`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`
`
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`
`
`
`
`
`
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`REALTIME DATA LLC,
`Patent Owner.
`
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2016-01365
`Patent 7,181,608
`
`
`
`
`
`
`
`
`
`
`PETITIONER’S REPLY TO PATENT OWNER’S RESPONSE
`
`
`
`
`I.
`II.
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`TABLE OF CONTENTS
`Introduction…...……………………………………………………………..1
`Claim Construction..………………………………………………………...1
`A. Preloading ................................................................................................. 1
` Realtime’s Preloading Construction Improperly Inserts
`Limitations into the Claims……………………………………2
` Realtime’s Preloading Construction is Inconsistent with
`Examples of Preloading in the ’608 Patent……………………3
` Third-Party Evidence Does Not Support Limiting
`Preloading to Realtime’s Construction ………………………..6
`III. The Applied Prior Art Renders the Challenged Claims Obvious………….11
`A. Sukegawa Describes Preloading (All Claims) ........................................ 11
` Realtime Wrongly Advances that Preloading Cannot
`Encompass Storage in Non-Volatile Memory ………..……...11
` Sukegawa’s Description of Preloading Matches the
`Approach Claimed and Described by the ’608 Patent………..14
`B. Sukegawa Describes Preloading Prior to Completion of Initialization
`(All Claims) ............................................................................................ 15
` Initializing and Preloading are not Limited to Being
`Performed During the Same Initialization Process…………...15
` Sukegawa Describes Preloading Prior to Initialization of the
`CPU…………………………………………………………...19
`C. Dye Describes a Plurality of Encoders (Claims 16-17, 20-21, 25-26, 30-
`31)……………………………………………………………………...20
`D. The Combination Applied in Ground 1 is Proper (All Claims) ............. 23
` Dye ’284 was Properly Incorporated by Reference into
` Dye……………………………………………………….23
` Dye (without Dye ’284) Provides Sufficient Motivation...25
` Grounds 2-4 Remain Unchallenged……………………...28
`IV. Conclusion………………………………………………………………….28
`
`
`
`
`
`
`
`i
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`EXHIBIT LIST
`
`APPLE-1001
`
`U.S. Patent No. 7,181,608 to Fallon, et al. (“the ’608 Patent”)
`
`APPLE-1002
`
`Excerpts from the Prosecution History of the ’608 Patent (“the
`Prosecution History”)
`
`APPLE-1003
`
`Declaration of Dr. Charles J. Neuhauser (“Dec.”)
`
`APPLE-1004
`
`Curriculum Vitae of Dr. Charles J. Neuhauser
`
`APPLE-1005
`
`U.S. Patent No. 5,860,083 (“Sukegawa”)
`
`APPLE-1006
`
`U.S. Patent No. 6,374,353 (“Settsu”)
`
`APPLE-1007
`
`Burrows et al., “On-line Data Compression in a Log-structured
`File System” (1992) (“Burrows”)
`
`APPLE-1008
`
`U.S. Patent No. 6,145,069 (“Dye”)
`
`APPLE-1009
`
`U.S. Patent No. 7,190,284 (“Dye ’284”)
`
`APPLE-1010
`
`Jeff Prosise, DOS 6 – The Ultimate Software Bundle?, PC
`MAGAZINE, Apr. 13, 1993 (“Prosise”)
`
`APPLE-1011
`
`Excerpts from John L. Hennessey & David A. Patterson,
`Computer Architecture a Quantitative Approach (1st ed. 1990)
`(“Hennessey”)
`
`APPLE-1012
`
`U.S. Patent No. 6,158,000 (“Collins”)
`
`APPLE-1013
`
`File, Microsoft Press Computer Dictionary (3d ed. 1997)
`
`APPLE-1014
`
`Excerpts from Tom Shanley & Don Anderson, PCI System
`Architecture, (4th ed. 1999) (“Shanley”)
`
`
`
`ii
`
`
`
`APPLE-1015
`
`APPLE-1016
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Jacob Ziv & Abraham Lempel, A Universal Algorithm for
`Sequential Data Compression, IT-23 No. 3 IEEE
`TRANSACTIONS ON INFORMATION THEORY 337 (1977)(“Ziv”)
`
`James A. Storer & Thomas G. Szymanski, Data Compression
`via Textual Substitution, 19 No. 4 JOURNAL OF THE
`ASSOCIATION FOR COMPUTING MACHINERY (1982)(“Storer”)
`
`APPLE-1017
`
`Transcript of June 20, 2017 Deposition of Dr. Back
`
`APPLE-1018
`
`U.S. Patent No. 6,658,492 (“Kawahara”)
`
`APPLE-1019
`
`U.S. Patent No. 5,815,718 (“Tock”)
`
`APPLE-1020
`
`U.S. Patent No. 6,463,509 (“Teoman”)
`
`APPLE-1021
`
`Encoder, Microsoft Press Computer Dictionary (5th ed. 2002)
`
`APPLE-1022
`
`Encoder, The Computer Desktop Encyclopedia (2nd ed. 1999)
`
`
`
`
`
`iii
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`
`I.
`
`Introduction
`Realtime’s arguments fail when longstanding legal principles are properly
`
`applied to the factual record. Indeed, a proper application of the broadest reasonable
`
`interpretation standard makes clear that Realtime overreached by pursuing overly
`
`broad claims in the ’608 Patent.
`
`II. Claim Construction
`In rendering its Institution Decision, the Board indicated that no express
`
`constructions were required. Institution Decision, 4. Despite this, and in an apparent
`
`attempt to manufacture patentability, Realtime now urges the Board to adopt an
`
`unduly narrow construction of “preloading” that improperly inserts limitations into
`
`the claims, and that is contradicted by examples within the specification, testimony
`
`of Realtime’s own expert, and third-party evidence. Patent Owner Response
`
`(“POR”), 12-16.
`
`A.
`Preloading
`Realtime proposes that “[t]he term ‘preloading,’ as used in claims 1, 3, 7-9,
`
`22, and 27, should mean transferring data from storage to memory in anticipation of
`
`immediate or near-in-time use.’” POR, 12. Realtime’s construction, however, is not
`
`“the broadest reasonable interpretation” because the construction improperly
`
`imports limitations that are missing from the ’608 Patent, excludes several examples
`
`of preloading provided in the ’608 Patent’s description, and is at odds with
`
`contemporaneous understanding of the term by persons of skill, as indicated by
`1
`
`
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`
`third-party evidence.
`
`
`
`Realtime’s Preloading Construction Improperly Inserts
`Limitations into the Claims
`Realtime’s “preloading” construction, which requires “transferring data from
`
`storage to memory in anticipation of immediate or near-in-time use,” improperly
`
`imports at least the “in anticipation of . . . use” and “immediate or near-in-time”
`
`limitations. These limitations are clearly absent from the claim language, and the
`
`specification nowhere describes preloading in these terms. Thus, importing the “in
`
`anticipation of . . . use” and “immediate or near-in-time” limitations into the
`
`construction of the term preloading is inconsistent with BRI.
`
`In an attempt to support its construction, Realtime asserts that “a POSITA
`
`would have understood that … movement of data from storage into memory is
`
`performed in anticipation of immediate or near-in-time use.” POR, 14. Realtime
`
`quotes the declaration of its expert, Dr. Back, ¶¶ 48-51. However, because the
`
`’608 Patent does not include the proposed language, Realtime provides no
`
`quotations to the ’608 Patent itself.
`
`And, although Dr. Back discusses portions of the ’608 Patent in ¶¶ 50-51,
`
`none of the portions cited by Dr. Back employ the terminology hand-picked by
`
`Realtime in an attempt to distinguish the prior art. See ’608, Abstract, 3:34-40,
`
`3:60-61, 21:50-54, 21:56-62, 22:9-19, 22:41-45, 32:13-38, FIGS. 7b, 8b. Indeed,
`
`the ’608 Patent nowhere describes a data transfer as being performed “in
`
`
`
`2
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`anticipation” of anything, let alone “immediate or near-in-time use.” Because the
`
`’608 Patent does not even use these terms to describe preloading, it would be
`
`improper to import them into the term preloading under BRI.
`
`
`
`Realtime’s Preloading Construction is Inconsistent with
`Examples of Preloading in the ’608 Patent
`Contrary to BRI, Realtime’s construction improperly excludes examples of
`
`preloading in the ’608 Patent. POR, 3-7, 12-16, 16-26. Specifically, the claims
`
`require preloading boot data prior to completion of initialization of the central
`
`processing unit. The ’608 Patent describes several examples where preloading
`
`occurs prior to completion of initialization because the preloading occurs prior to
`
`system reset or boot. In these examples, the ’608 Patent simply states that preloading
`
`occurs prior to system reset or boot; it does not qualify how far in advance of reset
`
`or boot the preloading occurs. Thus, these examples align with the claim language,
`
`but are not limited to preloading “in anticipation of immediate or near-in-time use,”
`
`as advocated by Realtime.
`
`For example, the ’608 Patent explains that “prior to host system reset, the data
`
`storage controller can proceed to pre-load the portions of the computer operating
`
`system from the boot device (e.g., hard disk) into the on-board cache memory.”
`
`’608, 21:45-65. In this example, “[s]ince the same portions of the operating system
`
`must be loaded upon each boot process, it is advantageous for the boot device
`
`
`
`3
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`controller to preload such portions and not wait until it is commanded to load the
`
`operating system.” Id.
`
`This example makes explicit that, contrary to Realtime’s arguments,
`
`preloading occurs “prior to host system reset” (i.e., prior to boot1). Indeed, the ’608
`
`Patent deems such an approach “advantageous.” ’608, 21:45-65.
`
`The ’608 Patent provides further confirmation that preloading may occur prior
`
`to a subsequent boot process through discussion of two alternative examples. In a
`
`first example, preloading is completed prior to boot and, in a second, alternative
`
`example, booting and preloading are performed simultaneously. ’608, 22:20-50.
`
`Specifically, the ’608 Patent explains that:
`
`It is to be understood that the depending on the resources of the given
`system (e.g., memory, etc.), the preloading process may be completed
`prior to commencement of the boot process, or continued after the boot
`process begins (in which case booting and preloading are performed
`simultaneously).
`’608, 22:40-50.
`
`This passage demonstrates that, in a system equipped with appropriate
`
`resources (e.g., non-volatile cache memory), preloading in the ’608 Patent is
`
`
`1 The ’608 Patent explains, at 11:26-30, that “the boot process begins when the
`
`CPU of the host system is released from external reset ….”
`
`
`
`4
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`completed before a boot process commences. Indeed, and as the passage makes
`
`clear by its alternative example, preloading during the boot process is only necessary
`
`in systems lacking appropriate resources.
`
`Along these lines, the ’608 Patent explains, and Dr. Back confirms2, that the
`
`cache memory device in which preloaded boot data is stored “may comprise…non-
`
`volatile memory” in which data is persistently stored across boot cycles. ’608, 6:60-
`
`63. Indeed, Dr. Back explained that, “if boot data were stored in the cache memory
`
`and if a form of non-volatile memory were used, then this data stored in the cache
`
`would be available after subsequent power on.” APPLE-1017 42, 76. Dr. Back
`
`added, “[t]hat’s what non-volatile means.” Id., 76.
`
`By describing examples of preloading prior to boot, including examples of
`
`preloading into non-volatile memory, the intrinsic record of the ’608 Patent runs
`
`contrary to Realtime’s attempt to narrow the term preloading. Indeed, Realtime’s
`
`“preloading” construction is improper because it excludes examples within the ’608
`
`Patent specification and is inconsistent with the testimony of Realtime’s own expert,
`
`Dr. Back.
`
`
`2 In his deposition on June 20, 2017, Dr. Back was asked whether the claimed cache
`
`memory “could be either volatile or non-volatile,” and he concluded that yes, “it
`
`could be either way.” APPLE-1017, 29-30.
`
`
`
`
`
`5
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Third-Party Evidence Does Not Support Limiting
`Preloading to Realtime’s Construction
`As discussed at Sections II.A.(1)-(2), the plain language of the claims and
`
`
`
`intrinsic record of the ’608 Patent do not support Realtime’s construction of
`
`preloading. Because the intrinsic record is clear that preloading should not be
`
`construed to import the hand-picked limitations offered by Realtime, consideration
`
`of extrinsic evidence is unnecessary. Third-party evidence, however, supports
`
`Apple’s interpretation of preloading.
`
`In the POR, Realtime attempts to equate the term “preload” with the term
`
`“prefetch” and use third-party definitions for the term “prefetch” as a surrogate for
`
`the actual claim term “preload.” Specifically, Realtime points to “step 77” of
`
`Figure 7b, in which boot data is said to be “preloaded into the onboard cache
`
`memory.” POR, 14 (citing ’608, 22:40-45, 22:20-39). Figure 7b labels “step 77”
`
`as “Prefetch Data Blocks Specified in List,” and Realtime asserts that this
`
`“disclosure indicates that ‘preloading’ has a meaning similar to ‘prefetching.’”
`
`POR, 14. Figure 7b, however, is the only part of the ’608 Patent that uses the term
`
`“prefetch.” Indeed, the ’608 Patent’s specification does not even mention the term
`
`“prefetching” or in any way establish that the terms “preloading” and “prefetching”
`
`have a similar meaning. The ’608 Patent’s use of the term “prefetch” in Figure 7b
`
`merely indicates that prefetching is an example of preloading; not that preloading
`
`is limited to only prefetching.
`
`
`
`6
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Realtime also attempts to use the prosecution history of the ’608 Patent to tie
`
`the term preload to the term prefetch, noting that the examiner initially rejected the
`
`’608 Patent based on a prior art reference that disclosed prefetching. Here again,
`
`Realtime’s evidence merely shows that preloading is broad enough to include
`
`prefetching as an example; not that preloading is limited to only prefetching.
`
`By providing only a single instance where the ’608 Patent uses the term
`
`prefetch as an example of preloading, Realtime falls far short in establishing that
`
`the ’608 Patent equates the term “preloading” with the term “prefetching.”
`
`Accordingly, the term “preloading” should not be limited by the definitions of the
`
`term “prefetching” provided by Realtime.
`
`Further, even assuming consideration of Realtime’s definitions of
`
`“prefetching,” Realtime’s “preloading” construction is inconsistent with these
`
`definitions and other third-party evidence. Realtime itself acknowledges that
`
`prefetching refers “to the process of retrieving data before it is needed.” POR, 15.
`
`Retrieving data “before it is needed” is not limited to “immediate or near-in-time”
`
`retrieval, and none of Realtime’s definitions even use these terms Realtime seeks
`
`to import into the claims.
`
`Additionally, third-party evidence of the actual claim term “preloading”
`
`supports an interpretation of preloading that is broader than transfer “in
`
`anticipation of immediate or near-in-time use.” For example, third-party evidence
`
`
`
`7
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`shows that the term “preload” was understood to cover loading before powering on
`
`of a computer system. APPLE-1018, 12:18-23 (“the preloaded executable module
`
`and boot time initiator 1320 are permanently stored in the read-only
`
`memory….Each time the…computer is powered on or rebooted, the boot time
`
`initiator 1320 is automatically executed”); APPLE-1019, 1:40-44 (“an offline class
`
`loader…classes are preloaded into memory without requiring runtime dynamic
`
`loading”), 2:1-7 (“an offline class loader to preload a browser in the client's read-
`
`only memory.…At system initialization or power up,…the browser is loaded from
`
`read-only memory”).
`
`In fact, as shown in FIGS. 9-10 (below), Teoman demonstrates that the term
`
`“preload” was understood to cover preloading of operating system and application
`
`data into a non-volatile cache for boot – the same operations performed by
`
`Sukegawa. APPLE-1020, 3:4-11, 7:43-63, 9:15-10:54, 13:23-59, 14:65-16:29.
`
`
`
`8
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`
`APPLE-1020, FIG. 9 (highlighted)
`
`
`
`
`
`9
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`
`APPLE-1020, FIG. 10 (highlighted)
`
`
`
`For the reasons discussed above, neither the ’608 Patent nor third-party
`
`evidence supports Realtime’s attempt to import the “in anticipation of . . . use” and
`
`“immediate or near-in-time” limitations into the term preloading.
`
`
`
`10
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`III. The Applied Prior Art Renders the Challenged Claims Obvious
`A.
`Sukegawa Describes Preloading (All Claims)
`As discussed in the Petition, Sukegawa’s controller 3 performs “a cache
`
`function” by “using the flash memory unit 1 as a cache memory,” and describes two
`
`techniques for preloading boot data into this cache memory: (1) user selection of
`
`data to preload, and (2) automatic selection of data to preload. Sukegawa, 1:50-61,
`
`4:7-10, 5:14-6:17, 6:19-58, 7:28-55; APPLE-1003, ¶¶55-58, 109-115. As Dr.
`
`Neuhauser explains, “Sukegawa makes use of the same techniques to identify data
`
`to be preloaded as the ’608 Patent.” APPLE-1003, ¶¶56-57. Because Realtime’s
`
`construction is improper and these examples fall within the ordinary meaning of
`
`preloading as set forth in Section II (above), Sukegawa describes preloading and
`
`Realtime’s arguments fail.
`
`
`
`Realtime Wrongly Advances that Preloading Cannot
`Encompass Storage in Non-Volatile Memory
`Realtime argues that Sukegawa does not disclose “preloading” but, rather,
`
`“Permanent Storage in Flash Memory for Later Access.” POR, 16. Specifically,
`
`Realtime construes preloading as involving “transferring data from storage to
`
`memory,” but contends, erroneously, that this transfer cannot encompass retrieving
`
`boot data from a boot device and “storing” it in a non-volatile memory device, for
`
`use in a subsequent boot process. Id., 4-5.
`
`
`
`11
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Indeed, Realtime’s attempt to distinguish Sukegawa is dependent on the
`
`misconception that preloading cannot involve storing data in non-volatile memory,
`
`as described in Sukegawa. POR, 18. This misconception, however, was effectively
`
`dispelled by Realtime’s expert during deposition where Dr. Back confirmed that the
`
`’608 Patent describes preloading data into a non-volatile cache3.
`
`Realtime argues that “the ’608 Patent draws a clear distinction between the
`
`concepts of ‘storing’ and ‘loading,’ and ‘storing’ and ‘preloading.’” POR, 18-20
`
`(citing ’608, Abstract, 1:15-21, 1:39-41, 3:41-51, 4:6-22). Yet, Realtime cites to
`
`only narrow slices of the ’608 Patent, ignoring broader and more relevant description
`
`that repeatedly contradicts its arguments. Id.
`
`For example, the ’608 Patent’s Abstract describes that, as part of the
`
`preloading process, “boot data is retrieved from a boot device and stored in a cache
`
`memory device.” Similarly, at 3:49-52, the ’608 Patent states that “[i]n a preferred
`
`embodiment, the boot data is retrieved from a boot device and stored in a cache
`
`
`3 Dr. Back confirmed that the claimed cache memory “could be either volatile or
`
`non-volatile,” that preloaded boot data is “stored in the cache memory device,” and
`
`that, “if boot data were stored in the cache memory and if a form of non-volatile
`
`memory were used, then this data stored in the cache would be available after
`
`subsequent power on.” APPLE-1017, 29-30, 38, 76.
`
`
`
`12
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`memory device,” and at 4:20-22, the ’608 Patent explains that “[t]he boot device
`
`controller further includes a cache memory device for storing the preloaded boot
`
`data.”
`
`In this way, and contrary to Realtime’s arguments, the ’608 Patent clearly and
`
`repeatedly discloses that preloading boot data involves “storing” that boot data in a
`
`cache memory device, just as Sukegawa does. APPLE-1003, ¶56 (citing Sukegawa,
`
`1:50-61, 5:10-14, 6:49-58, 7:7-9). Moreover, and as Dr. Back confirmed4, the ’608
`
`Patent clearly explains that the cache memory device in which preloaded boot data
`
`is stored may comprise “non-volatile memory,” similar to Sukegawa’s. ’608, 6:60-
`
`63, FIG. 1. The fact that the ’608 Patent describes other examples where the cache
`
`is volatile is inconsequential because the claims are not limited to a type of cache
`
`memory and the ’608 Patent quite clearly describes non-volatile memory, such as
`
`Sukegawa’s, as being applicable.
`
`Thus, Realtime’s attempt to distinguish Sukegawa on the basis of Sukegawa’s
`
`preloading processes involving storage of boot data in non-volatile memory fails.
`
`
`4 Dr. Back confirmed during deposition that, in the 2000 timeframe, “[t]here were
`
`high-speed non-volatile memories,” and that “boot data [could] indeed be stored on
`
`such memories.” APPLE-1017, 11.
`
`
`
`13
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Sukegawa’s Description of Preloading Matches the
`Approach Claimed and Described by the ’608 Patent
`Regardless of construction, the examples identified in Section II (above)
`
`
`
`demonstrate that, in a system equipped with appropriate resources (e.g., non-volatile
`
`cache memory), preloading in the ’608 Patent is completed before a boot process
`
`commences. ’608, Abstract, 3:49-52, 4:20-22, 6:60-63, 21:45-65, 22:40-50, FIG. 1.
`
`Indeed, as both the testimony of Dr. Back and the examples of the ’608 Patent make
`
`clear, preloading in the ’608 Patent encompasses storage of boot data in non-volatile
`
`memory, for use during a boot process that would follow a subsequent power
`
`off/power on cycle. APPLE-1017, 11, 29-30, 40-42, 76; ’608, Abstract, 3:49-52,
`
`4:20-22, 6:60-63, 11:26-30, 21:45-65, 22:20-50, FIG. 1.
`
`As Dr. Neuhauser explains, this is precisely the approach disclosed by
`
`Sukegawa, which describes preloading boot data into non-volatile cache memory, to
`
`ensure that the preloaded data is available immediately after power on. APPLE-
`
`1003, ¶¶56-57 (citing Sukegawa, 1:50-61, 4:7-11, 5:10-14, 6:49-58, 7:7-9). Thus,
`
`Sukegawa’s description matches preloading examples in the ’608 Patent and, as a
`
`consequence, a proper construction of preloading must encompass the preloading
`
`operations performed by Sukegawa.
`
`
`
`14
`
`
`
`B.
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Sukegawa Describes Preloading Prior
`to Completion of
`Initialization (All Claims)
`Realtime argues that the “plain language” of the claims “requires that the steps
`
`of ‘initializing’ and ‘preloading’ be performed during the same initialization
`
`process,” and that this, “in turn, requires that these steps be performed during the
`
`same power-on cycle of the computer system.” POR, 27-28. From this, Realtime
`
`concludes that Sukegawa cannot invalidate any claim because Sukegawa discloses
`
`initializing and preloading during different power-on cycles. POR, 28.
`
`Realtime is mistaken. Under BRI, there is no basis for requiring the claimed
`
`“initializing” and “preloading” to be performed during the same initialization
`
`process.
`
`
`
`Initializing and Preloading are not Limited to Being
`Performed During the Same Initialization Process
`Claim 1 recites “initializing a central processing unit of the computer system”
`
`(the “initializing” step) and “preloading the boot data into a cache memory prior to
`
`completion of initialization of the central processing unit of the computer system”
`
`(the “preloading” step). With this language, claim 1 does not tie “initialization” in
`
`the “preloading” step to the “initializing” step. Indeed, notably absent from claim
`
`1’s “preloading” step is reference to any particular initialization, as its preloading
`
`step recites initialization without reference to “the” initializing step, much less
`
`insertion of any requirement for antecedent basis (i.e., insertion of the article “the”)
`
`
`
`15
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`or other limitation on when preloading starts relative to the “initializing” step. To
`
`this point, as drafted, the timing of the “preloading” step is limited in only one way
`
`– it must occur “prior to completion of initialization of the central processing unit.”
`
`While the preloading must temporally occur before initialization completes, the
`
`claim is without temporal limitation on how far in advance the preloading must
`
`occur. In other words, the “preloading” step does not specify when the preloading
`
`must start. To this point, although the prior to “completion” language allows
`
`preloading to occur after initialization starts, applying the claim language, no basis
`
`exists for requiring preloading to occur contemporaneously with initialization.
`
`Realtime’s argument reinforces this point, as Realtime turns to antecedent
`
`language applied elsewhere in the claim to advance its position, tacitly recognizing
`
`the absence of similar language as a modifier for the critical term, “initialization.”
`
`Specifically, Realtime directs attention to “[t]he use of the definite article ‘the’ in
`
`the term ‘the central processing unit’,” in its attempt to argue that antecedent basis
`
`exists to support a requirement for “‘preloading prior to completion of initialization
`
`of the central processing unit’ and ‘initializing a central processing unit’
`
`...occur[ring] during the same initialization process—not two or more separate
`
`processes.” POR, 30. Although Petitioner agrees that use of the term “the” indicates
`
`antecedent basis for “central processing unit,” Petitioner disagrees that the presence
`
`of “the” as a modifier for central processing unit imparts any limitation on the
`
`
`
`16
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`initialization referenced within the preloading step. Rather, presence of “the” simply
`
`confirms that the central processing unit mentioned in the “preloading” step is the
`
`same central processing unit mentioned in the “initializing” step.
`
`Indeed, Realtime’s reliance on the definite article “the” to establish a
`
`connection between “the central processing unit,” as recited in the preloading step,
`
`and “a central processing unit,” as recited in the initialization step, suggests an
`
`awareness on Realtime’s part of the general requirements for establishing antecedent
`
`basis. And yet, when the claims were drafted and presented to the Patent Office,
`
`Realtime nevertheless demonstrates its conscious choice to omit a similar definite
`
`article or other limiting language in front of the term “initialization” in the
`
`“preloading” step.
`
`The drafter’s omission of that language, and its consequent non-existence in
`
`the claims as written, is glaring, and it leaves open as possibilities that “preloading
`
`prior to completion of initialization” and “initializing” might occur during the same
`
`initialization process, or might not. The drafter should bear the burden of this
`
`omission and the consequential breadth in claim scope, particularly given the lack
`
`of basis, either in the plain language of the claims themselves, or in the specification,
`
`for assuming, as Realtime urges, that “preloading prior to completion of
`
`initialization” and “initializing” must occur during the same initialization process.
`
`
`
`17
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Realtime argues that one particular embodiment from the specification, which
`
`describes “preloading the boot data upon initialization of the host system,” “confirms
`
`that preloading is designed to be performed with (and prior to completion of) an
`
`initialization that occurs during…the same power-on cycle.” POR, 32 (citing ’608,
`
`4:18-19). This embodiment, however, fails to support Realtime’s argument because,
`
`in plain contrast to this embodiment, the claims do not recite preloading “upon”
`
`initialization. Rather, this embodiment highlights the breadth of the claim language
`
`where “upon” was substituted for the broader “prior to completion of” language. By
`
`making this substitution, Realtime covered the “upon” embodiment, but also quite
`
`clearly intended coverage of other examples where preloading occurred “prior to,”
`
`but not “upon” initialization.
`
`Indeed, the examples identified in Section II (above) demonstrate that, in a
`
`system equipped with appropriate resources (e.g., non-volatile cache memory),
`
`preloading in the ’608 Patent occurs prior to the beginning of a boot process. ’608,
`
`Abstract, 3:49-52, 4:20-22, 6:60-63, 21:45-65, 22:40-50, FIG. 1. And, as Realtime’s
`
`own expert Dr. Back testified during deposition, the central processing unit of a
`
`computer system is initialized during the computer system’s boot process. APPLE-
`
`1017, 56. Because examples of preloading in the ’608 Patent occur prior to the
`
`beginning of a boot process that starts initialization of a central processing unit,
`
`
`
`18
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`Realtime’s position is inconsistent with the ’608 Patent specification and improperly
`
`excludes embodiments.
`
`For these reasons, Realtime’s argument that “preloading” and “initializing”
`
`must occur during the same initialization process is not supported by the plain
`
`language of the claims or the ’608 Patent specification. Accordingly, Realtime’s
`
`argument does not accord with the broadest reasonable interpretation.
`
`
`
`Sukegawa Describes Preloading Prior to Initialization of the
`CPU
`Realtime does not dispute that Sukegawa describes storing boot data in flash
`
`memory unit 1 prior to initialization of the host computer system’s CPU. Instead,
`
`Realtime limits its argument to the contention that the claimed “preloading” and
`
`“initializing” must be performed during the same initialization process. POR, 27-
`
`37. As explained above, Realtime is mistaken as to this alleged requirement.
`
`Indeed, a proper application of the broadest reasonable interpretation standard
`
`makes clear that Sukegawa describes preloading “prior to completion of
`
`initialization of the central processing unit,” and that Realtime was overreaching
`
`when pursuing the overly broad claims of the ’608 Patent. Sukegawa, 5:10-12, 6:19-
`
`26, 6:45-54; APPLE-1003, ¶¶111-114. In fact, as explained at Section III.A.(2),
`
`Sukegawa’s description matches preloading examples in the ’608 Patent and, as a
`
`consequence, a proper construction of “preloading prior to completion of
`
`initialization” must encompass the preloading operations performed by Sukegawa.
`
`
`
`19
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`C. Dye Describes a Plurality of Encoders (Claims 16-17, 20-21, 25-26,
`30-31)
`As discussed in the Petition, Dye “uses a ‘parallel lossless
`
`compression/decompression’ ‘designed to process stream data at more than a
`
`single byte or symbol (character) at one time.’” Petition, 60. Realtime contends
`
`that Dye’s disclosure is insufficient because Dye implements “a single algorithm”
`
`with a “single encoder that distributes the encoding calculations among several
`
`stages.” POR, 38 (emphasis in original). Although Realtime acknowledges that
`
`Dye includes a plurality of “stages” or “processes” in performing encoding,
`
`Realtime contends that each stage or process “perform[s] a subpart of the single
`
`encoder’s encoding operation, but do[es] not each output encoded data
`
`themselves.” POR, 38-42. Realtime is incorrect.
`
`As Dr. Neuhauser explains, the logic of Fig. 13 serves “as an encoder of the
`
`Input Stream” “because this logic encodes four bytes of the input stream 610
`
`against the four bytes in the history table...and produces encoded information
`
`representing the sequence in Output Mask and Output Count.” APPLE-1003,
`
`¶¶284-285. In fact, “there are 16 such encoders used in the system of Dye” and the
`
`output of these encoders “are sent to the selection logic shown in Dye Fig. 14.”
`
`APPLE-1003, 284-286. Dye’s selection logic processes “multiple compressed
`
`blocks” from the 16 such encoders and “generates a combined compress mask.”
`
`Dye, 23:58-24:15 (“a compressed block that includes one or more of the 4 symbols
`
`
`
`20
`
`
`
`Proceeding No.: IPR2016-01365
`Attorney Docket: 39521-0023IP1
`received on this cycle”). For these reasons, as Dr. Neuhauser explained, Dye
`
`includes a plurality of encoders that, a