner failed to meet its burden to show Rupp
`
`is prior art, (2) Rupp does not disclose a network communication bandwidth as
`
`-1-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`recited in the claims, (3) a POSITA would not combine Chandran and Rupp, and
`
`(4) Rupp is not analogous art.
`
`II. CLAIM CONSTRUCTION
`
`A POSITA considering the term “calculating a delay period” in view of the
`
`specification would understand that the term involves calculating a length of time.
`
`(Resp. 10–13.)
`
`Petitioner contends “calculating a period does not necessarily require
`
`determining a specific length of time under the broader standard.” (Reply 1.)
`
`Petitioner argues that “‘period’ … means an ‘interval’ (i.e., ‘set’) of time,” citing a
`
`dictionary definition of “period” in support. (Reply 1 (citing Ex. 1023).)
`
`Petitioner’s position fails for several reasons. First, Petitioner’s position
`
`contradicts itself. An interval is a length of time. So construing “period” as “interval”
`
`does not help Petitioner’s ultimate position, which is that calculating a delay period
`
`does not require calculating a length of time. Whether calculating a delay period
`
`within the meaning of the patent can be performed without calculating a length of
`
`time is the claim-construction dispute, and Patent Owner therefore focuses on
`
`Petitioner’s ultimate position.
`
`Petitioner begins with an extrinsic dictionary definition rather than the
`
`specification or claim language. That approach is reminiscent of the dictionary-
`
`based analysis the Federal Circuit rejected in PPC Broadband, Inc. v. Corning
`
`-2-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`Optical Commc’ns RF, LLC, 815 F.3d 747, 752–56 (Fed. Cir. 2016), because it led
`
`the Board to adopt an unreasonably broad construction divorced from the intrinsic
`
`evidence. The Board must account for how both the claims and the specification
`
`inform a POSITA’s understanding of the disputed term’s meaning in the context of
`
`the patent. Id. at 752. A proper analysis begins with the intrinsic evidence and
`
`involves extrinsic evidence only when the intrinsic evidence is insufficient to resolve
`
`the dispute. See Microsoft Corp. v. Proxyconn, Inc., 789 F.3d 1292, 1297 (Fed. Cir.
`
`2015) (since intrinsic record was clear, extrinsic evidence not considered), overruled
`
`in part on other grounds by Aqua Prod., Inc. v. Matal, 872 F.3d 1290 (Fed. Cir.
`
`2017); see also Power Integrations, Inc. v. Fairchild Semiconductor Int’l, Inc., 711
`
`F.3d 1348, 1360 (Fed. Cir. 2013)(extrinsic evidence appropriate where “intrinsic
`
`record is ambiguous”). Here, the claim language and intrinsic evidence are clear.
`
`(Resp. 10–13.)
`
`In any event, Petitioner’s dictionary definition supports Patent Owner’s
`
`position. Petitioner cites the first definition of “period”: “An interval of time marked
`
`by the occurrence of certain conditions or events <a period of six weeks>.”
`
`(Ex. 1023.) As this definition’s example illustrates, a period (or interval) connotes a
`
`length of time, e.g., “six weeks.”
`
`Petitioner also contends the specification supports its dictionary-based
`
`construction. (Reply 1.) Specifically, Petitioner cites lines 47–54 of column 9, which
`
`-3-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`describes an embodiment involving placing a packet in a timeslot of a ring buffer.
`
`(Id. at 2 (citing Ex. 1001 col. 9 ll. 47-55).) Petitioner then concludes: “So, calculating
`
`a delay period also includes calculating a specific timeslot or other interval for
`
`delaying packets.” (Id.)
`
`But the passage Petitioner cites does not support this conclusion or
`
`Petitioner’s construction. Petitioner cites selectively from column 9, ignoring
`
`passages showing that the ’857 patent distinguishes between the calculation of the
`
`delay period and the placement of a packet in a timeslot. Specifically, the patent
`
`describes calculating the delay period and then placing the packet in a virtual queue:
`
`The subscriber selectable bandwidth module 32 calculates the
`appropriate delay, if any, using the size (in bytes) of the current data
`packet, and the size and time of the previous packet delivered from the
`subscriber. For example, if the user/subscriber has paid for a downlink:
`bandwidth of 100 kilobits per second (kbps), and the gateway device
`12 receives a data packet with the size of 1,500 bytes (12,000 bits), it
`would schedule a delay between packets of 0.12 seconds (12,000 bits
`in a packet/100,000 bits per second bandwidth limit).
`If it is determined that the packet should be delayed, then the
`subscriber selectable bandwidth module 32 places the packet in
`memory in a virtual queue for later delivery.
`(Ex. 1001 col. 8 ll. 39–51 (emphases added).) The column 9 passage Petitioner cites
`
`merely describes a specific embodiment in which “the virtual queue is implemented
`
`by a ring buffer.” (Id. at col. 9 ll. 32–33.) Nothing about this embodiment changes
`
`the operation described in column 8: the bandwidth module calculates a delay period
`
`and separately places the packet in the queue.
`
`-4-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`Petitioner cites a second passage, from column 12, but this passage again
`
`conflicts with Petitioner’s position: “If the packet needs to be queued, then it is
`
`determined at block 350 how long the packet should be delayed, and then the
`
`packet is placed in the appropriate timeslot of the ring buffer.” (Ex. 1001 col. 12 ll.
`
`11–16 (emphases added), cited in Reply 2.) The passage thus distinguishes between
`
`identifying the appropriate timeslot and calculating the delay period. Patent Owner
`
`pointed out this distinction in its Response, yet Petitioner improperly conflates the
`
`two operations.
`
`Petitioner also ignores that the column 12 passage it relies on directly
`
`contradicts its ultimate position. The passage describes block 350 of figure 4B as
`
`determining “how long the packet should be delayed,” and block 350 plainly
`
`includes “DETERMINING LENGTH OF DELAY.” (Ex. 1001 fig. 4B.) The
`
`specification thus reinforces what the plain language of the claims requires:
`
`calculating a delay period involves calculating a length of time.
`
`Petitioner contends that, before the district court, “Patent Owner disagreed
`
`with the very construction it [now] proposes.” (Reply 1.) That is false. Before the
`
`district court, Petitioner proposed construing “calculate a delay period” as
`
`“mathematically determine a length of time to delay transmission of a received
`
`packet.” (Ex. 1025 at 26.) The parties focused on whether Petitioner’s proposal of
`
`“mathematically determine” was more helpful than the claim language “calculate.”
`
`-5-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`(Id.) Agreeing with Nomadix that it was not, the district court declined to construe
`
`“calculate.” (Id. at 26–27.) Notably, Petitioner itself proposed construing “period”
`
`as “length of time,” which Patent Owner did not dispute. (See id.) Although
`
`Petitioner filed the present petition on the last day for which the BRI standard
`
`applies, Petitioner fails to identify any reason that the “length of time” construction
`
`it proposed would be appropriate under the claim-construction principles applicable
`
`in the district court but not under the BRI standard.
`
`Petitioner further contends that the construction Patent Owner now offers is
`
`“narrower than in district court.” (Reply 1.) This too is false. Before the district court,
`
`Patent Owner argued that no construction was necessary; but to the extent
`
`construction was necessary, Patent Owner proposed “calculate a period of time
`
`based on which a received packet’s transmission will be delayed.” Patent Owner
`
`thus used the claim language “period” in its construction. But that does not make
`
`“length of time” narrower. Rather, it reflects that Patent Owner simply did not
`
`address whether “period” referred to a length of time. Petitioner is the party
`
`presenting conflicting positions: it proposed a “length of time” construction at the
`
`district court, and the difference in claim-construction standards cannot explain
`
`Petitioner’s reversal in positions.
`
`Petitioner also contends Patent Owner suggested in its Response that “the
`
`delay period must be the actual amount of time an associated packet is delayed.”
`
`-6-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`(Reply 2 (citing Resp. 20).) Patent Owner never suggested as much. Petitioner is
`
`correct that the claim language requires merely delaying packet transmission “based
`
`on” the calculated delay period and thus does not require the actual delay to be
`
`exactly the same as the calculated delay period. That is exactly what Patent Owner
`
`explained in the district-court litigation. (Ex. 1025 at 27.) But that observation is
`
`irrelevant to the present claim-construction issue: the actual transmission delay need
`
`only be “based on” the calculated delay period, but the calculated delay period is
`
`still a length of time. Conflating the act of delaying with the calculation of the delay
`
`period would read the “calculate a delay period” limitation out of the claims. (See
`
`id. at 28 (noting claims require both “(1) calculating a delay period and (2) delaying
`
`transmission based on the delay period”).)
`
`In summary, calculating a delay period requires calculating a length of time.
`
`III. THE BOARD SHOULD CONFIRM THE
`PATENTABILITY OF THE ’857 PATENT
`
`A.
`
`Petitioner fails to carry its burden for Grounds 1 and 3
`1.
`
`Bonomi does not calculate a delay period
`
`Bonomi fails to disclose “calculating a delay period” as required by the
`
`claims. (See Resp. 15–21.) Although Petitioner failed to specify in its Petition which
`
`aspect of Bonomi discloses this limitation, it now claims that two different parts of
`
`Bonomi disclose “calculating a delay period.” Petitioner cites the following passages
`
`as allegedly disclosing this limitation:
`
`-7-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`Whether the estimated arrival time X complies with the traffic contract
`is determined at step 33 where X is compared to t+1/ρ. If X is less than
`or equal to t+σ/ρ then the connection is complying with the traffic
`contract and the cell is conforming as shown in Step 34. In the case of
`a conforming cell, conformance time c equals the current time t. As
`shown in Step 35, if X is greater than t+σ/ρ the cell is non-conforming
`and the conformance time is set to comply with the contracted traffic
`parameters, c=X-σ/ρ.
`(Bonomi col. 8 ll. 30-38.)
`
`If the cell is non-conforming, the cell is enqueued on a sorting bin as
`shown in step 44. In an embodiment having b sorting bins of grain g in
`the sorting unit, the cell is enqueued onto sorting bin [(c mod bg)/g].
`(Id. at col. 8 ll. 49-52.)
`
`13. The method of shaping traffic in claim 1 wherein each said sorting
`bin corresponds to a single delay period between said conformance time
`and said current time.
`(Id. at claim 13.)
`
`Whether the steps in Bonomi are taken alone or in combination, Bonomi fails
`
`to disclose “calculat[ing] a delay period associated with a received packet.” First,
`
`Petitioner contends that the conformance time set in Bonomi “is a delay period.”
`
`(Reply 5.) However, as Dr. Dordal admitted, a future time, such as the conformance
`
`time, is “not a delay period.” (Ex. 2005 (Dordal Tr.) 94:22–95:6.) A discrete point
`
`in time is not a delay period. Nor is it an interval under Petitioner’s construction.
`
`Second, Petitioner contends “the total potential delay ... also qualifies as a
`
`‘delay period’ under Patent Owner’s proposed construction.” (Reply 6 (citing Dordal
`
`Reply Decl. ¶ 21).) But Petitioner relies on an example invented by its expert that
`
`-8-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`never appears in Bonomi. (Compare Reply 3–4 and Dordal Reply Decl. ¶¶ 13–19
`
`with Bonomi generally.) Bonomi never suggests “grain” refers to time nor provides
`
`an example of grain size (100ms), number of bins (18 bins), or “total potential delay
`
`period” (1800ms). And without any basis in Bonomi, Petitioner illustrates Bonomi’s
`
`bins as a “clock face.” (Reply 4; Dordal Reply Decl. ¶ 13.) Such evidence untethered
`
`to Bonomi is entitled to little weight, if any.
`
`Petitioner also appears to rely on a general delay resulting from placing the
`
`packet in a queue. But a delay in itself is not a calculation. Again, the claims
`
`separately require delaying packet transmission based on the delay period, so
`
`conflating the delay itself with “calculating a delay period” would read out the
`
`calculation limitations. Moreover, none of the calculations in Bonomi leading to
`
`identifying which bin to use is a calculation of a delay period. (See Resp. 16–21.) As
`
`noted by Petitioner, Bonomi “calculate[s] a bin” to place the packet—not a delay
`
`period. (Reply at 6.) Identifying a bin number is not calculating a delay period or
`
`interval. (Stubblebine Decl. ¶ 44.) Moreover, even if a bin somehow “corresponds
`
`to” a delay period, the bins are configured in advance of receiving any packets;
`
`choosing such a bin does not involve calculating a delay period associated with a
`
`received packet as required by the claims. (Stubblebine Decl. ¶ 40.)
`
`Bonomi fails to disclose calculating a delay period as required by the claims.
`
`Therefore, Petitioner has failed to meet its burden on Grounds 1 and 3.
`
`-9-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`2.
`
`A POSITA would not be motivated to combine Bonomi and
`Borella
`
`Bonomi and Borella are directed to two different network types—
`
`Asynchronous Transfer Mode (ATM) and Internet Protocol (IP), respectively. ATM
`
`networks have their own protocols that differ substantially from IP networks.
`
`(Stubblebine Decl. ¶ 53, 62–64.) ATM networks are connection-oriented, meaning
`
`ATM’s units of data—cells—are delivered in the sequence they were sent over the
`
`connection. (Id.; Bonomi col. 1 ll. 36–37.) The connection-oriented nature of ATM
`
`facilitates networked telephony and voice applications, which require low latency
`
`and low jitter (variation in packet delay). (Stubblebine Decl. ¶ 63.) In contrast, IP is
`
`connectionless, meaning data packets can travel different routes to their destination
`
`and may arrive out of sequence. (Id. ¶ 64.) This can cause delay and jitter that
`
`telephony applications do not handle well. (Id.) ATM networks are materially
`
`different from IP networks. Not only are the two types of networks materially
`
`different, Bonomi and Borella rely on features unique to their respective network
`
`types. The entire disclosure of Bonomi relies on a “contract” negotiated for each
`
`connection, where the contract specifies a bandwidth ρ and burstiness σ. (Bonomi
`
`col. 2 ll. 11 –20.) This contract negotiation is a feature specific to ATM networks.
`
`(Bonomi col. 2 ll. 11–17 (explaining ATM’s “connection admission control (CAC)”
`
`process).) The IP protocol does not provide for such contract negotiation.
`
`-10-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`(Stubblebine Decl. ¶ 67.) On
`
`the other hand, Borella’s mechanism for
`
`communicating quality-of-service parameters depends entirely on repurposing a
`
`specific byte in the IP protocol header, which is specific to IP networks and not
`
`present in ATM networks. (Borella col. 9 ll. 11–38; Stubblebine Decl. ¶¶ 61–62.)
`
`Petitioner contends “Bonomi is not limited to ATM networks, and does not
`
`teach away from implementing its system in IP networks.” (Reply 6.) Petitioner
`
`relies upon a single statement in Bonomi that its system is “optimally suited for, but
`
`not limited to, high speed switches in an Asynchronous Transfer Mode Network.”
`
`(Bonomi at col. 6, ll. 36–38.) As Dr. Stubblebine explained, Bonomi “looks like
`
`someone wrote an application related to ATM and then wrote that sentence as well
`
`to say it’s not limited to ATM.” (Ex. 1030 (Stubblebine Tr.) 181:4–7.) A conclusory
`
`assertion in a specification that the invention is “not limited to” switches in ATM
`
`networks does not provide enabling disclosure of use of the invention in an IP
`
`network. Nor does it alter the entire focus of Bonomi’s disclosure, which relies on
`
`features unique to ATM. Moreover, Bonomi notes that “per-connection queuing
`
`enables each connection to be treated separately, allowing the traffic shaper to take
`
`each connection’s unique traffic parameters into account when shaping traffic
`
`received from that connection.” (Bonomi col. 4 ll. 31–35 (emphasis added).) Thus,
`
`Bonomi treated each data type, such as voice, video, and data, with different traffic
`
`parameters, including bandwidth, to deliver the cells in a timely manner and ensure
`
`-11-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`that they are not dropped. (Stubblebine Decl. ¶ 75.) IP networks do not distinguish
`
`between different data types and do not have dedicated fixed connections for each
`
`communication. (Stubblebine Decl. ¶ 64.) Thus, a POSITA would understand
`
`Bonomi is directed to an ATM network and would not be motivated to combine
`
`Borella, which is directed to an IP network. (Stubblebine Decl. ¶¶ 49, 66.)
`
`Petitioner also contends “Borella is neither limited to IP networks, nor teaches
`
`away from ATM networks.” (Reply 6–7.) However, Petitioner cites a passage in
`
`Borella that specifically states it is directed to providing access on an IP network.
`
`(Borella col. 3 ll. 18–20.)
`
`Petitioner contends it “does not propose incorporating all of Borella’s IP
`
`technology into Bonomi.” (Reply 7.) Petitioner contends it is relying on “Borella for
`
`the limited purpose of teaching a user device and a user-selected bandwidth.” (Id.;
`
`see also Dordal Reply Decl. ¶ 32 “my proposed combination … assumes only the
`
`use from Borella of the general abstract idea of a user device and user-selected
`
`bandwidth.”).) However, Petitioner’s argument is based on a false premise that a
`
`POSITA can cherry-pick and generalize an idea mentioned in one reference, lift it
`
`out of context, and import it into the primary reference without explanation. In re
`
`Wesslau, 353 F.2d 238, 241 (C.C.P.A. 1965) (emphasis added) (holding it “is
`
`impermissible within the framework of section 103 to pick and choose from any one
`
`reference only so much of it as will support a given position, to the exclusion of
`
`-12-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`other parts necessary to the full appreciation of what such reference fairly suggests
`
`to one of ordinary skill in the art.”) cited in Eli Lilly and Co. v. Actavis Elizabeth
`
`LLC, 731 F. Supp. 2d 348, 365–66 (D.N.J. 2010), aff’d in relevant part, 435 Fed.
`
`App’x 917, 921 (Fed. Cir. 2011); Bausch & Lomb, Inc. v. Barnes-Hind/Hydrocurve,
`
`Inc., 796 F.2d 443, 448–49 (Fed. Cir. 1986) (citing Wesslau and vacating ruling of
`
`obviousness). This logic would eviscerate the need to provide an articulated
`
`motivation to combine and justification for a reasonable expectation of success.
`
`Moreover, Petitioner’s reasoning improperly relies upon hindsight bias, picking and
`
`choosing generalized pieces of Borella and ignoring the highly specific
`
`characteristics of Borella’s system. See, e.g., In re NTP, Inc., 654 F.3d 1279, 1299
`
`(Fed. Cir. 2011) (“The Board improperly relied on hindsight reasoning to piece
`
`together elements to arrive at the claimed invention.”); InTouch Techs., Inc. v. VGO
`
`Commc’ns, Inc., 751 F.3d 1327, 1351–52 (Fed. Cir. 2014); Merck Sharp & Dohme
`
`B.V. v. Warner Chilcott Co., LLC, 711 Fed. App’x 633, 637 (Fed. Cir. Oct. 19, 2017).
`
`Nor does Petitioner’s assertion withstand scrutiny. First, Bonomi’s objective
`
`is to differentiate between different types of traffic and shape traffic on a connection-
`
`by-connection basis. (Stubblebine Decl. ¶ 50.) Altering Bonomi to manage
`
`bandwidth per user rather than per connection would make Bonomi inoperative for
`
`its intended purpose. Bonomi and Borella “thus cannot serve as predicates for a
`
`-13-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`prima facie case of obviousness.” McGinley v. Franklin Sports, Inc., 262 F.3d 1339,
`
`1354 (Fed. Cir. 2001); In re Gordon, 733 F.2d 900, 902 (Fed. Cir. 1984).
`
`Second, Petitioner has failed to show a POSITA would reasonably expect to
`
`succeed in incorporating Borella’s teaching of user-selected bandwidth into Bonomi.
`
`As explained in detail in Patent Owner’s Response, Borella’s user-selected
`
`bandwidth is rooted in modifying a TOS byte in the IP packet header. (Resp. 27–
`
`30.) Borella’s solution is thus rooted in the specific format of packets that the Internet
`
`Protocol dictates. But the IP protocol is connectionless; does not have a built-in
`
`mechanism for end-to-end, per-connection negotiation of transmission rate or
`
`burstiness; does not guarantee delivery of IP datagrams in the order they were sent;
`
`and allows for packets of variable length. (Stubblebine Decl. ¶ 67.) The IP protocol
`
`therefore does not offer the underlying infrastructure critical to the operation of
`
`Bonomi’s ATM-based per-connection traffic-shaping solution. Thus, a POSITA
`
`would not be motivated to combine Bonomi and Borella.
`
`3.
`
`A POSITA would not be motivated to combine Bonomi and
`Teraslinna
`
`Petitioner contends it relies on “Bonomi for teaching delaying packets based
`
`on a calculated delay period.” (Reply 23.) Petitioner further contends it “does not
`
`rely on, and there is no need to incorporate” a list of features related to Bonomi’s
`
`traffic shaper into Teraslinna. (Id.) But Bonomi’s traffic shaper, including the
`
`-14-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`calculation of a delay period, is premised on “tak[ing] each connection’s unique
`
`traffic parameters into account when shaping traffic received from that connection.”
`
`(Bonomi col. 4 ll. 33 –35.) For example, Bonomi bases the bins and granularity of
`
`its traffic shaper on the negotiated traffic parameters for the ATM connections. (Id.
`
`at col. 7 ll. 38 –47.) Petitioner relies upon these bins as part of the “calculating a
`
`delay period” limitation. Thus, the features in Bonomi that Petitioner suggests
`
`incorporating into Teraslinna are premised on shaping traffic on a per-connection
`
`basis, according to parameters negotiated for specific connections.
`
`Petitioner contends that Patent Owner failed to raise and therefore waived its
`
`argument that “Bonomi teaches away from ‘using its system outside of a per-
`
`connection basis.’” (Reply 24 (citing Stubblebine Decl. ¶ 75).) However, Patent
`
`Owner cited the very paragraph that Petitioner points to for this argument. (Resp. 39
`
`(citing Stubblebine Decl. ¶¶ 72–77).) Patent Owner explained in more detail
`
`Bonomi’s per-connection traffic shaping in its Response. (Resp. 26–27.)
`
`Petitioner also contends that Apple Inc. v. Samsung Elecs. Co., 839 F.3d 1034,
`
`1051, n. 15 (Fed. Cir. 2016) is irrelevant. (Reply 25.) But the traffic shaper in
`
`Bonomi that allegedly discloses calculating a delay is premised on per-connection
`
`queuing. Its features and operation are based on traffic parameters negotiated per-
`
`connection in the ATM network. (See, e.g., Bonomi col. 7 ll. 38–47, col. 8 ll. 18–
`
`54.) In contrast, Teraslinna’s objective is to avoid the inefficiencies of per-
`
`-15-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`connection traffic shaping. (Teraslinna at col. 3 ll. 53–67.) Each reference teaches
`
`away from the critical features of the other reference, which is relevant to and
`
`strongly undercuts any suggestion that a POSITA would be motivated to combine
`
`the references’ teachings. See Apple, 839 F.3d at 1051, n. 15.
`
`Accordingly, a POSITA would not be motivated to combine Bonomi with
`
`Teraslinna. (Resp. 36–41.)
`
`B.
`
`Petitioner fails to carry its burden for Ground 3
`1.
`
`Petitioner did not meet its burden to show Rupp is prior art
`
`In its Preliminary Response, Patent Owner argued that Petitioner failed to
`
`meet its burden to show that Rupp is prior art. (Prelim. Resp. 29–35.) In response to
`
`the Preliminary Response, Petitioner submitted a new declaration for Gerard P.
`
`Grenier, Senior Director of Content Management at IEEE. (Ex. 1022.) Petitioner did
`
`not seek to introduce any other declarations. In its Response, Patent Owner again
`
`argued Patent Owner had failed to show Rupp was prior art. (Resp. 41–48.) In reply,
`
`Petitioner seeks to introduce entirely new evidence from two declarants—Edward
`
`W. Knightly (Ex. 1027) and Dr. Bjorn Rupp, a co-author of the reference (Ex. 1028).
`
`Petitioner does not offer any explanation why it could not obtain these declarations
`
`when it submitted the new declaration for Mr. Grenier. By waiting until this late
`
`stage of the proceeding, Petitioner has prejudiced Patent Owner. For example, Patent
`
`Owner is prohibited from presenting new testimony from other declarants
`
`-16-
`
`
`
`IPR2019-00211
`Patent 7,953,857
`
`contradicting Petitioner’s new declarants. (See PTAB Trial Prac. Guide at 14 (Aug.
`
`2018).)
`
`The late declarations of Mr. Knightly and Dr. Rupp emphasize the
`
`deficiencies in Petitioner’s previously submitted evidence. Petitioner’s properly
`
`submitted evidence fails to show Rupp is a prior-art printed publication. (Resp. 41–
`
`48.)
`
`2.
`
`Rupp does not disclose a user-selected network communication
`bandwidth or data transmission parameter associated with the
`user within the constraints of claims 1 and 9
`
`Petitioner also offers Dr. Rupp’s declaration in response to arguments that
`
`Rupp does not disclose a network communication bandwidth within the constraints
`
`of the claims. Claim 1 requires that the data storage system includes “an indication
`
`of a network communication bandwidth associated with the user device and selected
`
`by the user.” Claim 9 recites a similar limitation, but refers to a “data transmission
`
`parameter.” Petitioner relies upon Rupp for this limitation. (Pet. 44, 47–49.) As
`
`noted by Dr. Stubblebine, and admitted by Dr. Dordal, Rupp discloses a guarantee
`
`of the quality of service a user will experience on the network. (Resp. 50–51;
`
`Stubblebine Decl. ¶ 80–82; Dordal Tr. 36:21–37:1.) Rupp’s