`
`BEFORE THE PATENT TRIAL AND APPEALS BOARD
`
`INTERNATIONAL BUSINESS MACHINES CORPORATION
`Petitioner
`
`v.
`
`INTELLECTUAL VENTURES II LLC
`Patent Owner
`
`Case No.: IPR2015-01089
`Patent No.: 6,818,271
`
`PETITIONER’S REPLY TO PATENT OWNER’S RESPONSE
`
`
`
`
`
`
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`TABLE OF CONTENTS
`
`Page
`
`I.
`
`GLOVER DISCLOSES PCES OPERABLE TO RECEIVE A
`PORTION OF UNCOMPRESSED DATA (CLAIMS 8 AND 32-34)........... 1
`
`A. Data Output From The Subband Unit In Glover Is
`Uncompressed. ...................................................................................... 2
`
`B.
`
`Data Output From The DPCM In Glover Is Uncompressed. ................ 4
`
`II.
`
`GLOVER IN COMBINATION WITH CHU DISCLOSES PCES
`THAT COMPRESS THE UNCOMPRESSED DATA................................... 6
`
`A. Glover In Combination With Chu Discloses PCEs That
`Compress The Portion Of Uncompressed Data Using The PDA
`(Claim 8). ............................................................................................... 7
`
`B.
`
`C.
`
`Glover In Combination With Chu Discloses PCEs That Are
`Configured To Compare Symbols With Entries in A History
`Table (Claims 32-34)............................................................................. 9
`
`Glover In Combination With Ranganathan Discloses PDEs
`That Are Operable To Receive Compressed Data Comprising
`Tokens Describing Symbols In Uncompressed Data (Claims 72,
`110). ..................................................................................................... 10
`
`III. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES
`THAT IMPLEMENT PCAS (CLAIMS 8, 32-34), AND GLOVER IN
`COMBINATION WITH RANGANATHAN DISCLOSES PDES
`THAT IMPLEMENT PDAS (CLAIMS 72, 110). ........................................ 11
`
`IV. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES
`THAT OPERATE INDEPENDENTLY (CLAIMS 8, 32-34), AND
`GLOVER IN COMBINATION WITH RANGANATHAN
`DISCLOSES PDES THAT OPERATE INDEPENDENTLY
`(CLAIMS 72, 110)......................................................................................... 14
`
`A.
`
`IV’s Purported “Purpose” Of Glover Is Not Supported By The
`Evidence. ............................................................................................. 15
`
`i
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`B.
`
`IV’s Argument That Glover Does Not Disclose PCEs And
`PDEs That Operate Independently Is Contradicted By The
`Evidence. ............................................................................................. 17
`
`V. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES
`THAT COMPRESS IN A PARALLEL FASHION (CLAIMS 8, 33-
`34), AND GLOVER IN COMBINATION WITH RANGANATHAN
`DISCLOSES PDES THAT DECOMPRESS IN A PARALLEL
`FASHION (CLAIMS 72 AND 110). ............................................................ 22
`
`VI. CONCLUSION .............................................................................................. 25
`
`
`
`
`
`ii
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`TABLE OF AUTHORITIES
`
`Page(s)
`
`Cases
`
`DyStar Textilfarben GmbH & Co. Deutschland KG v. C.H. Patrick Co.,
`464 F.3d 1356 (Fed. Cir. 2006) ............................................................................13
`
`In re Kahn,
`441 F.3d 977 (Fed. Cir. 2006) ..............................................................................13
`
`KSR Int’l Co. v. Teleflex, Inc.,
`550 U.S. 398 (2007) ................................................................................. 12, 18, 22
`
`Randall Mfg. v. Rea,
`733 F.3d 1355 (Fed. Cir. 2013) ..................................................................... 12, 22
`
`Statutes
`
`35 U.S.C. § 103 .......................................................................................................... 1
`
`iii
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`IV’s arguments are solely focused on Glover—IV does not dispute that a
`
`person having ordinary skill in the art (“PHOSITA”) would have been motivated to
`
`combine Glover with either Chu or Ranganathan, or that Chu and Ranganathan
`
`disclose parallel compression/decompression algorithms. In order to make its
`
`arguments, IV misreads the Challenged Claims in a way not supported by the ’271
`
`patent, and reads a strained purported “purpose” into Glover that is not supported
`
`by its disclosure. IV’s arguments are both legally and technically flawed, and the
`
`differences between Glover and the ’271 patent “are such that the subject matter as
`
`a whole would have been obvious at the time the invention was made” to a
`
`PHOSITA. 35 U.S.C. § 103 (pre-AIA) (emphasis added).
`
`I.
`
`GLOVER DISCLOSES PCES OPERABLE TO RECEIVE A
`PORTION OF UNCOMPRESSED DATA (CLAIMS 8 AND 32-34).
`
`IV’s contention that Glover does not disclose parallel compression engines
`
`(“PCEs”) that “receive a different respective portion of the uncompressed data”
`
`(Ex. 1004 at claim 8) or “receive the uncompressed portion of the data” (Ex. 1004
`
`at claims 32-34) conflates processing data with compressing data. Subband
`
`transform means 4 (“subband unit”) and Differential1 Pulse Code Modulator 102
`
`
`1 Glover alternatively refers to this as Differential, Data, or Digital Pulse Code
`
`Modulation. Ex. 1010 at 2:55-57, 5:1-5, 5:19-21. The parties agree that Glover
`
`implements differential pulse code modulation. Ex. 1026 at 235:7-11.
`
`1
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`(“DPCM”) in Glover both process uncompressed data in some way, but they do
`
`not compress data.
`
`A. Data Output From The Subband Unit In Glover Is Uncompressed.
`
`While IV states in heading IV.2.i that the data output from the subband unit
`
`in Glover is “not uncompressed,” IV never actually contends that the subband unit
`
`compresses data. IV argues instead that the filtered subbands output from the
`
`subband unit in Glover are “not original, unmodified, and uncompressed data, but
`
`rather have been modified and filtered.” Paper 13 at 22 (quoting Ex. 2001 ¶ 78);
`
`see also Paper 13 at 20; Ex. 2001 ¶¶ 72, 77. But claims 8 and 32-34 of the ’271
`
`patent only require that the PCEs receive a portion of uncompressed data, not
`
`“original” and “unmodified” data. Ex. 1004 at claims 8, 32-34. IV’s argument is
`
`thus based upon a requirement that is not found in the ’271 patent claims.
`
`IV’s argument is also incorrect technically.
`
` The NASA technical
`
`memorandum co-authored by Dr. Glover that IV relies upon as “an example of
`
`how data is transformed” by the subband unit in Glover undermines IV’s argument
`
`that “modified and filtered” data is equivalent to “compressed” data. Paper 13 at
`
`21-22 (citing Ex. 2005). IV’s expert, Dr. Richardson, opined that the output of the
`
`subband unit would be similar to the subbands in Figure 5 of Dr. Glover’s memo.
`
`Ex. 2001 ¶ 78; Ex. 2005. Dr. Glover’s memo expressly states that after data has
`
`been split into the subbands, “[n]o compression has been achieved yet; the
`
`2
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`subbanding process prepares the data for compression coding.” Ex. 2005 at 3
`
`(emphasis added); see also Ex. 2020 at 5 (“transforms do not in themselves
`
`produce data compression, but prepare
`
`the data
`
`for quantization and
`
`compression”); Ex. 1025 ¶ 4. This could not be clearer: the subbands output from
`
`the subband unit are not compressed data.
`
`Glover also explained that the subband unit “separates the signal into a
`
`plurality of subbands,” not that it compresses data. Ex. 1010 at 5:17-19; see also
`
`id. at Abstract (the incoming data is “separated into subbands,” after which “it can
`
`be coded”); Ex. 1001 ¶ 57; Ex. 2017 at 43:24-44:12; Ex. 1025 ¶ 3. IV itself admits
`
`that the subband unit separates the incoming data into four data streams, filtered by
`
`frequency. Paper 13 at 20-22; Ex. 2001 ¶¶ 76-78. Dr. Richardson stated that
`
`“[t]he information content of band 0 is significantly reduced compared with the
`
`original image.” Ex. 2001 ¶ 56; see also id. ¶ 78. But considering the original
`
`image has been split into four subbands, it is entirely expected that each subband
`
`would have only a portion of the information contained in the original image. That
`
`is not evidence of compression. Ex. 2017 at 59:12-61:7 (the subband unit “doesn’t
`
`change the size of the data”). The ’271 patent expects this, as the PCEs each
`
`receive a portion of uncompressed data. Ex. 1004 at claims 8, 32-34.
`
`3
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`B. Data Output From The DPCM In Glover Is Uncompressed.
`
`The portion of uncompressed data output from the subband unit of Glover in
`
`subband 0 is put through an additional processor, the DPCM. Ex. 1010 at Fig. 1,
`
`5:19-21. IV contends that the differential pulse code modulation (“DPC
`
`modulation”) performed by the DPCM “is a well-known form of compression,”
`
`such that the data received by the quantizer in subband 0 is not uncompressed.
`
`Paper 13 at 22; Ex. 2001 ¶ 80. But IV offers nothing other than conclusory, ipse
`
`dixit statements from its expert in support of this contention. See id.
`
`Dr. Richardson cites to paragraphs 59-62 of his declaration as allegedly
`
`demonstrating
`
`that DPC modulation
`
`is a form of compression, but,
`
`in
`
`Dr. Richardson’s own words,
`
`those paragraphs only
`
`illustrate how DPC
`
`modulation “makes Band 0 more similar to the other Bands.” Ex. 2001 ¶¶ 59-62,
`
`80. There is nothing in the action of allegedly making subband 0 more similar to
`
`the other subbands that equates to compressing the data in subband 0.
`
`DPC modulation is a common processing step used in compression systems
`
`that can serve to enhance the amount of compression of the system. Ex. 1025 ¶ 5.
`
`DPC modulation itself, however, is not a form of compression. Ex. 1025 ¶¶ 6-7;
`
`Ex. 2017 at 111:4-112:8. The DPCM in Glover uses a predictor equation to
`
`perform DPC modulation. Ex. 1010 at 5:1-5. As Dr. Richardson himself describes
`
`it, this predictor outputs difference values so that “there is less variation in the
`
`4
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`output value from an expected mean.” Ex. 2001 ¶ 32. Glover expressly states that
`
`“[t]he quantizer is applied to subbanded image data that has not been scaled or
`
`rounded”—i.e., the output from the DPCM, to which quantization is subsequently
`
`applied, has not been compressed. Ex. 1010 at 3:23-33. The DPCM transforms
`
`the data in a signal by outputting difference values, but it does not compress the
`
`data or reduce the amount of information in the signal. Ex. 1025 ¶¶ 5-7.
`
`Dr. Richardson claims that DPC modulation “reduces the information in a
`
`signal.” Ex. 2001 ¶¶ 32, 73. Not only does Dr. Richardson admit that that is not
`
`true for all images, this statement is misleading. Ex. 1026 at 241:4-242:12; Ex.
`
`2023 at 68:19-21, 71:6-9 (Dr. Richardson testifying that reducing the information
`
`content does not necessarily compress the data). Dr. Richardson does not actually
`
`say that the reduction in the information of a signal is a reduction in the size of the
`
`signal, Ex. 2001 ¶¶ 32, 73, and, in any event, DPC modulation does not reduce the
`
`information of a signal. Ex. 1025 ¶ 7. Even if there are situations in which the
`
`DPCM reduces the amount of “information” in subband 0, Glover recognizes that
`
`this
`
`is not compression, as Glover distinguishes DPC modulation from
`
`compression. It is only after DPC modulation and quantization that “an adaptive
`
`Huffman coder or Lempel-Ziv based coder is used to perform compression.” Ex.
`
`1010 at 5:13-15.
`
`5
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`Most of IV’s argument about the DPCM is an effort to show that the DPCM
`
`is necessary in Glover, not that it actually compresses data. See Paper 13 at 22-27.
`
`IV’s argument is both irrelevant and incorrect. As IV acknowledges, Glover
`
`expressly discloses a scenario in which the compression algorithms are applied
`
`directly to the subbanded data, without first applying the DPCM (or quantizers).
`
`Ex. 1010 at 1:63-2:5; Paper 13 at 23; Ex. 2001 ¶ 73; see also Ex. 1025 ¶ 8. Claim
`
`1 of Glover does not even include a DPCM. Ex. 1010 at claim 1. IV’s argument
`
`also relies heavily on Dr. Richardson’s purported model of Glover applied to a
`
`sample image (the Lenna or Lena image). This model is of only one image, when
`
`Glover discloses that its system will have different results for different images. Ex.
`
`1010 at Table II; Ex. 1025 ¶ 9. Dr. Richardson admits that his model is “illustrative
`
`in nature” and “not intended to be an exact implementation of Glover.” Ex. 2001
`
`¶ 46; see also Ex. 2023 at 15:13-20. IV’s heavy reliance on this model is
`
`misplaced.
`
`II. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES THAT
`COMPRESS THE UNCOMPRESSED DATA.
`
`The arguments in sections IV.3, IV.4, and IV.5 of IV’s Patent Owner
`
`Response are all variations on a theme—that certain claim limitations are not
`
`disclosed by Glover in combination with Chu or Ranganathan because the
`
`uncompressed data is processed and compressed prior to reaching the parallel
`
`compression algorithm (“PCA”) in the PCE. See generally Paper 13 at 27-32.
`
`6
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`These arguments are both technically wrong and based upon unduly limited
`
`interpretations of the claim language.
`
`A. Glover In Combination With Chu Discloses PCEs That Compress
`The Portion Of Uncompressed Data Using The PDA (Claim 8).
`
`According to IV, the PCA disclosed by the combination of Glover and Chu
`
`“would not compress uncompressed data” because the uncompressed data in
`
`Glover—uncompressed video signal 2—is processed by the subband transform,
`
`DPCM, Quantizer, Run-Length coder, and Huffman coder before reaching the
`
`Lempel-Ziv (“LZ”) coder. Paper 13 at 27-29. IV’s argument fails because it
`
`assumes that the PCE must compress an uncompressed portion of the
`
`uncompressed data using the PCA. Paper 13 at 28-29. That requirement is found
`
`nowhere in claim 8.
`
`Claim 8 requires PCEs that are configured to “compress the different
`
`respective portion of the uncompressed data” using the PCA. Ex. 1004 at claim 8
`
`(emphasis added). The ’271 patent expressly contemplates that each PCE will
`
`receive only a portion of the uncompressed data, and thus that there is some
`
`processing done to divide the uncompressed data into portions before the each
`
`portion of data is provided to and compressed by a PCE. Ex. 1004 at claim 8, Fig.
`
`1 (illustrating that the uncompressed data goes through “splitting logic” before
`
`entering the PCEs). IV’s contention that each PCE must compress the
`
`uncompressed data (i.e., uncompressed video signal 2 in Glover) using the PCA is
`
`7
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`inconsistent with the claim and teachings of the ’271 patent. Claim 8 requires only
`
`that each PCE compress a respective portion—whether that portion is compressed
`
`or uncompressed—of the uncompressed data.
`
`Because claim 8 does not require that the portion of the uncompressed data
`
`be an uncompressed portion of the uncompressed data, it is irrelevant whether the
`
`components that process that portion of data prior to the PCA also compress the
`
`data. The PCA of each PCE in Glover combined with Chu is the sequence of Run-
`
`Length, Huffman, and parallel LZ compression algorithms. Paper 2 at 28-29. The
`
`PCE created by the combination of Glover and Chu is “operable to ... compress the
`
`different respective portion of the uncompressed data” because the PCA—the
`
`combination of Run-Length, Huffman, and parallel Lempel-Ziv algorithms—
`
`compresses “the different respective portion of the uncompressed data” (i.e., the
`
`subband) received by the PCE. Paper 2 at 27-29; Ex. 1001 ¶¶ 142-144, 148-150;
`
`Ex. 1025 ¶ 11. Whether the processing components (subband transform, DPCM,
`
`and quantizers) that come before the PCA compress “the different respective
`
`portion of the uncompressed data” is irrelevant. See Ex. 1004 at claim 8.
`
`Even if claim 8 did require that the PCA compress an uncompressed portion
`
`of the uncompressed data, IV’s argument fails. The subband transform and DPCM
`
`do not compress data, as explained above. See supra Section I; Ex. 1025 ¶¶ 3-8.
`
`The quantizers do not compress data either. Ex. 1025 ¶¶ 12-14; Ex. 2017 at 139:5-
`
`8
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`25; Ex. 2022 at 40:19-41:21. Glover distinguishes quantization from compression,
`
`stating that it is only after DPC modulation and quantization that “an adaptive
`
`Huffman coder or Lempel-Ziv based coder is used to perform compression.” Ex.
`
`1010 at 5:13-15, 5:24-27, claim 1 (“quantizers for mapping” and “coders for
`
`compressing”). And contrary to IV’s misstatement, the Run-Length and Huffman
`
`algorithms are part of the PCA. Paper 2 at 28-29.
`
`B. Glover In Combination With Chu Discloses PCEs That Are
`Configured To Compare Symbols With Entries in A History Table
`(Claims 32-34).
`
`IV’s argument that Glover in combination with Chu does not disclose PCEs
`
`configured to “compare the plurality of symbols with entries in the history table” is
`
`analogous to its arguments about whether the PCAs compress the uncompressed
`
`data and suffers from similar flaws. See Paper 13 at 29-30.
`
`IV does not dispute that, in compressing the “received uncompressed portion
`
`of the data,” the PCE of Glover combined with Chu compares a plurality of
`
`symbols with entries in a history table, where those entries are also symbols. See
`
`id.; Paper 2 at 22-23, 33-34. IV’s only argument is that the compared symbols are
`
`not the plurality of symbols that represent the uncompressed data, because the data
`
`has been compressed by a Quantizer, Run-Length coder and Huffman coder before
`
`the PCA compares it to entries in a history table. Paper 13 at 29-30. But even
`
`after it has been processed by the Quantizer, Run-Length coder, and Huffman
`
`9
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`coder, the data present in each subband is still in the form of symbols, which
`
`represent the uncompressed portion of data received by the PCE. Ex. 1025 ¶ 15. If
`
`the data compressed by the LZ coder no longer represented the uncompressed data,
`
`there would be no way to reverse the process and retrieve the original image.
`
`C. Glover In Combination With Ranganathan Discloses PDEs That Are
`Operable To Receive Compressed Data Comprising Tokens
`Describing Symbols In Uncompressed Data (Claims 72, 110).
`
`IV’s argument that Glover in combination with Ranganathan does not
`
`disclose parallel decompression engines (“PDEs”) operable to receive a “different
`
`respective portion of the compressed data [that] comprises tokens each describing
`
`one or more of the symbols in the uncompressed data” again stems from its
`
`arguments about whether the PCAs compress the uncompressed data, and is based
`
`upon a misinterpretation of the claim language. See Paper 13 at 30-32.
`
`Claims 72 and 110 recite that the PDEs are configured to receive a portion
`
`of the compressed data, where the compressed data comprises “a compressed
`
`representation of uncompressed data” and “tokens each describing one or more
`
`symbols in the uncompressed data.” Ex. 1004 at claims 72, 110. IV contends that
`
`the tokens of the compressed data in Glover combined with Ranganathan do not
`
`“describe one or more symbols in the uncompressed data” because the
`
`uncompressed data was “transformed, compressed, and processed” prior to the LZ
`
`coder, and the tokens “at best” describe the input to the LZ coder. Paper 13 at 31-
`
`10
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`32. This argument presumes that the tokens of the ’271 patent must do more than
`
`“describe” the symbols in the uncompressed data, they must directly correlate to
`
`the exact symbols of the uncompressed data. IV offers no justification for such a
`
`narrow reading of the claims.
`
`IV does not dispute that the “Data In” of Figure 2 of Glover is compressed
`
`data, that it is a representation of the uncompressed data, or that it is comprised of
`
`tokens. See Paper 13 at 30-32. Those tokens “describe” one or more symbols in
`
`the uncompressed data so that the compression process can be reversed and the
`
`uncompressed data—a video image in Glover—can be retrieved. Ex. 1025 ¶ 15.
`
`III. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES THAT
`IMPLEMENT PCAS (CLAIMS 8, 32-34), AND GLOVER IN
`COMBINATION WITH RANGANATHAN DISCLOSES PDES THAT
`IMPLEMENT PDAS (CLAIMS 72, 110).
`
`In its Petition, IBM explained that a PHOSITA would have been motivated
`
`to combine Glover and Chu, and would have used the parallel implementation of a
`
`LZ compression algorithm taught by Chu as the “LZ coder” in Glover. Paper 2 at
`
`25, 27-28. A PHOSITA would also have been motivated to combine Glover and
`
`Ranganathan, and would have used the parallel implementation of a LZ
`
`decompression algorithm taught by Ranganathan as the “LZ decoder” in Glover.
`
`Paper 2 at 48-49, 51-52. IV does not dispute that Chu and Ranganathan disclose
`
`parallel LZ algorithms, or that a PHOSITA would be motivated to combine Glover
`
`with Chu and Ranganathan. See Paper 13 at 33-36. There is thus no real dispute
`
`11
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`that the combination of Glover with Chu discloses PCEs that implement PCAs, or
`
`that the combination of Glover with Ranganathan discloses PDEs that implement
`
`PDEs.
`
`In an attempt to manufacture a dispute, IV raises a number of irrelevant
`
`issues. Contrary to IV’s suggestion, Glover does not need to use the words
`
`“parallel,” “parallel compression,” “concurrently,” or any variations of those terms
`
`to make it obvious to a PHOSITA to implement PCAs in the PCEs, and PDAs in
`
`the PDEs, of Glover. Paper 13 at 33-34. In an obviousness analysis, “it will be
`
`necessary for a court to look to . . . the background knowledge” of a PHOSITA.
`
`KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398, 417-18 (2007); Randall Mfg. v. Rea,
`
`733 F.3d 1355, 1363 (Fed. Cir. 2013). A PHOSITA may use that knowledge to
`
`apply a known technique to improve a device, or combine familiar elements
`
`according to known methods, each of which would have been obvious. KSR, 550
`
`U.S. at 416-17.
`
`Although IV does not dispute that a PHOSITA would have been motivated
`
`to combine Glover with Chu and Ranganathan, it suggests that using parallel
`
`algorithms with Glover would be an application of hindsight because Glover does
`
`not mention parallel compression. Paper 13 at 33-34. Glover’s disclosure,
`
`however, would have been understood by a PHOSITA as describing a flexible,
`
`inexpensive, and fast architecture. Ex. 2017 at 171:17-174:5; Ex. 1025 ¶¶ 17-18.
`
`12
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`In any event, a motivation “does not have to be found explicitly in the prior art.”
`
`In re Kahn, 441 F.3d 977, 987-988 (Fed. Cir. 2006). The question is whether a
`
`PHOSITA, faced with the problem sought to be solved by the inventor of the ’271
`
`patent, would have been motivated to combine the teachings of Glover with his or
`
`her knowledge that parallel processing increases speed. See id. at 988. The ’271
`
`patent says that its goal was a system and method “for the reduction of system
`
`bandwidth and improved efficiency.” Ex. 1004 at 2:36-41. A PHOSITA would
`
`have been motivated to improve the efficiency of Glover by using parallel
`
`algorithms, rendering the Challenged Claims obvious. See DyStar Textilfarben
`
`GmbH & Co. Deutschland KG v. C.H. Patrick Co., 464 F.3d 1356, 1368-69 (Fed.
`
`Cir. 2006) (“an implicit motivation to combine exists” when the combination
`
`results in a faster or more efficient process).
`
`IBM’s position in separate proceedings that a PHOSITA would have
`
`pipelined the coders and decoders in Glover is in no way inconsistent with IBM’s
`
`obviousness combinations in this proceeding. Ex. 1025 ¶ 19. It is also irrelevant.
`
`What is at issue here are the arguments IBM presented in its Petition in this
`
`proceeding, relating to a set of claims with different elements than in prior
`
`proceedings. IV does not dispute that Chu and Ranganathan disclose parallel LZ
`
`algorithms. When those LZ algorithms are used in the LZ coder and decoder of
`
`Glover, the combination discloses PCEs that implement PCAs and PDEs that
`
`13
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`implement PDAs. Whether “the various process stages would be performed
`
`serially” has no bearing on that combination or the conclusion that the combination
`
`discloses the claim elements. See Paper 13 at 35-36.
`
`IV’s obviousness analysis is also suspect because IV and Dr. Richardson
`
`performed the analysis as of the wrong date. IBM established in its Petition that
`
`the ’271 patent is only entitled to claim priority to its filing date, January 11, 2002,
`
`because that application is the only one that even arguably supports the claim
`
`element that the PCEs “operate[] independently.” Paper 2 at 10-11; Ex. 1015;
`
`Ex. 1016 at 32-33; Ex. 1017 at 2; Ex. 1018 at 8; Ex. 1021 at 2. IV and
`
`Dr. Richardson performed their analysis as of the original January 1999
`
`application, even though that application does not support the claims of the ’271
`
`patent. Paper 13 at 3; Ex. 2001 ¶ 22; Ex. 1021 at 5 (disclosure of separate history
`
`buffers in January 2002 application allegedly support “independently”); see
`
`generally Ex. 1028 (no disclosure of separate history buffers). IV’s attempt to cure
`
`this deficiency in a new footnote is ineffective. See Ex. 2001 ¶ 11.
`
`IV. GLOVER IN COMBINATION WITH CHU DISCLOSES PCES THAT
`OPERATE INDEPENDENTLY (CLAIMS 8, 32-34), AND GLOVER
`IN COMBINATION WITH RANGANATHAN DISCLOSES PDES
`THAT OPERATE INDEPENDENTLY (CLAIMS 72, 110).
`
`This Panel previously rejected IV’s attempt to argue that Glover, in
`
`combination with Chu or Ranganathan, does not disclose PCEs and PDAs that
`
`operate independently. Paper 8 at 17-20. IV attempts another bite at the apple by
`
`14
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`giving a new and strained interpretation of the “purpose” of Glover in support of
`
`its argument. See Paper 13 at 36-52. This has no more merit than IV’s first
`
`attempt. IV’s new arguments that Glover does not disclose PCEs and PDEs that
`
`operate independently are similarly meritless and should be rejected.
`
`A. IV’s Purported “Purpose” Of Glover Is Not Supported By The
`Evidence.
`
`IV argues the purpose of Glover is “to ‘provide an apparatus and method for
`
`compressing video image data into bands and subbands and grouping the data
`
`whereby a dictionary based compression scheme may be utilized.’” Paper 13 at 42
`
`(quoting Ex. 1010 at 2:63-66). As an initial matter, this “purpose” is not at all
`
`inconsistent with IBM’s explanation that “[t]he purpose of Glover is to split the
`
`video signal into subbands so that the subbands can be processed in parallel.”
`
`Paper 2 at 27 (citing Ex. 1001 ¶¶ 135-36); Ex. 1025 ¶ 32. There is no dispute that
`
`Glover discloses splitting the video signal into subbands and processing the
`
`subbands in parallel compression engines.
`
`IV and Dr. Richardson seize upon Glover’s reference to “grouping the data”
`
`at column 2, lines 63-66 as a basis to recharacterize the purpose of Glover as
`
`providing similarity across the four subbands, and, specifically, to make subband 0
`
`more similar to subbands 1 through 3. See Paper 13 at 11, 13, 15, 42-47. But
`
`making the subbands more similar to each other is not a purpose of Glover. Ex.
`
`1025 ¶¶ 36-37. Glover discloses “grouping” the data within a subband, not across
`
`15
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`the subbands. Data output from a Run-Length coder, as used in each subband of
`
`Glover, is often referred to as “grouped” data because a Run-Length coder encodes
`
`runs of data values as a single data value and count. Ex. 1025 ¶ 34; Ex. 1027 at 11.
`
`Dr. Richardson admitted that Glover’s Run-Length coder “groups together runs of
`
`data” in a single subband, not across subbands. Ex. 1026 at 234:4-14.
`
`IV and Dr. Richardson further attempt to support their mischaracterization
`
`and redefinition of Glover’s purpose by citing to a statement in Glover that “there
`
`is enough similarity in the grouped signals to use a compression technique like the
`
`Lempel-Ziv technique.” Ex. 1010 at 5:31-33; see Paper 13 at 11-13. They ignore
`
`the text immediately preceding this statement, however, which says that each of
`
`the Run-Length coders “group the data into values thereby producing grouped
`
`signals at 120, 220, 320 and 420.” Ex. 1010 at 5:27-30. In other words, the Run-
`
`Length coder in each subband groups the data in that subband, as Run-Length
`
`coders do, to produce a grouped signal that can then be compressed by the LZ
`
`coder in each subband. Id. at 5:27-33, Fig. 1; Ex. 1025 ¶ 34. Similarly, Glover
`
`explains that Run Length decoders “ungroup” the information in each subband to
`
`produce “ungrouped subbands at 155, 255, 355, and 455.” Ex. 1010 at 6:16-18,
`
`Fig. 2; Ex. 1025 ¶ 35.
`
`IV also uses circular logic when it claims this purported “purpose” of Glover
`
`shows that the PCEs and PDEs in Glover would share resources. According to IV,
`
`16
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`a PHOSITA would understand that Glover “would use a single, shared lookup
`
`table for Huffman coding/decoding across the subbands and a single, shared
`
`dictionary for Lempel-Ziv coding/decoding across the subbands” because the
`
`subbands “each have a similar statistical profile.” Paper 13 at 46-47; Ex. 2001
`
`¶¶ 102-03, 122-23. But the reason that IV contends that Glover would have similar
`
`statistical profiles across subbands is because statistical and dictionary based
`
`compression “rely on having the largest data set possible with similar statistical
`
`behavior.” Paper 13 at 12; Ex. 2001 ¶ 33. In other words, IV presumes that the
`
`algorithms in Glover would have a large data set that shares resources to support
`
`its argument that Glover’s purpose is to produce similar statistical profiles across
`
`bands, and then argues that because Glover produces similar statistical profiles
`
`across bands it must share resources.
`
`IV’s argument that Glover does not disclose PCEs and PDEs that operate
`
`independently is based on its manufactured “purpose” of Glover. Because that
`
`manufactured “purpose” is not supported by the evidence, IV’s arguments based
`
`upon that “purpose” must fail.
`
`B. IV’s Argument That Glover Does Not Disclose PCEs And PDEs That
`Operate Independently Is Contradicted By The Evidence.
`
`IV once again focuses on details about a specific implementation of Glover
`
`instead of answering the controlling question of law as to what Glover discloses to
`
`17
`
`
`
`Petitioner’s Reply to Patent Owner Response: IPR2015-01089
`
`
`a PHOSITA in view of his or her background knowledge. See KSR, 550 U.S. 417-
`
`18.
`
`The PCEs and PDEs in Glover each have their own inputs and outputs, and
`
`neither the figures nor the text in Glover make any reference to the PCEs or PDEs
`
`sharing resources. Paper 2 at 27, 51; Ex. 1010 at 5:16-39, 6:4-25, Figs. 1 and 2;
`
`Ex. 1001 ¶¶ 138-40, 271-73. Independent processing of data streams was well
`
`known in the art. Ex. 1001 ¶¶ 140, 273; Ex. 1022 at 4:4-34; Paper 8 at 18-19. An
`
`implementation of Glover with PCEs and PDEs that operate independently would
`
`have minimal, if any, impact on the storage and circuit area required. Ex. 1025
`
`¶¶ 28-29. A PHOSITA would combine the disclosure of Glover with this
`
`background knowledge to conclude that the PCEs and PDEs in Glover operate
`
`independently. See Ex. 2017 at 100:7-14 (Glover’s Figure 1 discloses “four
`
`independent parallel compression engines, each of them implementing a parallel
`
`compression algorithm with the traditional pipelining architecture. It’s very
`
`efficient, it’s very cost effective, very, very well known.”), 170:5-171:16.
`
`Further supporting a PHOSITA’s conclusion that the PCEs and PDEs in
`
`Glover operate independently are the disclosures in Glover that the subbands are
`
`optimized for compression of that subband. Gl