UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`DELL INC.; EMC CORPORATION; HEWLETT-PACKARD ENTERPRISE
`CO.; AND HP ENTERPRISE SERVICES, LLC, AND TERADATA
`OPERATIONS, INC.,
`Petitioners
`
`v.
`
`REALTIME DATA LLC d/b/a IXO,
`Patent Owner.
`
`
`
`Case: IPR2017-00179
`* * * * *
`Case: IPR2017-00808
`
`Patent No. 9,054,728
`[Consolidated]
`
`
`
`PETITIONERS’ REPLY TO PATENT OWNER RESPONSE
`
`Mail Stop PATENT BOARD
`Patent Trial and Appeal Board
`United States Patent and Trademark Office
`PO Box 1450
`Alexandria, Virginia 22313–1450
`Submitted Electronically via the Patent Review Processing System
`
`

`

`TABLE OF CONTENTS
`
`I. PRELIMINARY STATEMENT ..................................................................... 1 
`II. THE TEACHINGS OF THE PRIOR ART ARE UNDISPUTED; THE
`SOLE QUESTION IS MOTIVATION TO ARRIVE AT THE
`CLAIMED SUBJECT MATTER. ................................................................... 3 
`III. THE FULLY-DEVELOPED RECORD SHOWS A POSA WOULD
`HAVE BEEN LED TO THE CLAIMED SUBJECT MATTER
`EXERCISING ONLY ROUTINE SKILL. ..................................................... 5 
`A. 
`Patent Owner’s Theory Rests On The Legally-Unsupported
`Assumption That A POSA Would Have Only Seen One Obvious
`Way To Combine Teachings Of the Prior Art. ...................................... 5 
`
`B. 
`
`Patent Owner Makes An Improper Bodily Incorporation
`Argument. ............................................................................................... 7 
`
`C.  Dr. Creusere Never Testified That Hsu Would Always Be Used
`To Determine Data Type. ....................................................................... 7 
`
`D.  The Fully-Developed Record Continues to Show Good Reasons
`Why a POSA Would Not Use Hsu When Franaszek Lacked A
`Data Type Descriptor. ............................................................................. 8 
`
`1.  Both Experts Agree That Using Franaszek With Hsu When
`Franaszek Has A Data Type Is Advantageous. ............................. 9 
`
`IV. PATENT OWNER’S ARGUMENT THAT FRANASZEK AND HSU
`CANNOT MEET LIMITATION 1[F] IS IRRELEVANT. .......................... 14 
`V. PATENT OWNER’S ARGUMENTS ABOUT HINDSIGHT ARE
`MISGUIDED. ................................................................................................ 14 
`A.  The Board Need Not Look To Sebastian To Find The Claimed
`Subject Matter Obvious. Patent Owner’s Contentions About
`Inconsistencies In Motivations Concerning Sebastian Are Thus
`Irrelevant. .............................................................................................. 14 
`
`B. 
`
`Patent Owner’s Expert Agrees There Was An Advantage To
`Combining Franaszek’s and Hsu’s Teachings. .................................... 16 
`
`
`
`i
`
`

`

`C.  Dr. Creusere’s Consideration Of The Claims Does Not Reveal
`Hindsight. Reviewing The Prior Art In Light Of The Claims Is A
`Necessary Prerequisite To Showing The Claimed Invention Was
`Obvious. ................................................................................................ 18 
`
`D. 
`
`E. 
`
`Patent Owner’s Suggestion That The Obviousness Theory
`Advanced In The Petition Defines The Problem In Terms of Its
`Solution Showing Hindsight Is Incorrect. ............................................ 19 
`
`Even if Sebastian Is Evaluated With Franaszek and Hsu, Dr.
`Zeger Admitted That Considering Franaszek And Hsu Together
`Would Lead a POSA To Consider Speed Issues. ................................. 20 
`
`VI. THE NOTION THAT A POSA WOULD ALWAYS USE
`SEBASTIAN’S DEFAULT ENCODER IF COMBINED WITH
`FRANASZEK IS CONTRARY TO ALL REFERENCES UPON
`WHICH TRIAL WAS INSTITUTED. .......................................................... 21 
`VII. EXPERT TESTIMONY IS UNNECESSARY TO ESTABLISH
`EVERY FACT TO REACH A LEGAL CONCLUSION OF
`OBVIOUSNESS. ........................................................................................... 23 
`VIII. CONCLUSION .............................................................................................. 25 
`
`
`
`
`
`
`
`
`
`
`ii
`
`

`

`TABLE OF AUTHORITIES
`
` Page(s)
`
`Cases
`Application of Kronig, 539 F.2d 1300 (CCPA 1976) .............................................. 15
`
`Belden Inc .v. Berk-Tek LLC,
`805 F.3d 1064 (Fed. Cir. 2015) .......................................................................... 24
`
`CRFD Res., Inc. v. Dish Network Corp.,
`No. 2016-2198, Slip op. (Fed. Cir. Dec. 5, 2017) (Ex. 1032) ............................ 15
`
`Graham v. John Deere,
`383 U.S. 1 (1966) ........................................................................................ 2, 5, 18
`
`In re Keller,
`642 F.2d 413 (CCPA 1981) ............................................................................ 6, 10
`
`KSR Int’l Co. v. Teleflex Inc.,
`550 U.S. 398 (2007) .............................................................................................. 6
`
`Mobotix Corp. v. E-Watch, Inc.,
`IPR2013-00334, Paper 18 (PTAB Jan. 10, 2014) .............................................. 15
`
`In re Mouttet,
`686 F.3d 1322 (Fed. Cir. 2012) ............................................................................ 5
`
`TI Group Automotive Sys. (North America), Inc. v. VDO North
`America, LLC,
`375 F.3d 1126 (Fed. Cir. 2004) .................................................................... 18, 19
`
`
`
`
`
`
`
`iii
`
`

`

`LIST OF ABBREVIATIONS
`
`
`
`Meaning
`
`U.S. Patent No. 9.054,728
`
`Institution Decision
`
`Petition
`
`Patent Owner Response
`
`
`
`
`
`
`
`
`
`
`
`
`
`iv
`
`Abbreviation
`
`’728 Pat.
`
`Inst. Dec.
`
`Pet.
`
`POR
`
`
`
`
`
`
`
`

`

`PETITIONER’S UPDATED EXHIBIT LIST
`
`BRIEF DESCRIPTION
`U.S. Patent No. 9,054,728
`Declaration of Dr. Charles Creusere
`Curriculum Vitae of Dr. Charles Creusere
`U.S. Patent No. 5,870,036 to Franaszek et al. (“Franaszek”)
`W.H. Hsu, et al., Automatic Synthesis of Compression Techniques
`for Heterogeneous Files, Software Practice & Experience, Vol.
`25, No. 10 pp. 1097-1116 (Oct. 1995) (“Hsu”)
`Notice of Intent to Issue Reexamination Certificate dated Nov. 27,
`2013, U.S. Inter Partes Reexamination Control No. 95/001,926
`Office Action dated Dec. 15, 2009 in U.S. Inter Partes
`Reexamination Control No. 95/001,479
`Appeal Brief dated Oct. 1, 2012 in U.S. Inter Partes
`Reexamination Control No. 95/000,479
`MCGRAW-HILL DICTIONARY OF SCIENTIFIC AND TECHNICAL
`TERMS, Fifth Ed. (1993) (excerpts)
`MICROSOFT PRESS COMPUTER DICTIONARY, Third Ed. (1997) 
`(excerpts)
`Reply to Office Action in Inter Partes Reexamination in
`Reexamination Control No. 95/000,479 (Mar. 15, 2010)
`Action Closing Prosecution in Inter Partes Reexamination in
`Reexamination Control No. 95/000,479 (Aug. 27, 2010)
`Jury Verdict Form in Realtime Data, LLC d/b/a IXO v. T-Mobile
`U.S.A., Inc., No. 6:10-cv-493, Dkt. 660 (E.D. Tex. Feb. 11, 2013)
`Jury Instructions in Realtime Data, LLC d/b/a IXO v. T-Mobile
`U.S.A., Inc., No. 6:10-cv-493, Dkt. 659 (E.D. Tex. Feb. 11, 2013)
`Trial Transcript Vol. 5 from Realtime Data, LLC d/b/a IXO v. T-
`
`v
`
`EX. NO.
`1001
`1002
`1003
`1004
`
`1005
`
`1006
`
`1007
`
`1008
`
`1009
`
`1010
`
`1011
`
`1012
`
`1013
`
`1014
`
`1015
`
`
`
`

`

`EX. NO.
`
`1016
`
`1017
`
`1018
`
`1019
`
`1020
`
`1021
`
`1022
`
`1023
`
`1024
`
`1025
`
`1026
`
`1027
`
`BRIEF DESCRIPTION
`Mobile U.S.A., Inc., No. 6:10-cv-493(E.D. Tex. Feb. 8, 2013)
`Order Granting in Part Plaintiff’s Motion for Judgment as a
`Matter of Law as to Invalidity, Realtime Data, LLC d/b/a IXO v.
`T-Mobile U.S.A., Inc., No. 6:10-cv-49, Dkt. 662 (E.D. Tex. Mar.
`4, 2013)
`U.S. Patent No. 7,161, 506 (“the ’506 patent” or “the parent ’506
`patent”)
`Memorandum Opinion and Order, Realtime Data, LLC d/b/a IXO
`v. Packeteer, Inc., No. 6:08-cv-144, Dkt. 371 (E.D. Tex. Jun. 22,
`2009)
`Final Judgment, Realtime Data, LLC d/b/a IXO v. T-Mobile
`U.S.A., Inc., No. 6:10-cv-493, Dkt. 664 (E.D. Tex. Mar. 28, 2013)
`Decision on Appeal in Blue Coat Systems, Inc. v. Realtime Data
`LLC, Appeal 2012-002371, Reexamination Control No.
`95/000,479
`U.S. Patent No. 4,956,808 to Aakre et al. (“Aakre”)
`Notice of Allowance in U.S. Patent Application No. 14/495,574
`(Feb. 23, 2015)
`Amendment in U.S. Patent Application No. 10/668,768 (Aug. 25,
`2004)
`Decision to Institute Inter Partes Review of U.S. Patent No.
`7,378,992 (IPR2016-00373), Paper No. 7 (Jun. 27, 2016)
`Decision to Institute Inter Partes Review of U.S. Patent No.
`8,643,513 (IPR2016-00374) Paper No. 7 (Jun. 27, 2016)
`Declaration of Mr. Scott Bennett and Attachments 1a-h
`William Underwood, Extensions of the UNIX File Command and
`Magic File for File Type Identification, Technical Report
`ITTL/CSITD 09-02, Georgia Tech Research Institute (Sept.
`2009).
`
`
`
`vi
`
`

`

`EX. NO.
`
`1028
`
`1029
`1030
`
`1031
`
`1032
`
`BRIEF DESCRIPTION
`AT&T UNIX® PC UNIX System V User’s Manual, Volume 1
`(1986)
`File(1): FreeBSD General Commands Manual (Dec. 8, 2000)
`U.S. Patent No. 6,253,264 to Sebastian (“Sebastian”)
`Deposition of Kenneth A. Zeger, Ph.D. (Nov. 2, 2017) and errata
`sheet (taken in IPR2017-00176, -179, -806, -808).
`CRFD Research, Inc. v. Dish Network Corp., No. 16-2198 (Fed.
`Cir. Dec. 5, 2017) (precedential slip opinion)
`
`
`
`
`
`
`
`vii
`
`

`

`I.
`
`PRELIMINARY STATEMENT
`The Petition established a prima facie case of obviousness. The only
`
`difference between Franaszek and the challenged independent claims is the
`
`negative limitation—i.e., “exclud[ing] analy[sis] based solely on a descriptor that
`
`is indicative of one or more parameters or attributes of the data block.” Analysis
`
`of data blocks using something other than “solely” a descriptor was old; Hsu
`
`teaches it. Patent Owner does contest—and its expert admits—that a POSA would
`
`have seen an advantage in combining teachings of Franaszek and Hsu. The sole
`
`disputed issue is whether a POSA would have combined Franaszek and Hsu to
`
`arrive at the claimed subject matter. As demonstrated in the Petition, it would have
`
`been.
`
`Patent Owner’s arguments to the contrary are wrong on the law and facts.
`
`Patent Owner’s overarching argument presumes that a combination of prior art
`
`teachings must be the most obvious combination or that there be no other obvious
`
`ways to combine teachings. Such a conclusion is contrary to law. Next, Patent
`
`Owner misinterprets Dr. Creusere’s testimony that a POSA “could then apply Hsu
`
`to finding data type” as a POSA “would always then apply Hsu to finding data
`
`type.” “Could” does not mean “would always.”
`
`
`
`1
`
`

`

`Moreover, contrary to Patent Owner’s argument, a POSA would have found
`
`it obvious to avoid using Hsu when Franaszek lacked data type information
`
`because it could be the “most efficient” and “least time-intensive method to
`
`identify an effective encoder for that block.” Pet., 40-41. Patent Owner’s contrary
`
`argument means that a POSA would use Hsu to identify the wrong data type when
`
`Franaszek lacks a type descriptor. This makes little sense.
`
`Fourth, Patent Owner’s hindsight arguments are not well-founded. Although
`
`Patent Owner alleges that there would be “conflicting goals” if Petitioner’s logic is
`
`followed for combining Franaszek, Hsu, and Sebastian, the record evidence shows
`
`no inconsistencies. Instead the record reveals commonplace tradeoffs. Patent
`
`Owner’s concern that Dr. Creusere improperly looked to claim 1 also lacks merit.
`
`Dr. Creusere had to compare the prior art and the claim under Graham. That he
`
`reviewed the claim is no surprise.
`
`Fifth, Patent Owner’s argument that a POSA following the rationale in the
`
`Petition would have also used Sebastian’s generic encoder even when a correct
`
`data type was provided and an available encoder was linked to that data type is
`
`nonsense. All prior art of record explains that tailoring compression methods to
`
`data type is advantageous. A POSA would not have disregarded these teachings to
`
`follow the path Patent Owner ventures down.
`
`2
`
`
`
`

`

`Finally, while Patent Owner criticizes Petitioner for not including direct
`
`testimony from Dr. Creusere saying that a POSA would have found it obvious to
`
`forego using Hsu when Franaszek lacks data type, expert testimony is not the only
`
`way to prove a fact. The Board’s rules do not require any expert testimony. The
`
`Petition presented—and the fully-developed record further supports—a prima facie
`
`case of obviousness.
`
`II. THE TEACHINGS OF THE PRIOR ART ARE UNDISPUTED; THE
`SOLE QUESTION IS MOTIVATION TO ARRIVE AT THE
`CLAIMED SUBJECT MATTER.
`
`Many facts are uncontested. The Petition showed how Franaszek disclosed
`
`all aspects of challenged claims 1 and 24 except the negative limitation—i.e., that
`
`analyzing data within the data block “excludes analyzing based solely on a
`
`descriptor that is indicative of the one or more parameters or attributes of the data
`
`within the data block.” See Pet., 19-41, 49-56; Ex. 1001, 26:38-41, 28:21-23.
`
`During reexamination of the parent ’506 patent a very similar limitation was
`
`viewed as the reason that the claims were patentable over Franaszek and Sebastian.
`
`See Pet., 1-2; Ex. 1006, 3 (NIRC). This “excludes” limitation was not new; it was
`
`taught by Hsu. See Pet., 2-3.
`
`The Petition also established that a POSA would have combined Franaszek,
`
`Hsu, and Sebastian to arrive at the claimed subject matter. Patent Owner does not
`
`
`
`3
`
`

`

`genuinely dispute that a POSA would have combined Franaszek and Hsu. Dr.
`
`Zeger, Patent Owner’s expert, agrees with Petitioner that a POSA would have
`
`recognized that combining Franaszek’s and Hsu’s teachings would have “help[ed]
`
`Franaszek ‘pick a better compression technique and get more compression.’” Ex.
`
`2004, ¶24 (quoting Ex. 2003, 130:12-17). The undisputed facts thus show that a
`
`POSA would have been motivated to combine Franaszek and Hsu.
`
`Based on Patent Owner’s response, the crux of the dispute is whether a
`
`POSA would have combined references to arrive at the claimed subject matter—
`
`i.e., one in which a processor is configured to (1) compress data blocks using a
`
`“content dependent data compression” technique when “the parameters or
`
`attributes” are identified and (1) compress data blocks using a “single” or “default”
`
`data compression encoder when “the parameters or attributes” are not identified.
`
`While the Petition identified good reasons—supported by record evidence—why a
`
`POSA would have found the claims obvious, Patent Owner disagrees and argues
`
`that modifying Franaszek to look at something other than solely descriptors to
`
`arrive at the claimed subject matter is based only on hindsight.
`
`We show below why each of Patent Owner’s protestations miss the mark.
`
`
`
`4
`
`

`

`III. THE FULLY-DEVELOPED RECORD SHOWS A POSA WOULD
`HAVE BEEN LED TO THE CLAIMED SUBJECT MATTER
`EXERCISING ONLY ROUTINE SKILL.
`
`Patent Owner summarizes its argument as follows: “if the evidence
`
`demonstrates that a POSA would use Hsu to find a block’s type and/or calculate
`
`any of its redundancy metrics [when Franaszek lacks data type information], the
`
`Petition fails to show unpatentability.” POR, 9. Patent Owner then incorrectly
`
`contends that both experts agree that Hsu would have been used when Franaszek
`
`lacks a data type.
`
`There are four problems with Patent Owner’s argument.
`
`A.
`
`Patent Owner’s Theory Rests On The Legally-Unsupported
`Assumption That A POSA Would Have Only Seen One Obvious
`Way To Combine Teachings Of the Prior Art.
`
`Patent Owner presupposes that a POSA would only see one obvious way to
`
`combine prior art teachings. The law does not require a showing that a
`
`combination of prior art teachings is the most obvious way to combine teachings or
`
`that a POSA would not have seen other obvious ways to combine teachings. These
`
`are not found in Graham’s inquiries. Graham v. John Deere, 383 U.S. 1, 17
`
`(1966). When addressing the concept of teaching away (something Patent Owner
`
`does not rely on here), the Federal Circuit has explained that “just because better
`
`alternatives exist in the prior art does not mean that an inferior combination is inapt
`
`
`
`5
`
`

`

`for obviousness purposes . . . .” In re Mouttet, 686 F.3d 1322, 1333-34 (Fed. Cir.
`
`2012).
`
`To accept the premise underlying Patent Owner’s argument is to turn a blind
`
`eye on variations on the prior art that would have been obvious to those in the
`
`field—even if not the most obvious variations. This would unseat section 103 from
`
`its role in ensuring patents satisfy their constitutional purpose. Graham, 383 U.S.
`
`at 5-17 (“Congress may not authorize the issuance of patents whose effects are to
`
`remove existent knowledge from the public domain . . . .”). Patent Owner’s “if-
`
`then” proposition would allow patents on “advances that would occur in the
`
`ordinary course without real invention” thereby “retard[ing] progress . . . .” KSR
`
`Int’l Co. v. Teleflex Inc., 550 U.S. 398, 419 (2007). Showing one variant of the
`
`prior art would also have been obvious does not detract from a showing that a
`
`different variant on the prior art would also have been obvious.
`
`Here, even if it were shown that a POSA presented with Franaszek and Hsu
`
`would have found it obvious to use Hsu as Patent Owner proposes, that does not
`
`mean that it would not also have been obvious to use Hsu only when data blocks
`
`included Franaszek’s type information. As explained in the Petition, there were
`
`good reasons to implement a data compression system in the manner Petitioner
`
`proposed. Pet., 40-41. The fully-developed record does not alter that conclusion.
`
`
`
`6
`
`

`

`B.
`
`Patent Owner Makes An Improper Bodily Incorporation
`Argument.
`
`Patent Owner advances a theory about how a POSA might “combine Hsu
`
`into Franaszek’s Figure 4A,” POR, 11-12
`
`(emphasis added)
`
`(showing
`
`incorporating of Hsu a process disclosed in Franaszek). The relevant question,
`
`however, is what the teachings of the references would have suggested to a POSA,
`
`not whether one reference may be bodily incorporated into the structure of another.
`
`In re Keller, 642 F.2d 413, 425 (CCPA 1981). Patent Owner’s argument about
`
`what might happen when Hsu is placed into Franaszek should be rejected as a
`
`matter of law.
`
`C. Dr. Creusere Never Testified That Hsu Would Always Be Used To
`Determine Data Type.
`
`Dr. Creusere testified that when Franaszek’s data type information was
`
`missing “you could then apply Hsu to finding data type . . . .” Ex. 2003, 131:13-
`
`132:15 (emphasis added). Patent Owner and its expert conclude that when Dr.
`
`Creusere said that you “could” use Hsu to determine data type even if Franaszek
`
`lacked data type information, Dr. Creusere meant a person of ordinary skill “would
`
`always” determine a data type with Hsu. POR, 10 (“[A] POSA would always use
`
`Hsu to identify a block’s data type, i.e., a POSA would not refrain from using
`
`Hsu’s” techniques “when Franaszek did not have a data type . . . .” (emphasis
`
`
`
`7
`
`

`

`original)); Ex. 2004, ¶24. But Patent Owner’s position takes Dr. Creusere’s
`
`testimony too far.
`
`“Could” does not mean “would always.” Patent Owner’s argument that Dr.
`
`Creusere testified that “a person of skill in the art would not combine” Franaszek
`
`and Hsu as proposed in the Petition is not supported by his testimony.
`
`D. The Fully-Developed Record Continues to Show Good Reasons
`Why a POSA Would Not Use Hsu When Franaszek Lacked A
`Data Type Descriptor.
`
`The Petition explained that a POSA would have found it obvious to avoid
`
`using Hsu when Franaszek lacked a type descriptor because using “Franaszek’s
`
`default list of decoders [sic encoders] could be the most efficient and least time-
`
`intensive method to identify an effective encoder for that block.” Pet., 40-41
`
`(emphasis added) (citing Ex. 1029, 5; Ex. 1002, ¶¶66, 131-33, 146). The prior art
`
`recognized that as the number of available file types increased, the “file” program
`
`from which Hsu’s new-file program was derived would slow down. Pet., 40; Ex.
`
`1029, 5; Ex. 1002, ¶131.
`
`The fully-developed record shows that a POSA would have recognized that
`
`using Franaszek in combination with Hsu when Franaszek lacked a data type
`
`descriptor would have been less efficient and more time-intensive than using
`
`
`
`8
`
`

`

`Franaszek’s default encoding scheme for blocks with unknown data types as
`
`Franaszek intended.
`
`1.
`
`Both Experts Agree That Using Franaszek With Hsu When
`Franaszek Has A Data Type Is Advantageous.
`
`Both experts agree that a POSA would have seen an advantage in combining
`
`Franaszek’s and Hsu’s teachings. But Patent Owner misinterprets the Petition to
`
`depict what it believes is Petitioner’s “combination theory.”1 Petitioner’s position
`
`may be summarized as follows:
`
`If Franaszek has data type → Use Hsu to get more compression
`
`If Franaszek lacks data type → Do not use Hsu’s time-consuming
`
`technique to provide the wrong data type; use Franaszek’s default
`
`encoder for unknown data types.
`
`Both experts agree on the first point: if Franaszek has a data type, a POSA
`
`would have looked to Hsu to obtain more compression. See Ex. 2004, ¶24
`
`(agreeing with Dr. Creusere’s testimony “that using Hsu [when Franaszek has a
`
`
`
`1 Patent Owner says that the theory in the Petition is: “If Franaszek has data type
`
`→ Use Hsu to find type and three metrics” and “If Franaszek lacks data type → Do
`
`not use Hsu to find type or three metrics.” POR, 9 (emphasis original).
`
`
`
`9
`
`

`

`descriptor] would help Franaszek ‘pick a better compression technique and get
`
`more compression.’ [Ex. 2003] at 130:12-17”); Ex. 1002, ¶¶125-26, 130-33.
`
`Regarding the second point, the Petition presented ample evidence that a
`
`POSA would have found it obvious to not waste time using use Hsu where
`
`Franaszek lacks data type information. Ex. 1002, ¶¶ 130-33 (opining that Hsu’s
`
`data type identification process could be time consuming). Patent Owner’s
`
`position assumes that a POSA reading Franaszek and Hsu together—as they
`
`must2—would have understood Hsu to be capable of identifying the actual data
`
`type information Franaszek lacked. That is not so. Even assuming Hsu would
`
`always return some data type even when Franaszek lacked data type, Hsu would
`
`provide the wrong type information, rendering the system less efficient and
`
`slower—just as contended in the Petition.3
`
`
`
`2 In re Keller, 642 F.2d at 426 (“[O]ne cannot show non-obviousness by attacking
`
`references individually” when challenge is based “on combinations of references”).
`
`3 Patent Owner cites Dr. Creusere’s deposition for the proposition that he agreed
`
`that a POSA would use Hsu when Franaszek lacked data type to “pick a better
`
`compression technique and get more compression.” POR, 13 (quoting Ex. 2003,
`
`130:12-17). The question asked of Dr. Creusere, however, presupposes that Hsu
`
`
`
`10
`
`

`

`Dr. Zeger explained that Franaszek might lack a descriptor where the data
`
`was of an unknown type.
`
`Q: . . . What you said was – after you said mislabeled
`“It just was inadvertently left unlabeled. It could be that
`there was no accurate type that was available for it.”
`That’s what you said.
`So, could you explain that for me a little bit more?
`
`
`
`will “recognize the data type.” Ex. 2003, 130:12-17. Certainly, if Hsu provided
`
`accurate data type information to Franaszek and there was an encoder associated
`
`with that data type—as Dr. Zeger contends there always would be, see Ex. 1031,
`
`162:17-163:2, 178:10-16—then Hsu would help Franaszek chose the associated
`
`compression algorithms for that data. Dr. Creusere agreed. Ex. 2003, 130:12-17.
`
`But, the question does not address the circumstance in which Hsu is viewed as
`
`always identifying a data type even if it were wrong. Ex. 1002, ¶66 (noting that a
`
`POSA could understand Hsu to “guess” at data types, but recognizing there may be
`
`downsides with the approach, but noting other interpretations of Hsu); Ex, 1031,
`
`95:20-24. The testimony relied on by Patent Owner merely responds to the
`
`assumption in the question—i.e., that Hsu recognizes the actual data type; it does
`
`not address the situation in which Hsu guesses wrong.
`
`11
`
`
`
`

`

`A: That’s basically the situation where it’s an unknown type.
`Ex. 1031, 55:1-13; id., 53:24-54:17 (Franaszek might not have a descriptor because
`
`the block “might be an unknown data type”). Systems in which data types are
`
`unrecognized are a reality in the field of data compression. Id., 55:16-22.
`
`According to Dr. Zeger, a POSA would have used Hsu to determine the data
`
`type if Franaszek lacked this information. However, Dr. Zeger’s testimony about
`
`Hsu shows that Hsu would, at best, have returned incorrect data type information.
`
`When Hsu receives a block with a data type that is not specified in Hsu’s list, Hsu
`
`will guess at the data type knowing full well that the guess is wrong:
`
`Q: So would it be fair to say that when Hsu encounters a data block
`that doesn’t, you know, contain one of the classes of data that are
`spelled out here, it will do its best to classify it as one of these types,
`but it might be wrong?
`A: That’s generally the way any kind of recognition algorithm
`works. That’s how – anything that’s trying to tell you what
`category something fits in or any type of pattern matching will tell
`you that – its best guess from among a predefined set of values.
`And in this case there’s ten different types, and Hsu would try to
`give you its best guess. . . . .
`Ex. 1031, 97:18-98:9. When asked about how Hsu would handle “randomized
`
`data” (a type not specified by Hsu) Dr. Zeger testified that Hsu would classify that
`
`data into one of Hsu’s ten types—even though it would be wrong. Id., 98:10-25.
`
`
`
`12
`
`

`

`The same wrong guess would occur with video data (a type also not specified by
`
`Hsu). Id., 95:25-97:17. Dr. Creusere acknowledged that Hsu may “guess the file
`
`type regardless of whether the guess was made confidently or not.” Ex. 1002, ¶65.
`
`Taking the teachings of Franaszek and Hsu together, a POSA would have
`
`recognized that a system devised based on their combined teachings could
`
`encounter blocks of unknown data type and would need to figure out how to
`
`compress those blocks. A POSA would have clear reasons why she would not
`
`want to waste time using Hsu when Franaszek’s data type field is unpopulated.
`
`The data type could be missing because the file type is unrecognized. Ex. 1031,
`
`53:1-13, 53:24-54:17; Ex. 1002, ¶131 (Franaszek uses a “default” encoder to
`
`compress “a data block having an unknown data type”). Under Dr. Zeger’s view
`
`of Hsu, a POSA would have been left with a binary choice when a data type is
`
`unrecognized: (1) use Hsu’s time-consuming process to guess at the data type
`
`knowing it would be wrong and use Franaszek’s encoders tailored to the wrong
`
`data type, contrary to Franaszek’s teachings, see Ex. 1004, 5:49-54, or (2) use
`
`Franaszek’s “default list of compression methods,” which a POSA would have
`
`understood to be designed for use when “no data type is available,” see id., 5:53-
`
`54. Rather than apply Hsu knowing it was time consuming and would arrive at the
`
`wrong result, a POSA would have used the tool meant for unknown data types—
`
`
`
`13
`
`

`

`i.e., Franaszek’s single default encoder. Patent Owner’s argument to the contrary
`
`makes little sense.
`
`IV. PATENT OWNER’S ARGUMENT THAT FRANASZEK AND HSU
`CANNOT MEET LIMITATION 1[F] IS IRRELEVANT.
`
`Patent Owner contends that “Dr. Creusere’s testimony additionally dooms
`
`the Petition because his testimony establishes that whenever Hsu is used, limitation
`
`1[f] cannot be met.” POR, 14 (emphasis added). But, as explained in the Petition
`
`and further discussed above, a POSA would have found it obvious to refrain from
`
`using Hsu when Franaszek’s data blocks lacked a type descriptor; a POSA would
`
`have found it obvious to continue to rely on Franaszek’s single default encoder.
`
`Therefore, Patent Owner’s arguments about what happens “whenever Hsu is used”
`
`has no bearing on the analysis presented for limitation 1[f], which does not rely on
`
`Hsu’s techniques.
`
`V.
`
`PATENT OWNER’S ARGUMENTS ABOUT HINDSIGHT ARE
`MISGUIDED.
`A. The Board Need Not Look To Sebastian To Find The Claimed
`Subject Matter Obvious. Patent Owner’s Contentions About
`Inconsistencies In Motivations Concerning Sebastian Are Thus
`Irrelevant.
`
`Patent Owner contends that the motivations to combine Franaszek and Hsu
`
`and Franaszek and Sebastian are “internally
`
`inconsistent,” “diametrically
`
`oppos[ed],” and are based on “conflicting goals.” POR, 1-2, 19-22. While these
`
`
`
`14
`
`

`

`arguments are unsound, the Board need not even address these criticisms to the
`
`extent they are based on motivations related to Sebastian.
`
`The sole reason that Sebastian was relied upon in the Petition was to address
`
`a potential (but incorrect) argument by Patent Owner that Franaszek did not
`
`disclose the “single” or “default” data compression encoder. See Pet., 28-30, 51.
`
`The use of Sebastian as an alternative argument was set forth in the Petition, id.,
`
`recognized by the Board in its Institution Decision, Inst. Dec., 8 n.3 (interpreting
`
`Ground 1 as encompassing “two alternative combinations of references” and
`
`choosing the one that “encompasses the combination of Franaszek and Hsu”), and
`
`acknowledged by Patent Owner as an “alternative theory,” e.g., POR, 17; Ex.
`
`2004, ¶29. Patent Owner has not contended that Franaszek lacks the “single” or
`
`“default” encoder of claims 1 and 24. And rightly so.
`
`While the Board exercised its discretion and denied institution based on
`
`Franaszek and Hsu alone for efficiency reasons, see Inst. Dec., 8 n.3, the analysis
`
`of Franaszek and Hsu was intertwined with the alternative argument based on
`
`Franaszek, Hsu, and Sebastian. Barring Petitioners from pursuing a ground raised
`
`in the Petition and was “expressly incorporated into other proposed grounds of
`
`unpatentability on which the Board instituted would . . . unfairly prejudice”
`
`Petitioners. CRFD Res., Inc. v. Dish Network Corp., No. 2016-2198, Slip op., p.29
`
`
`
`15
`
`

`

`(Fed. Cir. Dec. 5, 2017) (Ex. 1032). At least one Board decision further confirms
`
`that the Board need not reach Sebastian’s teachings to find the challenged claims
`
`obvious. Mobotix Corp. v. E-Watch, Inc., IPR2013-00334, Paper 18 (PTAB Jan.
`
`10, 2014) (where trial instituted on three-way combination of references
`
`“Petitioner is free to prove the obviousness of claims 1-10 based on” one reference
`
`alone); cf. Application of Kronig, 539 F.2d 1300, 1302-03 (CCPA 1976) (affirming
`
`Board obviousness rejection based on three references where examiner relied on
`
`four reference combination because three reference combination was not a new
`
`grounds of rejection). Nonetheless, the combination of Franaszek, Hsu, and
`
`Sebastian renders the subject matter of claims 1 and 24 obvious as described in the
`
`Petition.4
`
`B.
`
`Patent Owner’s Expert Agrees There Was An Advantage To
`Combining Franaszek’s and Hsu’s Teachings.
`
`The Petition established several reasons why a POSA would have looked to
`
`Hsu’s teachings of analyzing something other than solely a descriptor as an
`
`obvious modification of Franaszek. Pet., 34-39. These reasons included: (1)
`
`
`
`4 Patent Owner has not challenged the showing of obviousness as to the challenged
`
`dependent claims apart from its objections to the proofs as to claims 1 and 24.
`
`Therefore, those claims are not addressed in this Reply.
`
`
`
`16
`
`

`

`“giving further insight into optimal methods for compressing” a data block, id., 36;
`
`(2) the ability to recognize and classify heterogeneous files, id., 36-38; and (3) the
`
`potential to speed up identification of data type, id., 38. Patent Owner’s expert, Dr.
`
`Zeger agreed that “using Hsu [when Franaszek has data type information] would
`
`help Franaszek ‘pick a better compression technique and get more compression.’”
`
`Ex. 2004, ¶24 (quoting Ex. 2003, 130:12-17). Dr. Zeger, therefore, agrees with at
`
`least one of Petitioner’s stated motivations to combine Franaszek and Hsu.5
`
`All evidence in this proceeding shows that a POSA would have had reasons
`
`to modify Franaszek with Hsu such that Franaszek did not rely “solely” on a
`
`
`
`5 Dr. Zeger’s agreement with Petitioner also casts doubt on the veracity of Dr.
`
`Z