Docket No.: 0100157-00268
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`____________________________________________
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`____________________________________________
`
`EMC Corporation
`Petitioner
`
`Lenovo (United States) Inc.
`Petitioner
`
`NetApp, Inc.
`Petitioner
`
`v.
`
`Intellectual Ventures I, LLC
`Patent Owner
`
`Case No. IPR2017-00429
`Patent No. 6,775,745
`
`
`PETITIONER’S SUPPLEMENTAL REPLY
`
`
`
`
`
`
`
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`____________________________________________
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`____________________________________________
`
`EMC Corporation
`Petitioner
`
`Lenovo (United States) Inc.
`Petitioner
`
`NetApp, Inc.
`Petitioner
`
`v.
`
`Intellectual Ventures I, LLC
`Patent Owner
`
`Case No. IPR2017-00429
`Patent No. 6,775,745
`
`
`PETITIONER’S SUPPLEMENTAL REPLY
`
`
`
`
`
`
`
`
`I.
`II.
`
`V.
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`
`TABLE OF CONTENTS
`
`i
`
`B.
`C.
`
`B.
`
`2.
`
`2.
`
`Introduction ...................................................................................................... 1
`Claim Construction .......................................................................................... 5
`A.
`“Frequency Factors Indicating How Often Each Of The
`Corresponding Files Are Requested By The Operating System” ......... 5
`“A Least Frequently And Least Recently Used File” ........................... 8
`“Scanning From A Frequency Factor Corresponding To A LRU
`File To A Frequency Factor Corresponding To A MRU File” ............. 9
`III. Ground 1: Burton/Karedla (claims 4-6) ....................................................... 15
`A.
`Claims 4 and 5 ..................................................................................... 15
`1.
`Burton Teaches the “Frequency Factor” Limitation
`Under the Broadest Reasonable Interpretation ......................... 15
`Burton Teaches the Limitation Even Under Patent
`Owner’s Construction ............................................................... 17
`Claim 6 ................................................................................................ 21
`1.
`Burton Teaches the “Scanning” Limitation Under the
`Broadest Reasonable Interpretation .......................................... 21
`Burton Teaches the Limitation Even Under Patent
`Owner’s Construction ............................................................... 21
`IV. Grounds 4-5 – Lee/Dharap/Karedla/Robinson/Sweeney .............................. 23
`A.
`Ground 4 (Claim 6) ............................................................................. 23
`B.
`Ground 5 (Claims 4 and 5) .................................................................. 24
`The Board did not Make a Factual Mistake in Instituting this IPR ............... 25
`
`
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`
`I.
`
`INTRODUCTION
`Patent Owner’s Supplemental Response (“POSR”) and the cross-
`
`examination of its expert confirm that claims 4-6 of the ʼ745 patent are
`
`unpatentable in view of newly instituted Grounds 1 and 4.1
`
`As an initial matter, the only true, new issue in this supplemental trial phase
`
`concerns claim 6. As explained below, claim 6 has never been tried and is clearly
`
`unpatentable in view of the record evidence that has developed during the first and
`
`supplemental trial phases of this IPR. Claims 4 and 5, in contrast, raise the same
`
`claim construction issues already tried (i) in this IPR in connection with Ground 3,
`
`and (ii) in IPR2016-01643 (where claim 4 was found unpatentable). If anything,
`
`
`
` 1
`
` Newly instituted Grounds 2, 5, and 6 include Sweeney in the obviousness
`
`combinations and were contingent on a construction of “an extended segment of
`
`data” as “large block of data.” Petition (“Pet.”), 20, 36, 62, 65. Patent Owner
`
`mooted these Grounds by stipulating that it will not contend that “an extended
`
`segment of data” connotes a “large block of data” under the broadest reasonable
`
`interpretation standard. Patent Owner Preliminary Response (“POPR”), 2-3.
`
`1
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`the supplemental phase of this IPR merely provides additional grounds for finding
`
`these claims unpatentable.2
`
`Thus, the issues in this phase of the case should be discrete and simple.
`
`Instead, however, Patent Owner has complicated the proceeding by attempting to
`
`reargue and rehabilitate arguments already tried in the initial phase of the case.3
`
`
`
` 2
`
` IPR2016-01643 found ʼ745 claims 1, 2, 4, 12 and 14 unpatentable over Karedla,
`
`Burton, and Otterness. Claim 4 was found anticipated by Karedla and obvious
`
`over Burton and Karedla. The Board did not consider claims 5 or 6. In this case,
`
`Patent Owner has stated that Petitioner’s challenge to claim 5 is an “insignificant”
`
`difference to IPR2016-01643’s challenge to claim 4, and that Patent Owner “is
`
`willing to tie the patentability of claim 5 to the patentability of claim 4 as
`
`determined in IPR2016-01643 and any appeals thereof.” POPR, 2.
`
`3 Indeed, Patent Owner represented to the Board that its supplemental response
`
`would not be “wide open supplemental briefing and evidence,” but instead would
`
`be “limited to the additional grounds and claims … [that] were not part of the
`
`trial previously.” Ex. 2020, 34:17-23; Paper 51 (Revised Sch. Order), 2 (allowing
`
`parties “to provide briefing on the newly instituted claim and grounds”) (all
`
`emphases used herein is added unless otherwise noted). Despite this, and for
`
`2
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`For claims 4 and 5, Patent Owner’s only responses to newly instituted
`
`Grounds 1 and 4 are premised on the same, fundamentally flawed claim
`
`construction for “frequency factor” that Patent Owner already argued in connection
`
`with Ground 3. POSR, 13-25; Patent Owner’s Response (“POR”), 7-25. As
`
`explained in the trial phase for Ground 3, Patent Owner’s proposed construction is
`
`inconsistent with the claim language and would exclude disclosed embodiments.
`
`Ex. 1026, 31:12-32:5, 44:15-46:6, 67:13-68:15, 72:12-73:3 (Patent Owner’s
`
`expert, Dr. Chong, admitting that its proposed construction excludes
`
`embodiments); see also EPOS Techs. Ltd. v. Pegasus Techs. Ltd., 766 F.3d 1338,
`
`1347 (Fed. Cir. 2014) (rejecting construction “because it reads out preferred
`
`
`
`reasons it never explains, Patent Owner does the opposite of what it represented to
`
`the Board. It spends the bulk of its POSR re-arguing its proposed constructions of
`
`“frequency factor” and “a least frequently and least recently used file, that Lee
`
`does not disclose the “frequency factor” and “a least frequently and least recently
`
`used file” limitations under Patent Owner’s erroneous constructions, and that the
`
`Board should exercise its discretion to terminate this IPR. Petitioner demonstrated
`
`that these arguments were wrong during the first trial. They remain wrong now
`
`and also should be rejected as improper and untimely.
`
`3
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`embodiments”). Moreover, the same proposed construction was already
`
`considered and rejected in IPR2016-01643. As in IPR2016-01643, Patent Owner
`
`provides no response to the Petition regarding claim 4 that does not rely on its
`
`erroneous construction. Thus, if the Board rejects Patent Owner’s construction, as
`
`it should, claims 4 and 5 are unpatentable. Regardless, even if the Board were to
`
`accept Patent Owner’s construction, the claims would still be unpatentable as
`
`explained below, in the Petition, and in the prior trial on Ground 3.
`
`As referenced above, claim 6 raises important issues not yet considered by
`
`the Board. Patent Owner premises its argument that claim 6 is patentable on
`
`another fundamentally flawed claim construction. Specifically, Patent Owner
`
`proposes construing claim 6’s “scanning” limitation as requiring scanning every
`
`file in the cache including the LRU file, the MRU file, and every file in between.
`
`Like Patent Owner’s constructions for claim 4, this construction is flatly
`
`inconsistent with the claim language and the specification. Indeed, Dr. Chong
`
`acknowledged that the patent includes embodiments which do not scan the MRU,
`
`and that the scanning could end as early as the LRU. See POR, 26 (arguing that
`
`the MRU file cannot be “a least frequency and least recently used file”); Ex. 2004,
`
`¶ 65 (confirming that MRU may not be considered during scanning (citing Ex.
`
`1001, 6:39-41)); Ex. 1027, 21:23-22:17. The broadest reasonable interpretation of
`
`the “scanning” limitation encompasses these embodiments by requiring a specific
`
`4
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`direction for the scanning—from LRU to MRU—but neither requiring nor
`
`precluding scanning every file in the cache. As explained below, both Burton
`
`(Ground 1) and Robinson (Ground 4) disclose this limitation, when properly
`
`construed.
`
`Moreover, Burton discloses this limitation even under Patent Owner’s
`
`construction. Specifically, Burton discloses scanning every file in the cache,
`
`including the LRU file, the MRU file, and every file in between, for certain cache
`
`configurations which Patent Owner’s expert admits are legitimate.
`
`II. CLAIM CONSTRUCTION
`A.
`“Frequency Factors Indicating How Often Each Of The
`Corresponding Files Are Requested By The Operating System”
`The parties addressed this term in their prior trial on Ground 3. For the
`
`reasons set out in the Petition and Reply, Patent Owner’s proposed construction—
`
`“frequency factors showing, with a fair degree of certainty, how often each of the
`
`corresponding files are requested by the operating system”—is inconsistent with
`
`the specification, would exclude preferred embodiments, and should be rejected (as
`
`it was in IPR2016-01643).
`
`Realizing the inconsistencies in its prior arguments, Patent Owner now
`
`concedes that the ʼ745 specification discloses “broader embodiments” than its
`
`proposed construction would allow. Patent Owner argues at length, however, that
`
`claim 4 excludes these “broader embodiments” because of the claim language
`
`5
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`requiring that the frequency factors “indicate how often each file is requested.”
`
`See POSR, 10-13. According to Patent Owner, “embodiments in which disclosed
`
`‘frequency factors’ are heavily weighted” cannot meet claim 4’s claim language.
`
`Id., 10.
`
`Patent Owner’s new arguments are untimely and contrary to Patent Owner’s
`
`representation to the Board and the Board’s Order, and should be stricken for that
`
`reason alone. Moreover, they are equally flawed on the merits. The specification
`
`uses the same term—“frequency factor”—to describe every embodiment, even
`
`those that are weighted using criteria other than frequency of use or access. See
`
`Ex. 1001, 5:65-6:4; 6:56-7:15; 8:65-9:3. Patent Owner’s proposed construction is
`
`clearly not reasonable (and certainly not the broadest reasonable interpretation) as
`
`it is premised on the same phrase “frequency factor” shifting its meaning from one
`
`sentence to the next in the specification. Moreover, as the Federal Circuit has
`
`repeatedly stated, a construction that excludes embodiments “is rarely, if ever
`
`correct.” EPOS, 766 F.3d at 1347. This is particularly true here where the
`
`embodiments that Patent Owner’s construction attempts to exclude are included in
`
`what the ʼ745 specification describes as “self-tuning” and explicitly claimed in
`
`dependent claim 7 (which depends from claim 4). See e.g., Ex. 1026, 44:15-45:8
`
`(conceding that Patent Owner’s construction “applies to some of the self-tuning
`
`embodiments but not necessarily all”). These self-tuning embodiments include
`
`6
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`frequency factors that do not calculate how often or even count how many times a
`
`file has been accessed, as Dr. Chong conceded. Ex. 1026, 20:14-21:4. For
`
`example, the frequency factors can be weighted “heavier, i.e., the frequency factor
`
`would be incremented by a factor of more than one for each time the [file] is
`
`accessed.” See Ex. 1001, 8:65-9:3. The self-tuning embodiments also disclose
`
`adjusting the frequency factors based on criteria other than frequency of use,
`
`including decrementing the factors over time, or adjusting factors in only a portion
`
`of the cache. See id., 6:56-7:51. The same language in claim 4 cannot
`
`simultaneously support excluding (as Patent Owner argues) and including (as
`
`claim 7 demands) the same “self-tuning” embodiments. See Tr. of Columbia Univ.
`
`v. Symantec Corp., 811 F.3d 1359, 1370 (Fed. Cir. 2016) (“construing the
`
`independent claim to exclude material covered by the dependent claim would be
`
`inconsistent”).4
`
`
`
` 4
`
` Patent Owner’s construction is also entirely subjective, as its own expert
`
`admitted. Ex. 1026, 28:14-29:9, 35:1-10; 40:6-17; 38:18-21. The inability of an
`
`expert to use his own proposed parameters “militates against the reasonableness of
`
`those parameters as delineating the metes and bounds of the invention.” Datamize,
`
`LLC v. Plumtree Software, Inc., 417 F.3d 1342, 1354 (Fed. Cir. 2005).
`
`7
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`Accordingly, for all the reasons already tried to the Board, Patent Owner’s
`
`proposed construction should be rejected.
`
`B.
`“A Least Frequently And Least Recently Used File”
`The parties also addressed this term in their prior trial on Ground 3. For the
`
`reasons set out in the Petition and Reply, Patent Owner’s proposed construction—
`
`requiring that this file “cannot be the most recently used file”—is inconsistent with
`
`the specification, would exclude preferred embodiments, and should be rejected.
`
`Despite characterizing it as “additional,” Patent Owner’s argument merely
`
`rehashes its prior argument that “every embodiment disclosed in the specification”
`
`purportedly excludes the MRU file from being the “least frequently and least
`
`recently used file.” POSR, 3-4. However, Patent Owner’s expert disagreed during
`
`cross-examination in the trial on Ground 3. Ex. 1026, 103:19-104:6 (“Q. …
`
`There’s at least one embodiment, you would concede, that the MRU is an eviction
`
`candidate. A. Yes.”).
`
`The broadest reasonable interpretation of “a least frequently and least
`
`recently used file” considers a balance between both frequency and recency. The
`
`specification discloses embodiments which identify and evict files which are not
`
`both the LFU and the LRU, at least one which identifies and evicts a file which is
`
`neither, and one in which the MRU can be evicted. POR, 29 (citing Ex. 1001,
`
`Figs. 2A-2B); Ex. 1026, 103:19-104:6. The broadest reasonable interpretation of
`
`8
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`the claims should, at a minimum, be broad enough to include these embodiments.
`
`EPOS, 766 F.3d at 1347.
`
`Accordingly, for all the reasons already tried to the Board, Patent Owner’s
`
`proposed construction should be rejected.
`
`C.
`
`“Scanning From A Frequency Factor Corresponding To A LRU
`File To A Frequency Factor Corresponding To A MRU File”
`Patent Owner’s proposed construction for this term—that “the scanning
`
`process must scan all files from the LRU file to the MRU file, inclusive” (POSR, 5
`
`(emphasis in original))—has the same flaws as its other proposed constructions.
`
`Once again, it is inconsistent with the specification, inconsistent with the claim
`
`language, inconsistent with Patent Owner’s other constructions, would exclude
`
`preferred embodiments and is inconsistent with Dr. Chong’s cross-examination.
`
`This conclusion is supported by the declaration of Dr. John D. Kubiatowicz, filed
`
`herewith as Ex. 1028. Ex. 1028, ¶ 88.
`
`The broadest reasonable interpretation of the “scanning” limitation does not
`
`require scanning all of the files in the cache, including the LRU file, the MRU file,
`
`and all files in between, as the POSR insists. To the contrary, claim 6 adds a
`
`directional limitation, i.e., that the scanning take place in a direction from the LRU
`
`to the MRU, as indicated by the scanning arrows in Figs 2A-C. Indeed, Patent
`
`Owner’s expert conceded there are embodiments which do not scan the MRU, and
`
`that the scanning could end as early as the LRU. Ex. 2004, ¶ 65 (confirming that
`
`9
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`MRU may not be considered during scanning (citing Ex. 1001, 6:39-41)); Ex.
`
`1027, 21:23-22:17. Moreover, claim 15 explicitly confirms the existence of
`
`embodiments which scan “from . . . a LRU file to . . . a MRU file” and exclude the
`
`MRU file. See Ex. 1001, claim 15; id., 6:39-41, 7:3-6. And finally, Patent
`
`Owner’s construction directly contradicts its argument that claim 4 (from which
`
`claim 6 depends) cannot include the MRU file as the “least frequently and least
`
`recently used file.” See, e.g., Ex. 1026, 99:13-100:15. Read together, Patent
`
`Owner’s constructions illogically require that the frequency factor for the MRU
`
`file must be scanned even though it cannot be a candidate for eviction. See POSR,
`
`5. Ex. 1028, ¶ 89-90.
`
`The broadest reasonable interpretation should include the disclosed
`
`embodiments in which some files in the cache are not scanned, so long as the
`
`scanning direction proceeds from the LRU file towards the MRU file, rather than
`
`in the opposite direction.
`
`1.
`
`Patent Owner’s Construction Is Inconsistent With the
`Specification
`The parties agree that claim 6 narrows the scope of claim 4 by requiring a
`
`specific direction (from the LRU to the MRU) when scanning the frequency
`
`factors. However, Patent Owner attempts to narrow claim 6 even further by citing
`
`embodiments in the specification in which “the MRU file acts as a stop for the
`
`scanning process.” POSR, 6 (citing Ex. 1001, 6:49-50). While these embodiments
`
`10
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`are certainly covered by claim 6, Patent Owner’s attempt to limit the claim to these
`
`embodiments is improper, and would necessarily exclude other disclosed
`
`embodiments reasonably within the claim’s scope. Ex. 1028, ¶ 91.
`
`Patent Owner’s construction ignores embodiments in which its own expert
`
`acknowledged that the MRU is “simply not considered” when scanning the
`
`frequency factors. Ex. 2004, ¶ 65 (citing Ex. 1001, 6:39-41); see also Ex. 1001,
`
`claim 15, 7:3-6. Indeed, as described below, Patent Owner argued in the trial
`
`phase on Ground 3 that “every embodiment disclosed in the specification”
`
`excludes the MRU file from being the “least frequently and least recently used
`
`file.” POSR, 3-4; see Section II.C.3, below. Patent Owner thus contends that no
`
`embodiment discloses considering the MRU file for eviction, but that claim 6
`
`nevertheless demands that the MRU file must always be scanned. This
`
`inconsistency demonstrates Patent Owner’s views are unreasonable and not the
`
`broadest reasonable interpretation. Ex. 1028, ¶ 92.
`
`Patent Owner’s self-contradictory logic similarly excludes the self-tuning
`
`embodiments. As described above, these embodiments include “features which
`
`may enhance the tuning,” such as “how far forward to search from the LRU file,
`
`…” Ex. 1001, 7:3-6. This language confirms that a user may customize which
`
`files are scanned for potential eviction, and clearly shows that scanning from the
`
`11
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`LRU may proceed by a specified amount (“how far forward”) and need not include
`
`the MRU. Id., 7:1-12.5 Ex. 1028, ¶ 93.
`
`Claim 6’s language does not require excluding these embodiments. To the
`
`contrary, claim 6 specifies the direction for the scanning—“from the LRU file to
`
`the MRU file”—but contains no requirement regarding “how far forward to search
`
`from the LRU file.” Id., 7:3-6. Ex. 1028, ¶ 94.
`
`2.
`
`Patent Owner’s Construction Is Inconsistent with the Claim
`Language
`The ʼ745 claim language similarly confirms that claim 6 does not require
`
`scanning all files in the cache. First, the claims show that, when the applicants
`
`wanted to include “all files” in the scan, the claims said so explicitly. See Ex.
`
`1001, claim 8 (“scanning the frequency factor for each of the files, …”); compare
`
`ResQNet, Inc. v. Lansa, Inc., 346 F.3d 1374, 1377, 1382 (Fed. Cir. 2003) (“each
`
`field” referred to “all fields”). Unlike claim 8, claim 6 does not use “each of”
`
`
`
` 5
`
` Claim 4 is broader than claim 6 in that it includes scanning in both directions—
`
`“back … towards the LRU file” from the MRU, and “forward … from the LRU
`
`file” to the MRU. Ex. 1001, 7:1-12.
`
`12
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`when referring to scanning the frequency factors, confirming that claim 6 does not
`
`require “each of” the frequency factors to be scanned.6 Ex. 1028, ¶ 95.
`
`Second, claim 15 explicitly recites that “the frequency factors are scanned
`
`from a frequency factor for a LRU file to a frequency factor for a MRU file, not
`
`including the frequency factor for the MRU file.” POSR, 6 (emphasis in
`
`original); see also Ex. 2004, ¶ 65. Twisting logic and law, Patent Owner argues
`
`that “by omitting the emphasized language, claim 6 requires scanning all
`
`frequency factors … including the MRU file.” Id. To the contrary, the
`
`emphasized language from claim 15 only confirms that claim 6 is broadly referring
`
`to a directional limitation, without any additional limitation on which files are
`
`included or excluded from the scan. In other words, claim 6 neither requires nor
`
`precludes scanning all files in the cache. Ex. 1028, ¶ 96.
`
`3.
`
`Patent Owner’s Construction is Inconsistent with its
`Proposed Constructions for Claim 4
`Finally, when read together, Patent Owner’s claim constructions for claim
`
`4’s limitation of “a least frequently and least recently used file” and dependent
`
`claim 6’s scanning limitation are not only inconsistent, they are mutually
`
`
`
` 6
`
` Claims 1, 4, 8, 12 and 15 also use the phrase “each of” to claim that all files in the
`
`cache are assigned a frequency factor.
`
`13
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`exclusive. According to Patent Owner, the only requirement for claim 4’s
`
`limitation of identifying “a least frequently and least recently used file” is that the
`
`MRU file is not included. Ex. 1026, 99:13-100:15. At the same time, Patent
`
`Owner insists that dependent claim 6’s scanning limitation requires scanning all
`
`files in the cache, so that the MRU is included. POSR, 5. Patent Owner provides
`
`no reason to require scanning the frequency factor of the MRU, when according to
`
`Patent Owner, the MRU cannot be considered a candidate for eviction. Indeed,
`
`Dr. Chong admitted on cross-examination that, once the frequency factor for the
`
`least frequently and least recently used file is scanned, there is no need to continue
`
`scanning. Ex. 1027, 21:23-22:17. Ex. 1028, ¶ 97.
`
`To the contrary, claim 6 makes perfect sense under the broadest reasonable
`
`interpretation of these terms, which includes embodiments where (i) the MRU file
`
`is included in the scan and is a possible “least frequently and least recently used
`
`file,” and (ii) other embodiments where the MRU file is not included in the scan
`
`and is not a possible “least frequently and least recently used file.” Thus, a proper
`
`construction of the scanning limitation in claim 6 adds a directional requirement to
`
`the scanning in claim 4, but neither requires nor precludes scanning all files in the
`
`cache. Ex. 1028, ¶ 98.
`
`14
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`III. GROUND 1: BURTON/KAREDLA (CLAIMS 4-6)
`As referenced above, Patent Owner’s only response to newly instituted
`
`Ground 1 is premised on fundamentally flawed claim constructions. Patent Owner
`
`does not dispute that claims 4-6 are unpatentable under Petitioner’s proposed
`
`constructions. Moreover, Burton discloses the disputed limitations even under
`
`Patent Owner’s flawed constructions. Ex. 1028, ¶ 99.
`
`A. Claims 4 and 5
`1.
`Burton Teaches the “Frequency Factor” Limitation Under
`the Broadest Reasonable Interpretation
`Patent Owner does not even attempt to dispute that Burton teaches the
`
`“frequency factor” limitation if its proposed construction is rejected (as it was in
`
`IPR2016-01643). Ex. 1028, ¶ 100.
`
`As set out in the Petition, the ʼ745 patent describes a “hybrid” cache eviction
`
`policy that considers both “recency” of use and “frequency” of use in identifying
`
`candidates for elimination from the cache. Pet., 8 (quoting Ex. 1001, 2:31-53).
`
`Burton teaches the same core ideas as the patent including the “frequency factor”
`
`limitation. Although Burton refers to its factors as “LRU ranks,” Burton is clear
`
`that its LRU ranks are likewise a hybrid metric by indicating “both how frequently
`
`the entry is accessed and the last time of access.” Burton, 3:11-13; see also id.,
`
`2:18-22 (confirming LRU ranks indicate “a time when the entry was last accessed
`
`and the number of accesses relative to other cache entries”); 2:3-5 (“For each
`
`15
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`entry in cache, a variable indicates both a time when the cache entry was last
`
`accessed and a frequency of accesses to the cache entry.”). Indeed, even
`
`Burton’s title discloses a “method . . . for demoting [i.e., evicting] data from cache
`
`based on least recently accessed and least frequently accessed data.” Id., title.7
`
`Ex. 1028, ¶ 101.
`
`As in the ʼ745 patent, Burton’s LRU ranks are “increased whenever the
`
`cache entry is accessed,” and “used in determining which entries to demote [i.e.,
`
`evict] from the cache to make room for subsequently accessed data.” Id., 2:16-17;
`
`2:22-23; compare Ex. 1001, 5:65-6:4, 6:47-55. Specifically, in Burton “those
`
`entries having the lowest LRU rank [are] marked for demotion [and] are both the
`
`least recently accessed and among those entries least recently accessed recently,
`
`are less frequently accessed.” Burton, 4:19-22; 2:24-35; title (“demoting data
`
`from cache based on least recently accessed and least frequently accessed data”).
`
`Ex. 1028, ¶ 102.
`
`
`
` 7
`
` During cross-examination, Dr. Chong either refused or was unable to reconcile
`
`his opinions with these explicit disclosures from Burton, providing grounds to
`
`assign little or no weight to his opinions on Burton. Ex. 1027, 53:12-72:18.
`
`16
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`Indeed, Patent Owner admits that in Burton, as in the ʼ745 patent, “[e]ach
`
`time a track is requested,” “the LRU rank for the most recently accessed cache
`
`entry is incremented.” POSR, 16 (discussing Burton, Fig. 2); see also Ex. 1027,
`
`91:7-15; Ex. 1001, 5:65-6:4. Patent Owner also concedes that Burton’s LRU rank
`
`takes frequency into account such that, between two files that “were last accessed
`
`at about the same time” (i.e., having the same recency), “the more frequently
`
`accessed entry will have a greater weighting” and thus a higher LRU rank than the
`
`less frequently accessed entry. POSR, 14-15 (citing Burton, 3:62-4:2); id. 15
`
`(conceding that in Burton “frequency of use serves as tiebreaker” and “plays a
`
`meaningful role” in at least some circumstances (emphasis in original)). Ex. 1028,
`
`¶ 103.
`
`2.
`
`Burton Teaches the Limitation Even Under Patent Owner’s
`Construction
`Patent Owner attempts to distinguish Burton based on an improper
`
`hypothetical access pattern used in conjunction with its claim construction
`
`requiring that the “frequency factor” indicate frequency “with a fair degree of
`
`certainty.” POSR, 13-25. But Burton discloses the frequency factor limitation
`
`even under this construction, because Burton’s LRU rank indicates how often a file
`
`is accessed with more certainty than the embodiments disclosed in the ʼ745 patent.
`
`Ex. 1028, ¶ 104.
`
`17
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`Dr. Chong’s analysis and testimony confirms this point. For example, Dr.
`
`Chong admitted that the frequency factors in the ʼ745 specification could deviate
`
`considerably, that the “ʼ745 patent is not specific” on how much considerations
`
`other than frequency had affected the frequency factors shown in Fig. 2A, and that
`
`he could not testify with any certainty “how often” the corresponding files in the
`
`Fig. 2A embodiment had been accessed. Ex. 1026, 53:1-22; 83:10-84:20.
`
`Burton’s LRU rank, in contrast, does indicate how often files have been accessed.
`
`Ex. 1028, ¶ 105.
`
`Analyzing Dr. Chong’s own proposed data access pattern—A, A, A, B, B,
`
`C, D, E, F, C—after initialization (for example when Burton’s time counter value
`
`is 100), confirms that Burton’s LRU ranks actually indicate frequency with a fair
`
`degree of certainty:
`
`Entry
`
`A
`
`B
`
`C
`
`D
`
`E
`
`F
`
`Ex. 1028, ¶ 107.
`
`Access Count
`
`3
`
`2
`
`2
`
`1
`
`1
`
`1
`
`LRU Rank
`303
`207
`214
`106
`107
`108
`
`18
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`These entries are proportionally weighted with every additional access, so
`
`entries D, E, and F with one access each have a value in the hundreds, entries B
`
`and C with two accesses have a value in the two hundreds, and entry A with three
`
`accesses has a value in the three hundreds. These results are exactly what one of
`
`ordinary skill would expect, and consistent with Burton’s express disclosure that
`
`its LRU ranks indicate frequency as well as recency: Entry A was accessed the
`
`most frequently (three times) and the least recently, and has the highest LRU rank,
`
`while entry D (accessed the least frequently and least recently of the least
`
`frequently accessed files) has the lowest LRU rank. Compare with POSR, 21
`
`(concluding that “track A has the lowest LRU rank even though it has been
`
`accessed more than any other track”). Similarly, Entries B and C were accessed at
`
`the same level of frequency (twice), but C has the higher LRU rank because it was
`
`accessed more recently. Entries D, E, and F were also accessed at the same level
`
`of frequency (once), and their respective LRU ranks correspond to their recency
`
`(e.g., F, being the most recently accessed, has the highest LRU rank). During
`
`cross-examination, Dr. Chong conceded that he “would not be surprised” by such a
`
`result. Ex. 1027, 105:17-20. Ex. 1028, ¶ 108.
`
`Dr. Chong comes to a different conclusion because he purports to apply
`
`Burton’s algorithm immediately after a cache first initializes. POSR, 17-21. But
`
`analyzing Burton’s LRU rank immediately after initialization, as Dr. Chong
`
`19
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`proposes, is not a reliable hypothetical because the ʼ745 patent and claims focus on
`
`identifying candidates for elimination, many cycles later, when the cache is full.
`
`See Ex. 1027, 116:15-18 (agreeing that “eviction is not needed until many cycles
`
`later after initialization, when the cache is full”). Ex. 1028, ¶ 106.
`
`Burton’s strong correlation between its variable “LRU rank” and frequency
`
`of access during the relevant time period (i.e., after initialization when the cache is
`
`full and eviction is needed) shows that Burton discloses the claimed frequency
`
`factor even under Patent Owner’s narrow construction, to at least the same degree
`
`as the ʼ745 patent itself.8 Ex. 1028, ¶ 111-118.
`
`
`
` 8
`
` Dr. Chong further argued that because “an LRU rank of 210 could correspond to
`
`any number of file accesses,” the LRU rank “does not provide an indication, or
`
`even a rough estimate, of how often a file is accessed” POSR, 21-24. However,
`
`during cross-examination Dr. Chong agreed that under the weighting embodiment
`
`of the ’745 patent, the same frequency factor could correspond to different
`
`numbers of file accesses (Ex. 1027, 87:14-88:25), which confirms that Burton’s
`
`LRU rank teaches the “frequency factor” of the ’745 patent.
`
`20
`
`
`
`IPR2017-00429
`U.S. Patent No. 6,775,745
`
`B. Claim 6
`1.
`Burton Teaches the “Scanning” Limitation Under the
`Broadest Reasonable Interpretation
`The parties agree that the “scanning” limitation requires scanning in the
`
`direction from the LRU to the MRU, and there is no dispute that Burton teaches
`
`scanning in that direction. Pet., 32-33. Burton is clear that, “[w]hen a track or
`
`page of data is staged into [the] cache …, a new entry [] is added to the top of the
`
`LRU linked list [],” i.e., the top corresponding to a MRU entry. Burton, 2:58-60.
`
`Similarly, cache entries “whose pointers are at the bottom of the list, represent[]
`
`[entries] that were accessed the longest time ago relative to other tracks in cache.”
`
`Id., 1:47-49. Accordingly, a person of ordinary skill would understand that
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
