`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_______________
`
`ORACLE AMERICA, INC., HP INC., HEWLETT PACKARD
`ENTERPRISE CO., AND HP ENTERPRISE SERVICES, LLC,
`Petitioner,
`
`v.
`
`REALTIME DATA LLC,
`Patent Owner.
`_______________
`
`Case IPR2016-00374
`Patent 8,643,513 B2
`_______________
`
`Before GEORGIANNA W. BRADEN, J. JOHN LEE, AND JASON J.
`CHUNG, Administrative Patent Judges.
`
`BRADEN, Administrative Patent Judge.
`
`DECISION
`Institution of Inter Partes Review
`37 C.F.R. § 42.108
`
`
`
`
`
`
`
`Teradata, Exh. 1025, p. 1 of 28
`
`
`
`I.
`
`INTRODUCTION
`
`A. Background
`Oracle America, Inc. (“Petitioner”) filed a Petition (Paper 2, “Pet.”) to
`institute an inter partes review of claims 1, 2, 4, 6, 11–16, 18–20, and 22 of
`U.S. Patent No. 8,643,513 B2 (Ex. 1001, “the ’513 patent”). Realtime Data
`LLC, (“Patent Owner”) timely filed a Preliminary Response (Paper 6,
`“Prelim. Resp.”). We have jurisdiction under 35 U.S.C. § 314(a), which
`provides that an inter partes review may not be instituted “unless . . . there is
`a reasonable likelihood that the petitioner would prevail with respect to at
`least 1 of the claims challenged in the petition.”
`Upon consideration of the Petition, the Petition’s supporting evidence,
`and Patent Owner’s Preliminary Response, we conclude Petitioner has
`established a reasonable likelihood it would prevail with respect to at least
`one of the challenged claims. Accordingly, for the reasons that follow, we
`institute an inter partes review.
`B. Related Proceedings
`Petitioner informs us of the following co-pending litigation matters
`that would affect or could be affected by a decision in this proceeding:
`Realtime Data LLC v Actian Corporation et al., E.D. Tex. Case No. 6:2015-
`cv-00463, Realtime Data LLC v Dropbox, Inc., E.D. Tex. Case No. 6:2015-
`cv-00465, Realtime Data LLC v EchoStart Corporation et al., E.D. Tex.
`Case No. 6:2015-cv-00466, Realtime Data LLC v Oracle America, Inc.,
`Hewlett-Packard Co. and HP Enterprise Services, LLC, E.D. Tex. Case No.
`6:2015-cv-00467, Realtime Data LLC v Riverbed Technology, Inc. et al.,
`E.D. Tex. Case No. 6:2015-cv-00468, Realtime Data LLC v SAP America,
`Inc. et al., E.D. Tex. Case No. 6:2015-cv-00469, Realtime Data LLC v
`
`Teradata, Exh. 1025, p. 2 of 28
`
`
`
`Teradata Corporation et al., E.D. Tex. Case No. 6:2015-cv-00470, all filed
`on May 8, 2015, and still pending currently. Pet. 3.
`Petitioner also informs us of concurrently filed IPR2016-00373
`(challenging U.S. Patent No. 7,378,992); IPR2016-00375 (challenging U.S.
`Patent No. 7,415,530); IPR2016-00376 (challenging U.S. Patent No.
`7,415,530); and IPR2016-00377 (challenging U.S. Patent No. 9,116,908).
`Id.
`
`C. The ’513 Patent
`The ’513 patent, titled “Data Compression Method and System,”
`discloses systems and methods for analyzing a data block and selecting a
`compression method to apply to that block. Ex. 1001, Title, Abst. The ’513
`patent further discloses “fast and efficient data compression using a
`combination of content independent data compression and content
`dependent data compression.” Id. at 3:55–58. One embodiment of the ’513
`patent is illustrated in Figure 13A reproduced below.
`
`As shown above in Figure 13A of the ’513 patent, the system receives an
`input data stream of data blocks. Id. at 15:63–16:5. Content dependent data
`recognition module 1300 analyzes the incoming data stream to recognize
`
`
`
`Teradata, Exh. 1025, p. 3 of 28
`
`
`
`“data types” and other parameters indicative of the “data type/content.” Id.
`at 16:15–21. If module 1300 recognizes the data type of a given data block,
`module 1300 routes the block to content dependent encoder module 1320
`(id. at 16:24–26); if not, it routes the block to “content independent” (or
`“default”) encoder module 30 (id. at.3:66–67, 4:30–35, 15:56–63, 16:26–27,
`18:17–25).
`Content dependent encoder module 1320 comprises lossy or lossless
`compression encoders (id. at 16:28–37); content independent encoder
`module 30 comprises only lossless encoders (id. at 16:43–50). Lossy
`encoders provide for an “inexact” representation of the original
`uncompressed data (id. at 2:4–7); lossless encoders provide for an “exact”
`representation (id. at 2:18–20). The ’513 patent teaches that “[e]ncoding
`techniques” may be selected “based upon their ability to effectively encode
`different types of input data.” Id. at 12:54–56.
`
`Another embodiment of the ’513 patent is illustrated in Figure 13B
`reproduced below.
`
`
`As shown above in Figure 13B of the ’513 patent, “compression ratio
`module 1340, operatively connected to the content dependent output
`
`Teradata, Exh. 1025, p. 4 of 28
`
`
`
`builder/counters 1330 and content independent buffer/counters 40
`determines the compression ratio obtained for each of the enabled encoders
`El . . . En.” Id. at 17:28–42. The compression ratio is set “by taking the
`ratio of the size of the input data block to the size of the output data block
`stored in the corresponding buffer/counters BCD1, BCD2, BCD3 . . . BCDm
`and/or BCE1, BCE2, BCE3 . . . BCEn.” Id. at 17:39–42
`
`D. Challenged Claims
`
`As noted above, Petitioner challenges claims 1, 2, 4, 6, 11–16, 18–20,
`and 22 of the ’513 patent, of which claims 1 and 15 are the only independent
`claims. Claims 1 and 15 are representative of the challenged claims and are
`reproduced below (with paragraphing):
`1. A method of compressing a plurality of data blocks,
`comprising:
`analyzing the plurality of data blocks to recognize when an
`appropriate content independent compression algorithm is to be
`applied to the plurality of data blocks;
`applying the appropriate content independent data compression
`algorithm to a portion of the plurality of data blocks to provide a
`compressed data portion;
`analyzing a data block from another portion of the plurality of
`data blocks for recognition of any characteristic, attribute, or
`parameter that is indicative of an appropriate content dependent
`algorithm to apply to the data block; and
`applying the appropriate content dependent data compression
`algorithm to the data block to provide a compressed data block
`when the characteristic, attribute, or parameter is identified,
`wherein the analyzing the plurality of data blocks to recognize
`when
`the appropriate content
`independent compression
`algorithm is to be applied excludes analyzing based only on a
`descriptor indicative of the any characteristic, attribute, or
`parameter, and
`
`Teradata, Exh. 1025, p. 5 of 28
`
`
`
`wherein the analyzing the data block to recognize the any
`characteristic, attribute, or parameter excludes analyzing based
`only on the descriptor.
`Ex. 1001, 26:21–46.
`15. A device for compressing data comprising:
`a first circuit configured to analyze a plurality of data blocks to
`recognize when an appropriate content independent compression
`algorithm is to be applied to the plurality of data blocks;
`a second circuit configured to apply the appropriate content
`independent data compression algorithm to a portion of the
`plurality of data blocks to provide a compressed data portion;
`a third circuit configured to analyze a data block from another
`portion of the plurality of data blocks for recognition of any
`characteristic, attribute, or parameter that is indicative of an
`appropriate content dependent algorithm to apply to the data
`block; and
`a fourth circuit configured to apply the appropriate content
`dependent data compression algorithm to the data block to
`provide a compressed data block when the any characteristic,
`attribute, or parameter is identified,
`wherein the first circuit is further configured to analyze the
`plurality of data blocks to recognize when the appropriate
`content independent compression algorithm is to be applied by
`excluding analyzing based only on a descriptor indicative of the
`any characteristic, attribute, or parameter, and
`wherein the third circuit is further configured to analyze the data
`block to recognize the any characteristic, attribute, or parameter
`by excluding analyzing based only on the descriptor.
`Id. at 27:32–28:19.
`
`E. The Evidence of Record
`Petitioner relies upon the following references, as well as the
`Declaration of Dr. James A. Storer (Ex. 1002):
`
`Teradata, Exh. 1025, p. 6 of 28
`
`
`
`Reference
`
`Patent/Printed Publication
`
`Hsu
`
`W. H. Hsu and A. E. Zwarico,
`“Automatic Synthesis of
`Compression Techniques for
`Heterogeneous
`Files,” Software—Practice and
`Experience, Vol. 25(10), 1097–
`1116 (October 1995)
`Franaszek US Patent No. 5,870,036
`
`Published/
`Issued Date
`Oct. 1995
`
`Exhibit
`
`1003
`
`Feb. 9, 1999
`
`1004
`
`F. The Asserted Grounds of Unpatentability
`Petitioner challenges the patentability of the ’513 patent based on the
`following grounds:
`References
`Hsu and Franaszek
`
`Basis
`§ 103
`
`Claims Challenged
`1, 2, 4, 6, 11–16, 18–20,
`and 22
`
`II. DISCUSSION
`
`A. Claim Construction
`In an inter partes review, claim terms in an unexpired patent are
`interpreted according to their broadest reasonable construction in light of the
`specification of the patent in which they appear. 37 C.F.R. § 42.100(b); see
`Cuozzo Speed Techs., LLC v. Lee, No. 15-446, 2016 WL 3369425, at *10–
`14 (U.S. June 20, 2016) (“We conclude that the regulation represents a
`reasonable exercise of the rulemaking authority that Congress delegated to
`the Patent Office.”). Under that standard, and absent any special definitions,
`we give claim terms their ordinary and customary meaning, as would be
`understood by one of ordinary skill in the art at the time of the invention. In
`re Translogic Tech., Inc., 504 F.3d 1249, 1257 (Fed. Cir. 2007).
`
`Teradata, Exh. 1025, p. 7 of 28
`
`
`
`Petitioner proposes constructions for “content independent
`compression algorithm” and “wherein the analyzing of the data within the
`data block to identify one or more data types excludes analyzing based only
`on a descriptor that is indicative of the data type of the data within the data
`block.” Pet. 13–16. At this stage of the proceeding, Patent Owner does not
`contest Petitioner’s proposed constructions. Prelim. Resp. 10–11. For
`purposes of this Decision and based on the record before us, we need not
`provide express constructions for any claim terms at this stage of the
`proceeding.
`B. Principles of Law
`A claim is unpatentable under 35 U.S.C. § 103(a) if the differences
`between the subject matter sought to be patented and the prior art are such
`that the subject matter as a whole would have been obvious at the time the
`invention was made to a person having ordinary skill in the art to which said
`subject matter pertains. KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 406
`(2007). The question of obviousness is resolved on the basis of underlying
`factual determinations, including: (1) the scope and content of the prior art;
`(2) any differences between the claimed subject matter and the prior art;
`(3) the level of skill in the art; and (4) objective evidence of nonobviousness,
`i.e., secondary considerations. See Graham v. John Deere Co., 383 U.S. 1,
`17–18 (1966).
`We analyze the challenges presented in the Petition in accordance
`with the above-stated principles.
`C. Level of Ordinary Skill in the Art
`In determining whether an invention would have been obvious at the
`time it was made, we consider the level of ordinary skill in the pertinent art
`
`Teradata, Exh. 1025, p. 8 of 28
`
`
`
`at the time of the invention. Graham, 383 U.S. at 17. “The importance of
`resolving the level of ordinary skill in the art lies in the necessity of
`maintaining objectivity in the obviousness inquiry.” Ryko Mfg. Co. v. Nu-
`Star, Inc., 950 F.2d 714, 718 (Fed. Cir. 1991).
`Petitioner’s Declarant, Dr. Storer, opines that a person of ordinary
`skill in the art relevant to the ’513 patent would have been a person with “an
`undergraduate degree in computer science and two years’ industry
`experience or a graduate degree in the field of computer science.” Ex. 1002
`¶ 20. Patent Owner does not offer any contrary explanation regarding who
`would qualify as a person of ordinary skill in the art relevant to the ’513
`patent. See generally Prelim. Resp.
`Based on our review of the ’513 patent, the types of problems and
`solutions described in the ’513 patent and cited prior art, and the testimony
`of Petitioner’s Declarant, we adopt and apply Dr. Storer’s definition of a
`person of ordinary skill in the art at the time of the claimed invention for
`purposes of this Decision. We also note that the applied prior art reflects the
`appropriate level of skill at the time of the claimed invention. See Okajima
`v. Bourdeau, 261 F.3d 1350, 1355 (Fed. Cir. 2001).
`D. Alleged Obviousness of Claims 1, 2, 4, 6, 11–16, 18–20, and 22 in
`View of Hsu and Franaszek
`Petitioner contends claims 1, 2, 4, 6, 11–16, 18–20, and 22 of the ’513
`patent are unpatentable under 35 U.S.C. § 103 in view of Hsu and
`Franaszek. Pet. 24–60. Patent Owner disputes Petitioner’s contention.
`Prelim. Resp. 11–49. For the reasons that follow, we determine Petitioner
`has demonstrated a reasonable likelihood of prevailing as to these claims.
`
`Teradata, Exh. 1025, p. 9 of 28
`
`
`
`1. Overview of Hsu
`Hsu is titled “Automatic Synthesis of Compression Techniques for
`Heterogeneous Files,” and discloses systems and methods for compressing
`“heterogeneous files”—files that contain “multiple types of data such as text,
`images, binary, audio, or animation.” Ex. 1003, Title, 1097. Hsu teaches a
`heterogeneous compressor that automatically chooses the best compression
`algorithm to use on a given variable-length block of a file, based on both the
`qualitative and quantitative properties of that segment, and “treats a file as a
`collection of fixed size blocks (5K in the current implementation), each
`containing a potentially different type of data and thus best compressed
`using different algorithms.” Id. at 1102. Hsu further teaches a two phase
`system. Id. In the first phase, the system uses statistical methods based on
`analysis of the data within each block to determine a data type of the block
`and to determine the optimal encoder to use in compressing a block. Id. at
`1097; see also id. at 1103 (“The compressibility of a block of data and the
`appropriate algorithm to do so are determined by the type of data contained
`in a block . . . .”). The second phase comprises the actual compression and
`an optimization that maximizes the size of a segment of data to be
`compressed using a particular algorithm. In this optimization, which is
`interleaved with the actual compression, adjacent blocks for which exactly
`the same method have been chosen are merged into a single block. Id. at
`1102. Specifically, during the second phase, the heterogeneous compressor
`applies the selected algorithms to the blocks separately. Id. at 1098.
`The approach in Hsu uses a program synthesis technique, meaning
`that a compression plan, consisting of instructions for each block of input
`data, is generated based on the statistical properties of the input data. Id.
`
`Teradata, Exh. 1025, p. 10 of 28
`
`
`
`According to Hsu, its system chooses each algorithm (as well as the duration
`of its applicability) before compression begins, rather than modifying the
`technique for each file during compression. Id. at 1100. The heterogeneous
`compressor in Hsu bases its compression upon statistics gathered from larger
`blocks of five kilobytes. Id. Hsu states that “[t]his allows us to handle much
`larger changes in file redundancy types. This makes our system less
`sensitive to residual statistical fluctuations from different parts of a file.” Id.
`Hsu further teaches that “in-depth statistical analysis in order to make a
`more informed selection from the database of algorithms” when looking at
`the blocks of data to be compressed, and such computations are considered
`for an entire block (as opposed to sporadic or random sampling from parts of
`each block). Id. at 1101.
`Hsu teaches that the compressibility of a block of data and the
`appropriate algorithm to do so are determined by the type of data contained
`in a block and the type of redundancy (if any) in the data. Id. at 1103.
`These two properties are represented by four parameters: the block type, and
`the three redundancy metrics. Id. The block type describes the nature of a
`segment of input data. Id. The redundancy metrics are quantitative
`measures that are used to determine the compressibility of a block of data.
`They are: the degree of variation in character frequency or alphabetic
`distribution, MAD; the average run length of the block, MRL; and the string
`repetition ratio of the block, MSR. Id. at 1104. According to Hsu, these three
`manifestations of redundancy are independent, and each of the redundancy
`types is exploited by different compression algorithms. Id.
`The compression algorithms and attendant heuristics of Hsu are
`organized into the 10 by 3 table shown Table 1, reproduced below.
`
`Teradata, Exh. 1025, p. 11 of 28
`
`
`
`
`As shown in Table 1, the 10 file descriptors are the row indices and the 3
`metrics are the column indices. Id. at 1106. Each entry of the table contains
`descriptors that are used to access the code for an algorithm-heuristic pair.
`Id. Hsu teaches the use of four basic compression algorithms to be used in
`its system: arithmetic coding, Lempel-Ziv, run length encoding (RLE), and
`JPEG for image/graphics compression. Id. An optimal algorithm is selected
`for each data block of a file, and the system creates a record of each data
`block and its optimal algorithm, which Hsu refers to as the file’s
`“compression plan.” Id. at 1109.
`2. Overview of Franaszek
`Franaszek teaches systems and methods for compressing and
`decompressing data blocks using a plurality of optimal encoders. Ex. 1004,
`Abst. Franaszek teaches that representative samples of each block are tested
`to select an appropriate encoder to apply to the block. Id. Franaszek teaches
`recognizing the data type of incoming data blocks and then compressing the
`collection of data blocks using a plurality of optimal encoders for the
`different types of data. Id. at 4:30–36, 5:49–53.
`
`Teradata, Exh. 1025, p. 12 of 28
`
`
`
`In one embodiment, Franaszek teaches a set of “default” compression
`algorithms, which are shown in Figure 2, reproduced below.
`
`
`Figure 2, above, illustrates data compressor 220 and data de-compressor
`270, with uncompressed data blocks 210 that can contain type information
`205. Id. at 4:25–31. According to Franaszek, the type information can be,
`for example, image data encoded in a given format, source code for a given
`programming language, etc. Id. at 4:32–34. Data blocks 210 are input to
`data compressor 220. Data compressor 220 and data de-compressor 270
`share a compression method table 240 and a memory 250 containing a
`number of dictionary blocks. Id. at 4:34–38. Compressor 220 selects a
`compression method to compress the data. Id. at 4:52–53. The compressor
`outputs compressed data blocks 230, with an index (M) 232 identifying the
`selected compression method. Id. at 4:55–57. De-compressor 270 de-
`compresses the block using the specified method found in compression
`method table 240 (using the compression method identifier as an index), and
`outputs uncompressed data blocks 280. Id. at 5:1–7. For example,
`
`Teradata, Exh. 1025, p. 13 of 28
`
`
`
`compression method table 240 is shown in Figure 2 implementing a Lempel-
`Ziv compression method.
`Figure 4A of Franaszek, reproduced below, shows the operation of
`data compressor 220 illustrated in Figure 2.
`
`
`As shown in Figure 4A, in step 401 when data compressor 220 receives an
`uncompressed data block, it first determines whether data “type”
`information (e.g., text, image, etc.) is available for the data block. Id. at
`5:49–50. If such information is available, then at step 404, the compression
`method list (CML) is set to a list of compression methods that have been
`preselected for that data type. Id. at 5:50–53. Otherwise, if no data type is
`available, in step 407 the CML is set to a default list of compression
`methods. Id. at 5:53–54. In instances when the data “type” information is
`
`Teradata, Exh. 1025, p. 14 of 28
`
`
`
`available, then data compressor 220 uses the compression method “table”
`240 shown in Figure 2. See id. at 5:49–53.
`3. Analysis
`a. Cited Art as Applied to Claims 1
`Petitioner contends that the combined teachings of Hsu and Franaszek
`would have rendered each limitation of claim 1 in the ’513 patent obvious to
`a person of ordinary skill in the art at the time of the invention. Pet. 28–43.
`Patent Owner disputes Petitioner’s contention. Prelim. Resp. 13–19, 38–39.
`Patent Owner specifically argues that Petitioner fails to show: (1) that Hsu is
`a printed publication; (2) that Hsu and Franaszek teach the combination of
`content dependent and content independent data compression; (3) that Hsu
`and Franaszek teach to “exclude[ ] analyzing based only on a
`descriptor . . . ,” as recited in challenged claim 1; and (4) that a person of
`ordinary skill in the art would have had a reason to combine Hsu and
`Franaszek. After considering the parties’ arguments and evidence, we
`determine that Petitioner presented sufficient evidence to establish a
`reasonable likelihood of prevailing on the ground of obviousness in view of
`Hsu and Franaszek. We address the issues disputed by Patent Owner in
`more detail.
`
`(1) Hsu as a Printed Publication
`Whether a document is a “printed publication” is “a legal
`determination based on underlying fact issues,” and involves a case-by-case
`inquiry into the facts and circumstances surrounding the reference’s
`disclosure to the public. In re Klopfenstein, 380 F.3d 1345, 1350 (Fed.
`Cir.2004); In re Hall, 781 F.2d 897, 899 (Fed. Cir. 1986). To qualify as a
`printed publication, a document must be available generally. Northern
`
`Teradata, Exh. 1025, p. 15 of 28
`
`
`
`Telecom, Inc. v. Datapoint Corp., 908 F.2d 931, 936 (Fed. Cir. 1990). A
`reference is publicly accessible upon a satisfactory showing that such
`document has been disseminated or otherwise made available to the extent
`that persons interested and ordinarily skilled in the subject matter or art,
`exercising reasonable diligence, can locate it. Kyocera Wireless Corp. v.
`Int’l Trade Comm’n, 545 F.3d 1340, 1350–51 (Fed.Cir. 2008).
`Patent Owner raises two general issues: (1) whether Hsu is
`admissible evidence; and (2) if admissible, whether Hsu, and the information
`presented on its face, is sufficient evidence of public accessibility. As to the
`first issue, admissibility of evidence in most cases is addressed more
`appropriately after trial is instituted and a motion to exclude is filed. As set
`forth in our Rules, all objections to evidence submitted during a preliminary
`proceeding must be filed within ten business days of the institution of trial.
`37 CFR § 42.64(b)(1). This allows the party relying on the evidence an
`opportunity to correct any deficiency by providing supplemental evidence.
`Id. Thus, we defer all questions as to the admissibility of Hsu until trial.1
`With respect to Patent Owner’s arguments about the sufficiency of
`Petitioner’s public accessibility evidence, we are persuaded that the
`information presented about Hsu demonstrates sufficiently for purposes of
`this Decision that Hsu was publicly accessible as of October 1995. On its
`face, Hsu appears to be an article published in the October 1995 issue of a
`journal titled “Software Practice & Experience.” Ex. 1003, 1. The cover of
`
`1 We note that Federal Rule of Evidence 803(16) provides that a document
`over twenty years old and whose authenticity is established is an exception
`to the hearsay rule, and that Rule 902(6) provides that newspapers and
`periodicals are self-authenticating documents.
`
`Teradata, Exh. 1025, p. 16 of 28
`
`
`
`Hsu appears to be stamped by a university library. Id. Hsu’s format and
`appearance are consistent with that of a published article in a technical
`journal. See generally Ex. 1003. At this stage of the case, we are satisfied
`these indicia are sufficient evidence to proceed. Therefore, on this record
`and for purposes of this Decision, we are satisfied Petitioner has made a
`threshold showing that Hsu is a “printed publication.”
`(2) “content independent data compression algorithm”
`and “content dependent data compression algorithm”
`Petitioner argues Hsu and Franaszek teach “applying the appropriate
`content independent data compression algorithm to a portion of the plurality
`of datablocks to provide a compressed data portion,” and “applying the
`appropriate content dependent data compression algorithm to the data block
`to provide a compressed data block” as required by challenged claim 1. Pet.
`34–36, 39–42. Petitioner argues Franaszek expressly teaches analyzing data
`blocks to determine when to apply either a content dependent compression
`algorithms (e.g., run-length or LZ1) or content independent compression
`algorithms (e.g., a dictionary-based algorithm). Id. at 31–32 (citing Ex.
`1004, Figs. 2, 4A). According to Petitioner, Franaszek teaches that content
`dependent compression occurs when the Franaszek system is able to identify
`a data block’s data type, each type of which is associated in the Franaszek
`system with one or more optimal compression algorithms. Id. (citing Ex.
`1004, 5:49–54). Petitioner explains that “[i]f the Franaszek system
`determines that a given block’s data ‘type’ information is not available, the
`system selects an appropriate content independent algorithm . . . , and then
`compresses that block with that algorithm.” Id. at 35 (citing Ex. 1004, 6:8–
`11 (“if no data type is available, a ‘default’ list of dictionary-based
`algorithms will be used, which have ‘block identifiers’ associated with
`
`Teradata, Exh. 1025, p. 17 of 28
`
`
`
`‘compression method M’”), 6:45–47 (“[data block] B is compressed using
`M”)).
`Petitioner further argues that Hsu teaches applying the appropriate
`content dependent data compression algorithm when the characteristics or
`attributes of the data (e.g., its “data type” and highest “redundancy metric”)
`that are indicative of an appropriate content dependent algorithm to apply
`are identified. Id. at 39–40 (citing Ex. 1003, 1102 (“[U]sing the block type
`and largest metric, the appropriate compression algorithms . . . are chosen
`. . . .[ ] The compression method for the current block is then recorded in a
`small array-based map of the file[.]”), 1109 (“An identifying tag for the
`selected algorithm is written to the ‘compression plan’, an array which stores
`one complete compression instruction per block (if the current block is
`deemed uncompressible, a ‘skip’ instruction is recorded).”).
`Patent Owner contests Petitioner’s position, arguing that Hsu and
`Franaszek fail to teach the limitation “applying the appropriate content
`independent data compression algorithm to a portion of the plurality of
`datablocks to provide a compressed data portion.” Prelim. Resp. 16–19.
`According to Patent Owner, Hsu’s lossless compression algorithms do not
`disclose content independent data compression. Id. at 18. Patent Owner
`argues that a lossless encoder may be for data for which a specific data type
`or content is identified or recognized, in which case it would not be content
`independent compression. Id. Patent Owner supports its arguments by
`citing to the ’513 patent description of how lossless compression algorithms,
`such as Lempel-Ziv, can be content dependent. Id. (citing Ex. 1001, 16:28–
`37).
`
`Teradata, Exh. 1025, p. 18 of 28
`
`
`
`At this stage of the proceeding, we are persuaded by Petitioner’s
`position that a person of skill in the art would have relied on Franaszek’s
`teachings regarding “content independent compression algorithms” in
`combination with the heterogeneous compressor system and content
`dependent compression algorithms in Hsu. See Pet. 34.
`(3) “wherein the analyzing the plurality of data blocks to
`recognize when the appropriate content independent
`compression algorithm is to be applied excludes
`analyzing based only on a descriptor indicative of the
`any characteristic, attribute, or parameter,” and
`“wherein the analyzing the data block to recognize
`the any characteristic, attribute, or parameter
`excludes analyzing based only on the descriptor”
`Claim 1 includes two wherein clauses, each limiting a prior analyzing
`step. Generally, each wherein clause limits the analyzing step to “exclude[]
`analyzing based only on a descriptor indicative of the any characteristic,
`attribute, or parameter.” Ex. 1001, 26:39–46.
`Petitioner argues the combination of Hsu and Franaszek teaches or at
`least suggests both wherein clauses recited in challenged claim 1 because the
`references teach “wherein the analyzing” of data blocks “excludes analyzing
`based only on a descriptor indicative of the any characteristic, attribute, or
`parameter.” Pet. 40–43. Petitioner argues that “Franaszek discloses
`analyzing a ‘type’ entry optionally appended to data blocks to determine
`whether content dependent or content independent compression algorithm
`is to be applied, and to determine the appropriate algorithm” (Ex. 1004,
`4:30–35, 6:1–2, Fig. 2), while “Hsu discloses directly analyzing the data
`within the data block (the first, middle, and last 512 bytes) to determine an
`appropriate content dependent compression algorithm to apply” (Ex. 1003,
`1104, 1007, Table I). Id. at 40–41.
`
`Teradata, Exh. 1025, p. 19 of 28
`
`
`
`Petitioner then argues that Hsu teaches analyzing data within the data
`block to identify the data’s type by analyzing something other than a
`descriptor, tag, or header appended to the block. Additionally, Hsu uses a
`procedure called “new-file” to determine the type of data in a file’s blocks
`by directly analyzing the first, middle, and last 512 bytes of the block,
`because such data are indicative of an appropriate content dependent
`algorithm to apply for compression. Id. at 42–43 (citing Ex. 1003, 1104;
`Ex. 1002 ¶ 122). Petitioner explains that the Hsu system then compares the
`pattern of data contained in those bytes to a collection of known data
`patterns for various different data types from Unix and other operating
`systems. Id. at 43(citing Ex. 1003, 1104). Petitioner further argues that Hsu
`also teaches analyzing data within the data block to determine the data’s
`type and degree of redundancy (i.e., compressibility) of the data within the
`data blocks. Id. (citing Ex. 1003, 1104).
`Patent Owner disputes Petitioner’s position, arguing that Hsu and
`Franaszek fail to teach “exclude[] analyzing based only on a descriptor
`indicative of the any characteristic, attribute, or parameter,” as recited
`generally in challenged claim 1. Prelim. Resp. 38–40. According to Patent
`Owner, based on Petitioner’s proffered construction of the disputed claim
`limitation,2 Petitioner must show that Hsu teaches analyzing two types of
`information: the data block and data appended to the data block. Id. at 39.
`Patent Owner specifically argues
`
`2 Petitioner proposes that the claim limitation “exclude[] analyzing based
`only on a descriptor indicative of the any characteristic, attribute, or
`parameter” be construed as “involves/involving analyzing data other than
`data appended to the data block.” Pet. 15.
`
`Teradata, Exh. 1025, p. 20 of 28
`
`
`
`taught,
`is
`Rather, without discussing where appension
`[Petitioner] alleges: “Hsu discloses analyzing information other
`than a data descriptor appended to the data block (e.g., sample
`bytes of data) to recognize characteristics or attributes of the
`data.” (Id., 42.) Then, [Petitioner] describes Hsu’s teaching that
`it analyzes the first, middle, and last 512 bytes. (Id., 42-43.) But
`[Petitioner] does not explain how analysis of the first, middle,
`and last 512 bytes is analysis of data other than data appended to
`the data block because [Petitioner] does not cite to any disclosure
`in Hsu of appended data that is outside of the data sampled. (Id.)
`
`Prelim. Resp. 39–40.
`Patent Owner, thus, concludes that absent an explanation regarding
`analysis of appended data, Petitioner has failed to meet its burden of
`showing that Hsu teaches its proposed claim construction of “analyzing data
`other than data appended to the data block.” Id. at 40.
`We are persuaded by Petitioner, at this stage o