`
`Filed on behalf of EMC Corporation
`
`By: Peter M. Dichiara, Reg. No. 38,005
`
`David L. Cavanaugh, Reg. No. 36,476
` WILMER CUTLER PICKERING HALE AND DORR LLP
`
`peter.dichiara@wilmerhale.com
`
`david.cavanaugh@wilmerhale.com
`
`Tel.: 617-526-6466
`
`Fax: 617-526-5000
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`____________________________________________
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`____________________________________________
`
`EMC CORPORATION,
`Petitioner
`
`v.
`
`Patent Owner of
`U.S. Patent No. 8,001,096 to Farber et al.
`
`
`IPR Case No. IPR2013-00087
`
`
`PETITIONER’S REPLY
`
`
`
`
`
`
`
`
`

`

`
`
`
`
`
`TABLE OF CONTENTS
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`INTRODUCTION ............................................................................................... 1
`I.
`II. CLAIM 1 IS OBVIOUS OVER KANTOR AND SATYANARAYANAN ... 1
`ARGUMENT 1: PersonalWeb’s contention that “Kantor fails to disclose
`digital part identifiers for all parts of a data item, and also fails to disclose a data
`item identifier that is based on at least all data of the data item.” (Resp. 3-18)
`(Claim 1)………… ………………………………………………………………1
`ARGUMENT 2: PersonalWeb’s contention that Kantor’s ‘y’ procedure
`“emphasizes the deficiencies in Kantor’s ‘zipfile contents signature’ procedure”
`(Resp. 18-21) (Claim 1) .......................................................................................... 4
`ARGUMENT 3: PersonalWeb’s contention that “Kantor is incapable of
`sections (A4)-(A5)” (Resp. 21-22) (Claim 1) ....................................................... 5
`ARGUMENT 4: PersonalWeb’s contention that “Kantor fails to disclose a
`determining part identifiers by applying a first hash function to parts of the ZIP
`file.” (Resp. 23-29) (Claim 1) ................................................................................ 5
`ARGUMENT 5: PersonalWeb’s contention that Kantor fails to disclose
`determining the part identifiers “in the file system” (Resp. 29-31) (Claim 1) ....... 9
`ARGUMENT 6: PersonalWeb’s contention that “Kantor teaches directly away
`from the alleged modification of adding zipfiles to multiple servers of the
`system.” (Resp. 31-34) (Claim 1) ........................................................................... 9
`ARGUMENT 7: PersonalWeb’s contention that “[i]t would not have been
`obvious to have modified Kantor so that read and download BBS commands
`would accept contents-signatures to identify files on which to operate.” (Resp.
`35-42) (Claim 1)… ……………………………………………………………...11
`III. CLAIMS 81 AND 83 ARE OBVIOUS OVER KANTOR AND
`SATYANARAYANAN ..........................................................................................13
`IV. SECONDARY CONSIDERATIONS ............................................................13
`V. KANTOR IS A PRINTED PUBLICATION .................................................14
`
`
`
`
`ii
`
`

`

`
`
`I.
`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`INTRODUCTION
`
`PersonalWeb presents seven separate arguments in an effort to support its
`
`position that Kantor does not invalidate the challenged ’096 claims. However,
`
`Kantor not only meets the requirements of the claims, but operates like the ’096
`
`embodiments. PersonalWeb’s arguments to the contrary are premised on overly-
`
`narrow interpretations of the claims that ignore the Board’s constructions and
`
`Kantor’s explicit disclosures. Moreover, PersonalWeb improperly focuses on
`
`Kantor’s software, ignoring the broader disclosures in Kantor’s publication. The
`
`Board accordingly should reject the challenged claims for the same reasons
`
`identified in its initial institution decision (“Decision”) and in view of the
`
`comments below.
`
`II. CLAIM 1 IS OBVIOUS OVER KANTOR AND SATYANARAYANAN
`ARGUMENT 1: PersonalWeb’s contention that “Kantor fails to disclose
`digital part identifiers for all parts of a data item, and also fails to disclose a
`data item identifier that is based on at least all data of the data item.” (Resp.
`3-18) (Claim 1)
`
`
`
`PersonalWeb devotes the first 15 pages of its response to the argument that
`
`Kantor fails to disclose digital part identifiers for “each part” of a data item—or a
`
`data item identifier “based, at least in part, on the contents of the data item”—
`
`because Kantor’s “zipfile contents-signatures” are based on “contents-signatures”
`
`of the inner files of a zipfile, and not on hashes of other information in the zipfile
`
`about the inner files (e.g., file names, directory information, time and date
`
`information, etc.). (Resp. 3, 15; see also ’096 patent, claim 1; Ex. 1001.)
`
`1
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`PersonalWeb premises this argument on an overly-narrow interpretation of the
`
`claim that is not the broadest reasonable construction. Indeed, it is not even a
`
`reasonable construction because it would exclude the ’096 preferred embodiments.
`
`PersonalWeb’s argument is thus easily dismissed for at least two reasons.1
`
`First, the Board construed a “data item”—consistent with the patent’s
`
`lexicography and both parties’ proposed constructions—as “any [ ] entity which
`
`can be represented by a sequence of bits,” including, among other things, “(2) a
`
`portion of a file.” (Decision 11 (emphasis added).) There can be no reasonable
`
`argument that the inner files of a zipfile are anything other than “a portion of a
`
`file.” (Reply Clark Decl. ¶¶ 4-9; Ex. 1089.) Kantor thus squarely meets this
`
`requirement. Kantor’s inner files are a “data item” (“(2) a portion of a file”) and
`
`consist solely of a “sequence of non-overlapping parts.” (Id. ¶ 9.)
`
`PersonalWeb attempts to argue that, because a data item is a “sequence of
`
`bits,” it cannot include some of the bits in a zipfile (the inner files) and exclude
`
`other “intervening” bits. (Resp. 16-17.) It purports to rely on Dr. Clark for this
`
`argument, but he explicitly confirmed in his deposition that there are many
`
`examples of sequences with intervening gaps including Fibonacci sequences,
`
`
`1 Moreover, PersonalWeb’s expert could not say whether he had even read the
`
`Board’s Decision in connection with his work on the case. (Dewar Tr. 23-24; Ex.
`
`1083.)
`
`2
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`random sequences, odd sequences, and even sequences. (Clark Tr. 191-93; Ex.
`
`2016.) The broadest reasonable construction of the term clearly would encompass
`
`such sequences, which have gaps. Moreover, PersonalWeb admits that the inner
`
`files (i.e., the “parts”) are “separated from each other” (Resp. 16), thus squarely
`
`falling within the meaning of “non-overlapping” parts.
`
`Second, even if the data item were viewed as all data in the zipfile,
`
`PersonalWeb bases its argument on a fundamentally flawed view of the claim that
`
`would exclude the preferred embodiments. The ’096 patent is explicit that the
`
`identifiers are based on the contents of the data item. (’096 patent 3:55-59; Ex.
`
`1001; Reply Clark Decl. ¶¶ 10-14; Ex. 1089; Dewar Tr. 48; Ex. 1083.) As both
`
`parties’ experts have confirmed, the ’096 patent purposefully and intentionally
`
`does not hash information about the contents in the data item (e.g., filename,
`
`directory information, time and date information, location information, etc.)
`
`because, if it did, the identifiers could not be used to accomplish the patent’s goals
`
`(e.g., the identifiers would change when pathname, directory, time of last access,
`
`or location change, even though the contents did not change). (Reply Clark Decl. ¶
`
`11; Ex. 1089; Dewar Tr. 48-53; see also 62-63; Ex. 1083.) PersonalWeb’s expert
`
`has admitted that he considers this independence a “fundamental property” of the
`
`invention. (Dewar Tr. 51-53; Ex. 1083.) The Summary of the Invention thus
`
`states that “the identity of a data item is independent of its name, origin, location,
`
`address, or other information not derivable directly from the data, and depends
`
`3
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`only on the data itself.” (’096 patent 3:55-59; Ex. 1001.) As the Federal Circuit
`
`has repeatedly confirmed, a construction that excludes the preferred embodiment is
`
`“rarely, if ever” correct. See, e.g., Accent Packaging, Inc. v. Leggett & Platt, Inc.,
`
`707 F.3d 1318 (Fed. Cir. 2013); see also SciMed Life Systems, Inc. v. Adv.
`
`Cardiovascular Sys., Inc., 242 F.3d 1337, 1345 (patent cannot cover disclaimed
`
`structure).
`
`
`
` In fact, Kantor not only meets the elements of the claim—including the
`
`requirement of “identifiers” for “each part” of a data item—it operates just like the
`
`‘096 patent. (Compare ’096 patent 3:55-59; Ex. 1001 (“[T]he identity of a data
`
`item is independent of its name, origin, location, address, or other information not
`
`derivable directly from the data.”); with Kantor Preface 2; Ex. 1004 (“a ‘zipfile
`
`contents signature’, (‘zcs’) . . . is independent of . . . zipfile name, zipfile date, . . .
`
`the name and dates of files in the zipfile, zipped path information.”); see also
`
`Reply Clark Decl. ¶¶ 10-11; Ex. 1089 (confirming that Kantor creates zipfile
`
`contents signatures independently of information not directly derivable from the
`
`data contents).) PersonalWeb’s attempt to distinguish Kantor is therefore
`
`misplaced. Kantor accomplishes the same goal as the ’096 patent (identifying and
`
`eliminating duplicate data items) in the same way (by basing the identifiers on the
`
`contents of the data items rather than the associated metadata).
`
`ARGUMENT 2: PersonalWeb’s contention that Kantor’s ‘y’ procedure
`“emphasizes the deficiencies in Kantor’s ‘zipfile contents signature’
`
`4
`
`

`

`
`
`procedure” (Resp. 18-21) (Claim 1)
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`
`
`As PersonalWeb admits, Kantor discloses an “optional ‘y’ procedure” that
`
`“calculates a CRC value [hash] based on ‘every byte in the zipfile.’” (Resp. 18.)
`
`PersonalWeb contrasts this procedure with Kantor’s “z” procedure, which
`
`calculates contents-signatures for the inner files but not for the metadata,
`
`suggesting that this is a deficiency of the “z”’ procedure. (Id. 18-19.) In fact,
`
`PersonalWeb’s argument illustrates that whether or not to hash metadata is a mere
`
`design choice, and that Kantor recognized and implemented both choices. Kantor
`
`expresses a preference not to hash metadata for precisely the same reasons that the
`
`’096 does not hash metadata (See supra Argument 1.) Dr. Clark confirms there is
`
`absolutely nothing inventive about hashing metadata, even if that were required by
`
`the ’096 claims (which it is not). (Reply Clark Decl. ¶¶ 15-17; Ex. 1089.)
`
`ARGUMENT 3: PersonalWeb’s contention that “Kantor is incapable of
`sections (A4)-(A5)” (Resp. 21-22) (Claim 1)
`
`
`
`PersonalWeb bases this argument on the same theory as Argument 1—
`
`namely, that Kantor does not hash “all parts” of a zipfile and thus is “incapable of
`
`sections A4 and A5” of claim 1. (Resp. 21.) As explained above, however, the
`
`most reasonable interpretation of the claim, and certainly the broadest reasonable
`
`interpretation of the claim, does not demand hashing of metadata. (See supra
`
`Argument 1; see also Clark Reply Decl. ¶ 18; Ex. 1089.) Likewise a “data item”
`
`can be a “(2) a portion of a file.” PersonalWeb’s argument accordingly fails.
`
`ARGUMENT 4: PersonalWeb’s contention that “Kantor fails to disclose a
`
`5
`
`

`

`IPR2013-00087
`
`
`Docket No. 0100157-00240
`
`
`determining part identifiers by applying a first hash function to parts of the
`ZIP file.” (Resp. 23-29) (Claim 1)
`
`PersonalWeb bases this argument on a new construction, arguing that claim
`
`1 “requires applying a first function comprising a first hash to parts of the data
`
`item.” (Resp. 23 (emphasis added).) It then spends seven pages attempting to
`
`argue that Kantor does not “apply” a hash function to the inner files of a zipfile
`
`even though it admits this very fact in one of the earlier sections of its response: “a
`
`CRC . . . is separately applied to each of a plurality of uncompressed files.” (Id. 6
`
`(emphasis added).) The Board has similarly recognized that Kantor “applies” a
`
`CRC hash function to each of the inner files. (Decision 16.)
`
`Even if PersonalWeb’s construction were correct (which it is not), its
`
`arguments that Kantor does not satisfy the claim are fundamentally flawed. First,
`
`PersonalWeb contends that Kantor’s CRC hashes are not applied to the “parts” of a
`
`zipfile because they are applied to uncompressed files before they are compressed
`
`and packaged into the zipfile. (Resp. 23-35.) But nothing in Kantor limits the
`
`“inner files” of the zipfile to compressed files. In fact, PersonalWeb tacitly admits
`
`to this when it states that “[t]he inner files within a ZIP file are almost always
`
`compressed” (i.e., not “always”) (id. 18 (emphasis added)), and both parties’
`
`experts agree that zipfiles can have uncompressed inner files. (Reply Clark Decl.
`
`¶¶ 19-21; Ex. 1089; Dewar Tr. 263-64; Ex. 1083.) Indeed, Kantor is clear that his
`
`software works with zipfiles of all forms, and like the ’096 patent, creates the same
`
`6
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`identifiers whether or not the files are compressed. (See, e.g.,’096 patent Fig. 28;
`
`9:39-43 and 23:41-46; Ex. 1001.) For example, Kantor explains that the resulting
`
`zipfile identifier depends on neither “the method nor amount of compression.”
`
`(Kantor 9; Ex. 1004.) Consistent with this statement, the zipfile standard - cited in
`
`Kantor - confirms that “compression method 0” has no compression: “0 – The file
`
`is stored (no compression).” (Ex. 2007 at 4 (emphasis added); see also Kantor
`
`Preface 2; Ex. 1004; Reply Clark Decl. ¶ 20; Ex. 1089; Dewar Tr. 262-63; Ex.
`
`1083.) Thus, regardless of whether or not its analysis is correct for inner files that
`
`are compressed before being packaged into a zipfile (which it is not), PersonalWeb
`
`has no argument or rebuttal for the situation in which Kantor processes zipfiles
`
`with uncompressed inner files. (Reply Clark Decl. ¶¶ 21-22; Ex. 1089.)
`
`Moreover, even if Kantor only used compressed inner files (which it does
`
`not), it would still satisfy the claim because the “part values” would still be based
`
`on a “function” that is “comprising” (i.e., including) a hash. The function would
`
`include both the hash and the compression. (Id. ¶ 23.)
`
`Second, PersonalWeb asserts that “Kantor reads the CRC values from the
`
`ZIP file and reads the uncompressed file lengths (sizes) from the ZIP file,” arguing
`
`that such “read[ing of] CRC values from the ZIP file” does not amount to the
`
`“applying a first function comprising a hash to each of a plurality of parts of the
`
`first data item.” (See Resp. 27.) This argument focuses on the operation of
`
`Kantor’s software. However, the issue is not how Kantor’s software operates, but
`
`7
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`rather, what Kantor’s publication discloses. Kantor’s publication clearly discloses
`
`that the CRC values of inner files, as one would expect, are applied by a
`
`“computer-implemented method,” which is all the claim requires. The publication
`
`makes plain that PKZIP software creates the CRC values for the inner files, and
`
`thus the “part values” are indeed generated by applying a first hash function to
`
`each part of the first plurality of parts (i.e., to the inner files). (See Kantor 6, 15,
`
`71-72, 192; Ex. 1004; Reply Clark Decl. ¶¶ 24-25; Ex. 1089; see also Ex. 2007 at
`
`3.) And as stated above, PersonalWeb admits in an earlier argument that Kantor
`
`discloses “a CRC (cyclic redundancy check) [that] is separately applied to each of
`
`a plurality of uncompressed files…”) (Resp. 6.)
`
`Moreover, even if the issue were Kantor’s software rather than Kantor’s
`
`publication (which it is not), Kantor is clear that the software is capable not only of
`
`reading CRC values but also of generating them. Kantor reads CRC values from a
`
`zipfile (previously generated by PKZIP software)—when they are available—
`
`because there is no reason to recompute what is already there. (See Clark Tr. 75;
`
`Ex. 2016; see also Reply Clark Decl. ¶¶ 21, 26; Ex. 1089.) Kantor clearly
`
`discloses, however, that its software will generate hash values if there is a need to
`
`do so (i.e., if they are not already there). (See Kantor 51 (“Make a ‘file contents
`
`signature’ for (each) Plain file (non-zip). In this case, unlike the case of a file in a
`
`zipfile, FWKCS can’t just look up the 32-bit CRC: it calculates it.”); Ex. 1004; see
`
`also Reply Clark Decl. ¶ 26; Ex. 1089.) Moreover, there is nothing inventive in
`
`8
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`calculating versus reading a CRC value, particularly given Kantor’s ability to
`
`generate a hash if one is needed.2
`
`ARGUMENT 5: PersonalWeb’s contention that Kantor fails to disclose
`determining the part identifiers “in the file system” (Resp. 29-31) (Claim 1)
`
`
`
`PersonalWeb bases this argument on two new constructions, contending that
`
`the preamble is limiting, and that the “determining” step (A1) must be performed
`
`on particular computers (the BBS) and not on other computers that communicate
`
`with the BBS. Whether or not the preamble is limiting, however, any person of
`
`ordinary skill in the art would appreciate that “a file system” could constitute the
`
`BBS, or it could constitute the BBS in combination with the computers
`
`communicating with the BBS which form the part identifiers, each of which is
`
`disclosed in Kantor. (See, e.g., Clark Decl. ¶¶ 85 and 99; Ex. 1009; see also Reply
`
`Clark Decl. ¶ 27; Ex. 1089.) For example, Kantor’s “Lookup” command, among
`
`others, illustrates that the combination of a BBS and a user computer operate as a
`
`file system. (See Kantor 96; Ex. 1004; Reply Clark Decl. ¶ 27; Ex. 1089.)
`
`ARGUMENT 6: PersonalWeb’s contention that “Kantor teaches directly
`away from the alleged modification of adding zipfiles to multiple servers of
`the system.” (Resp. 31-34) (Claim 1)
`
`
`
`PersonalWeb argues that Kantor teaches away from a modification to store
`
`“parts” on multiple servers. In doing so, it significantly distorts Kantor’s
`
`2 PersonalWeb also argues that a zipfile may include information in addition to the
`
`“inner files.” (Resp. 17.) This argument fails for the same reason as Argument 1.
`
`9
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`teachings. Kantor is not concerned with avoiding all duplicate files; instead, it
`
`focuses on avoiding unwanted duplicates. (Kantor Preface 2; Ex. 1004.) In fact,
`
`Kantor allows users to retain duplicates, if desired, by allowing them to control the
`
`MULTIS file which indicates which duplicates to save and which to delete. (Id.
`
`189-90; see also Reply Clark Decl. ¶ 28; Ex. 1089.) As Dr. Clark confirmed, it
`
`accordingly would have been obvious to modify Kantor to store “mirrored” files in
`
`a redundant manner to enhance reliability. (Reply Clark Decl. ¶ 29; Ex. 1089.) As
`
`the ’096 patent explains, this “mirroring” of files does not create unwanted
`
`duplicates, but instead intentionally stores two or more desired copies on two or
`
`more disks, in case one disk were to fail. (See ’096 patent 4:18-21; Ex. 1001;
`
`Reply Clark Decl. ¶ 29; Ex. 1089.) As Dr. Clark further confirms, Kantor is silent
`
`about mirroring because storage devices and mirroring techniques are an issue for
`
`the underlying BBS operator, not the Kantor system. (Reply Clark Decl. ¶ 29; Ex.
`
`1089.) Consequently, there is no teaching away, and it would have been obvious
`
`to utilize the mirroring techniques of Satyanarayanan to increase the reliability and
`
`response time of requests for files stored by the BBS systems.3 (Id.; see also
`
`Satyanarayanan II 450;Ex. 1028.) Indeed, Dr. Dewar agreed that mirroring
`
`
`3 PersonalWeb also argues that there is no “second mapping data” for filenames,
`
`directory information, time and date information, etc. This argument fails for the
`
`same reasons as Arguments 1 and 3.
`
`10
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`technology was old. (Dewar Tr. 114-15; Ex. 1083.)
`
`ARGUMENT 7: PersonalWeb’s contention that “[i]t would not have been
`obvious to have modified Kantor so that read and download BBS commands
`would accept contents-signatures to identify files on which to operate.” (Resp.
`35-42) (Claim 1)
`
`
`
`The Board has recognized “that it would have been obvious to allow
`
`download and read commands of the system to identify a file by a contents-
`
`signature,” (Decision 17.) PersonalWeb’s arguments to the contrary are flawed.
`
`
`
` First, PersonalWeb argues that neither Kantor nor Satyanarayanan (Ex.
`
`1028) discloses using anything other than a conventional file name for read and
`
`download commands, so there would be no motivation for the combination. (Resp.
`
`35.) This statement is misplaced. The point is that it would be obvious to modify
`
`read or download commands to utilize hash-based identifiers. Contrary to
`
`PersonalWeb’s assertions, Dr. Clark has provided examples where Kantor shows
`
`that modifying read and download commands to accept a contents-signature as
`
`input would be easy. (Clark Decl. ¶ 83; Ex. 1009; see also Reply Clark Decl.
`
`¶¶ 30-35; Ex. 1089). Identifying files based on their contents-signatures as input
`
`already existed as part of the “Lookup” and “Exclude” commands. (Reply Clark
`
`Decl. ¶¶ 31, 35; Ex. 1089.) He has explained that the precise functionality needed
`
`for the modified read or download command (i.e., providing the part identifiers in
`
`response to a zipfile contents signature) already existed as part of the “Lookup”
`
`command. (See Kantor 96, 113; Ex. 1004; Clark Decl. ¶¶ 83, 93-95; Ex. 1009;
`
`11
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`Clark Tr. 212-213; Ex. 2016; see also Reply Clark Decl. ¶ 32; Ex. 1089.)
`
`Moreover, Dr. Clark has confirmed that a person of ordinary skill would have been
`
`motivated to modify Kantor, for example, to facilitate integrity checking. (Clark
`
`Decl. ¶ 83; Ex. 1009; see also Reply Clark Decl. ¶¶ 30-33; Ex. 1089.)
`
`PersonalWeb grossly mischaracterizes both Dr. Clark’s declaration and
`
`Kantor in suggesting that the separate documents LOOKUP.DOC and
`
`PRECHECK.DOC reveal “fatal” flaws in Dr. Clark’s analysis. (See Resp. 36.)
`
`Dr. Clark did not rely on the contents of these files, as PersonalWeb suggests.
`
`Instead, as Dr. Clark made clear in his deposition, he relied on and cited only the
`
`disclosure within Kantor that mentions the “Lookup” and “Precheck” functions.
`
`(See Clark Dec. ¶ 83; Ex. 1009; Clark Tr. 115-17; Ex. 2016; see also Reply Clark
`
`Dec. ¶ 36; Ex. 1089.) The Kantor reference itself provides ample disclosure to
`
`show how it uses content-based identifiers as input to the BBS running the Kantor
`
`system. (See Clark Decl. ¶ 83; Ex. 1009; see also Clark Tr. 201; Ex. 2016; Kantor
`
`96-98, 173; Ex. 1004.)
`
`Second, PersonalWeb argues that the cited art fails to disclose any problem
`
`with the use of conventional file names for BBS commands. To the contrary,
`
`Kantor explicitly discloses certain deficiencies from using conventional file names.
`
`For example, its Exclude command utilizes a contents-signature to identify a file,
`
`rather than a conventional file name, because the contents-signature can more
`
`correctly identify a file that should be excluded by the BBS than a conventional
`
`12
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`name can. (See Reply Clark Decl. ¶ 35; Ex. 1089.)
`
`Third, PersonalWeb reiterates the irrelevant argument that Kantor may have
`
`different files with the same signature. This issue was addressed (see Arguments 1
`
`and 3) and will not be repeated here for brevity, other than to restate that the claim
`
`simply requires the same file to have the same signature. Indeed, PersonalWeb
`
`does not contest—and its expert admits—that Kantor satisfies this element.
`
`(Dewar Tr. 252-53; Ex. 1083.) Likewise, Dr. Dewar acknowledged that the ’096
`
`patent also can have different files with the same signature. (Id. at 60-61.)
`
`III. CLAIMS 81 AND 83 ARE OBVIOUS OVER KANTOR AND
`SATYANARAYANAN
`
`PersonalWeb provides no new arguments for claims 81 and 83. Instead, it
`
`relies on the same claim limitations and same arguments already discussed above
`
`for claim 1. (Resp. 42-50; see also Reply Clark Decl. ¶ 37; Ex. 1089.)
`
`PersonalWeb’s arguments accordingly fail for the same reasons as its arguments
`
`for claim 1. (See supra Section II, Arguments 1-7.)
`
`IV.
`
` SECONDARY CONSIDERATIONS
`
`PersonalWeb relies on three prior licenses to support its argument that claim
`
`1 is not obvious. However, it has failed to establish the required nexus. See In re
`
`GPAC, Inc., 57 F.3d 1573, 1580 (Fed. Cir. 1995). The agreements grant licenses
`
`to an array of U.S. and foreign patents, and PersonalWeb has not provided any
`
`evidence that the claims-at-issue motivated the decision to license. (See, e.g., Ex.
`
`13
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`2012 at § 1.1.4.) Moreover, the cross-examination of PersonalWeb’s declarant,
`
`Kevin Bermeister, revealed that each of the three licenses involved related parties
`
`with interlocking ownership and/or business interests, undermining any inference
`
`that the agreements represent a recognition of the merits of the claimed inventions.
`
`(See Bermeister Tr. 54-57, 63-66, 85-89, 121-144; Ex. 1087.) Similarly, while
`
`PersonalWeb alleges that the consideration paid for the licenses exceeded any
`
`reasonable litigation cost, on cross-examination, it became clear that Mr.
`
`Bermeister had no basis for the “consideration” values contained in his declaration.
`
`(Id. 28-37, 53-57, 72-76, 92-98, 112-113, 117-121.)
`
`V. KANTOR IS A PRINTED PUBLICATION
`
`
`
`Kantor is a printed publication that was widely distributed to the public more
`
`than a year before the critical date. As the Board noted, Kantor itself explains how
`
`to get new versions of the software from The Invention Factory BBS (“BBS”), and
`
`“the title page of Kantor clearly shows the posted date of August 10, 1993.”
`
`(Decision 19-20.) The sworn testimony of Michael Sussell, the system operator of
`
`the BBS, corroborates the distribution (See Sussell Decl. ¶¶ 15-27; Ex. 1050.)
`
`
`
`Petitioner also has established that the Kantor publication was distributed to
`
`the public in October 1993 on a CD-ROM by a company called Walnut Creek CD-
`
`ROM. Jason Sadofsky, a technology archivist, obtained the October 1993 CD-
`
`ROM (Exhibit 1049) in 2009, authenticated it, verified that it contained Kantor,
`
`14
`
`

`

`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`
`and identified publications prior to the critical date advertising its distribution.
`
`(See Sadofsky Decl. ¶¶ 3, 16-17; Ex. 1078; Sadofsky Tr. 56-66; Ex. 2013.)
`
`
`
`Despite all of this evidence, PersonalWeb argues that Petitioner has failed to
`
`prove Kantor was published before the critical date because it has not submitted a
`
`declaration from the author or from someone who read Kantor prior to the critical
`
`date, but cites no authority that such declaration is needed. (Resp. 53.) “It is
`
`hornbook law that direct evidence of a fact is not necessary. Circumstantial
`
`evidence is not only sufficient, but may also be more certain, satisfying and
`
`persuasive than direct evidence.” Moleculon Research Corp. v. CBS, Inc., 793 F.2d
`
`1261, 1272 (Fed. Cir. 1986) (internal quotations omitted).
`
`
`
`PersonalWeb also argues that Kantor was not publicly accessible because
`
`Kantor “intentionally buried” the Kantor publication in a zipfile. (Resp. 34.)
`
`PersonalWeb relies on the declaration of Todd Thompson, but Mr. Thompson was
`
`not a person skilled in the art in 1994. (Thompson Decl. ¶¶ 2-3; Ex. 2014;
`
`Thompson Tr. 7-8, 13-14; Ex. 1086.) He also did not follow the installation
`
`instructions or use appropriate software for the 1993-1994 timeframe. (Thompson
`
`Tr. 32-35, 39-42, 51-56; Ex. 1086.) Moreover, even Mr. Thompson ultimately
`
`succeeded in accessing the Kantor publication. (See Thompson Decl. ¶ 22; Ex.
`
`2014.) Petitioner submits a Reply Declaration from Mr. Sadofsky (Ex. 1088)
`
`showing the ease by which a person skilled in the art could access Kantor by
`
`following the included directions and using compatible software.
`
`15
`
`

`

`
`
`
`
`
`
`
`IPR2013-00087
`Docket No. 0100157-00240
`Respectfully Submitted,
`
`Dated: October 2, 2013
`
`/Peter M. Dichiara/
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Peter M. Dichiara
`Registration No. 38,005
`Cynthia Vreeland
`WILMER CUTLER PICKERING
`HALE AND DORR LLP
`60 State Street
`Boston, Massachusetts 02109
`peter.dichiara@wilmerhale.com
`Tel.: 617-526-6466
`Fax: 617-526-5000
`
`16
`
`

`

`
`
`
`
`
`IPR2013-00087
`
`Docket No. 0100157-00240
`
`CERTIFICATE OF SERVICE
`
`I hereby certify that on October 2, 2013, I caused a true and correct copy of
`
`the following materials:
`
`• Petitioner’s Reply
`
`• Exhibits 1080-1090
`
`• List of Exhibits
`
`to be served via email on the following counsel of record for Patent Owner:
`
`Joseph A. Rhoa, Lead Counsel
`USPTO Reg. No. 37,515
`NIXON & VANDERHYE P.C.
`901 North Glebe Road, 11th Floor
`Arlington, VA 22203-1808
`jar@nixonvan.com
`Tel.: 703-816-4043
`
`Updeep S. Gill, Backup Counsel
`USPTO Reg. No. 37,344
`NIXON & VANDERHYE P.C.
`901 North Glebe Road, 11th Floor
`Arlington, VA 22203-1808
`usg@nixonvan.com
`Tel.: 703-816-4030
`
`/Owen K. Allen/
`
`Owen K. Allen
`Reg. No. 71,118
`
`
`
`17
`
`
`
`
`
`
`
`
`
`
`
`
`