`
`
`IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`
`PATENT: 6,415,280
`
`INVENTOR: DAVID A. FARBER
`AND RONALD D. LACHMAN
`
`
`
`
`
`FILED: APR. 1, 1999
`
`ISSUED: JUL. 2, 2002
`
`
`
`TITLE: IDENTIFYING AND
`REQUESTING DATA IN A
`NETWORK USING IDENTIFIERS
`WHICH ARE BASED ON THE
`CONTENT OF THE DATA
`
`Mail Stop PATENT BOARD
`Patent Trial and Appeal Board
`U.S. Patent & Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`
`DECLARATION OF DOUGLAS W. CLARK, PH.D.
`
`
`
`I, Douglas W. Clark, declare as follows:
`
`1.
`
` My name is Douglas Clark. I received my B.S. in engineering and
`
`applied science from Yale University in 1972 and my Ph.D. in computer science
`
`from Carnegie-Mellon University in 1976. Since receiving my doctorate, I have
`
`devoted my professional career to the research, design, development, study, and
`
`teaching of numerous aspects of computer systems architecture and design.
`
`
`
`- 1 -
`
`EMCVMW 1009
`
`
`
`2.
`
`I am currently a Professor of Computer Science at Princeton
`
`University. I have held this position since 1993. My teaching experience at
`
`Princeton includes a large introductory lecture course for engineering freshmen,
`
`advanced courses in computer architecture for upper-level undergraduates and
`
`graduate students, advanced graduate seminars in various topics in computer
`
`systems, a course on the great papers of computer science, and an introductory
`
`course for nonscientists. I have taught most of these courses several times. I have
`
`also taught as a Visiting Lecturer in the Division of Applied Sciences at Harvard,
`
`and as a Visiting Professor of Computing and Information Science at the
`
`University of Pennsylvania, in 2003.
`
`3.
`
`As part of my research at Princeton, I have worked in a number of
`
`areas, including in the multiprocessor SHRIMP project, specializing in
`
`performance monitoring issues; in simulation-based architecture investigations of
`
`modern dynamic processors, focusing on branch prediction; in video camera-based
`
`automatic alignment strategies for large multi-projector displays; and in various
`
`low-power techniques for modern microprocessors, including clock speed
`
`management in separate clocking domains.
`
`4.
`
`In addition to my experience in academia, I have over 14 years of
`
`industrial experience designing computer systems. From 1976 to 1980, I was a
`
`Member of the Research Staff at the Xerox Palo Alto Research Center, where I
`
`
`
`- 2 -
`
`
`
`worked chiefly on the design of the Dorado, one of the earliest high-performance
`
`workstations. From 1980 to 1990, I worked for the Digital Equipment Corp., first
`
`as a Principal Engineer in the Systems Architecture Group, and then as a
`
`Consulting and Senior Consulting Engineer in both the Advanced VAX Systems
`
`Engineering and Alpha Advanced Development groups. I worked mainly on the
`
`architecture, organization, design, simulation, and performance analysis of VAX
`
`and Alpha computers. I was one of the principal designers of the VAX 8700 and
`
`VAX 8800 – both highly successful machines of the late 1980’s.
`
`5.
`
`I have authored or co-authored about 60 academic publications in the
`
`fields of computer science and engineering. In addition, I have been a referee or
`
`associate editor for the following academic journals: ACM Transactions on
`
`Computers, IEEE Transactions on Computers, and IEEE Computer.
`
`6.
`
`I have also been a program committee member or co-chair at
`
`numerous national and international conferences/symposiums, including the
`
`International Conference on Computer Design, SIGMETRICS Conference on
`
`Measurement and Modeling of Computer Systems, and International Symposium
`
`on Computer Architecture.
`
`7.
`
`8.
`
`A copy of my latest curriculum vitae (CV) is attached as Appendix A.
`
`I have reviewed and understand the specification, claims, and file
`
`history of the ‘280 patent. I have been informed that the ‘280 patent claims
`
`
`
`- 3 -
`
`
`
`priority to U.S. Patent Application No. 08/425,160, filed on April 11, 1995. I
`
`understand this means the ‘280 patent is considered to have been filed on April 11,
`
`1995 for the purposes of determining whether a reference will qualify as prior art.1
`
`9.
`
`I have also reviewed and understand the following references:
`
`(cid:120) S. Browne et al., “Location-Independent Naming for Virtual
`Distributed Software Repositories,” University of Tennessee
`Technical Report CS-95-278 (Feb. 1995) (“Browne ”, Ex. 1002);
`S. Browne et al., “Location-Independent Naming for Virtual
`Distributed Software Repositories,”
`http://www.netlib.org/utk/papers/lifn/main.html (Nov. 11, 1994)
`(“Browne 1994”, Ex. 1006); and K. Moore et al., “An Architecture
`
`
`1 I have also been informed that the ‘280 patent is a division of U.S. Application
`
`No. 08/670,079, filed on Oct. 24, 1997, now Pat. No. 5,978,791, which is a
`
`continuation of U.S. Patent Application No. 08/425,160 and, as such, may not be
`
`entitled to the priority date of April 11, 1995 for all claims. This does not affect
`
`my analysis below which uses the April 11, 1995 date for each of the challenged
`
`claims.
`
`
`
`
`
`- 4 -
`
`
`
`for Bulk File Distribution,” Network Working Group Internet
`Draft (July 27, 1994) (“Moore 1994”, Ex. 1007)2
`
`(cid:120) Woodhill et al., U.S. Patent No. 5,649,196, entitled “System and
`Method For Distributed Storage Management on Networked
`Computer Systems Using Binary Object Identifiers,” filed Nov. 9,
`1995 as a continuation of application 85,596, filed July 1, 1993
`(“Woodhill”, Ex. 1005).
`
`(cid:120) Legent Software, Inc., “ESM: Product Information,” Legent
`Corporation (April 1994) (“ESM Manual”, Ex. 1026).3
`
`(cid:120) Satyanarayanan, “Scalable, Secure, and Highly Available
`
`Distributed File Access,” IEEE Computer, vol. 23, no. 5 (May
`
`1990), pp. 9–21 (“Satyanarayanan,” Ex. 1029)
`
`(cid:120) Albert Langer, “Re: dl/describe (File descriptions),” ),” article
`<1991Aug7.225159.786@newshost.anu.edu.au> in Usenet
`
`2 These three references – Browne, Browne 1994, and Moore 1994 – have
`
`substantially the same disclosure with respect to the challenged claims of the ‘280
`
`patent. Solely for the purposes of simplicity, my discussion will refer to only the
`
`Browne reference. The Browne 1994 and Moore 1994 references are equally
`
`relevant to the patentability of the challenged claims.
`
` 3
`
` I understand the ESM Manual is a printed publication, published by Legent
`
`Corporation in April 1994.
`
`
`
`
`
`- 5 -
`
`
`
`newsgroups “alt.sources.d” and “comp.archives.admin” (August 7,
`1991) (“Langer”, Ex. 1003)
`
`(cid:120) Kantor, “The Frederick W. Kantor Contents-Signature System
`Version 1.22,” FWKCS122.REF (August 10, 1993) (“Kantor”, Ex.
`1004). 4
`
`10.
`
`I have further reviewed and understand the claim charts submitted by
`
`EMC Corporation and VMware, Inc. (“Petitioner”) in the above-captioned inter
`
`partes review (Exs. 1032-1035). I understand these claim charts were submitted
`
`in an ongoing litigation involving the Petitioner and PersonalWeb Technologies
`
`LLC. In my opinion, a person of ordinary skill in the art would find that these
`
`claim charts identify representative subject matter from the identified references
`
`that teach each and every limitation of the challenged patent claims.
`
`11.
`
`I am being compensated at my normal consulting rate for my work.
`
`My compensation is not dependent on and in no way affects the substance of my
`
`statements in this Declaration.
`
`12.
`
`I have no financial interest in Petitioner. I have been informed that
`
`PersonalWeb Technologies, LLC (“PersonalWeb”) and Level 3 Communications,
`
`LLC (“Level 3”) each purport to own 50% of the ‘280 patent. I have no financial
`
`
`4 I understand Kantor is a printed publication that has been available to the public
`
`since August 1993.
`
`
`
`- 6 -
`
`
`
`interest in PersonalWeb or Level 3, and I have had no contact with either company.
`
`I similarly have no financial interest in the ‘280 patent, and have had no contact
`
`with the named inventors of the ‘280 patent: David A. Farber and Ronald D.
`
`Lachman.
`
`The ‘280 Patent
`
`13.
`
`I have reviewed and understand the overview of the ‘280 patent set
`
`forth in Section IV of the Petition for Inter Partes Review. In my opinion, the
`
`overview accurately describes that the claims of the ‘280 patent are directed to data
`
`storage systems that use “substantially unique data identifiers” – based on all the
`
`data in a data item and only the data in the data item – to identify and access data
`
`items.
`
`14.
`
`I understand that claims 36 and 38 of the ‘280 patent are being
`
`challenged in the above-referenced inter partes review.
`
`Technical Basis Underlying the Grounds of Rejections Set Forth in the
`Petition for Inter Partes Review
`
`15.
`
`I now turn to the references applied in the grounds of rejections set
`
`forth in Sections V(A–D) of the Petition for Inter Partes Review.
`
`16. For ease of reference, the following chart will be used when referring
`
`to portions of claims 36 and 38 of the ‘280 patent:
`
`
`
`
`
`- 7 -
`
`
`
`Claim 36
`
`A method of delivering a data file in a network comprising a plurality
`of processors, some of the processors being servers and some of the
`processors being clients, the method comprising:
`
`storing the data file is [sic] on a first server in the network and storing
`copies of the data file on a set of servers in the network distinct from
`the first server; and
`
`responsive to a client request for the data file, the request including a
`hash of the contents of the data file, causing the data file to be
`provided to the client.
`
`Claim 38
`
`A method of delivering a data file in a network comprising a plurality
`of processors, some of the processors being servers and some of the
`processors being clients, the method comprising:
`
`storing the data file is [sic] on a first server and storing copies of the
`data file on a set of servers distinct from the first server; and
`
`responsive to a client request for the data file, the request including a
`value determined as a given function of the contents of the data file,
`providing the data file to the client.
`
`Claim
`Portion
`[36a]
`
`[36b]
`
`[36c]
`
`Claim
`Portion
`[38a]
`
`[38b]
`
`[38c]
`
`
`
`Grounds of Invalidity for Challenged Claims 36 and 38 Based on Browne as a
`Primary Reference
`
`17.
`
`I have reviewed and understand the description of the Browne set
`
`forth in Section V(A) of the Petition for Inter Partes Review and think it
`
`
`
`- 8 -
`
`
`
`accurately summarizes its disclosure. In my opinion, a person of ordinary skill in
`
`the art would find Browne to be an enabling disclosure of the subject matter it
`
`discusses.
`
`18.
`
`I have reviewed and understand the claim chart in Exhibit 1034. In
`
`my opinion, a person of ordinary skill in the art would agree that the chart
`
`identifies and discusses representative subject matter from Browne that teaches
`
`each and every claim limitation of claims 36 and 38 of the‘280 patent.
`
`Ground 1: Browne Anticipates Challenged Claims 36 and 38
`
`19. Browne discloses claim portions [36b] and [38b]. For example,
`
`Figure 2 of Browne illustrates the steps involved in “publishing” a file on
`
`Browne’s distributed mirrored file repository:
`
`
`
`- 9 -
`
`
`
`
`
`The “Publisher”, depicted in Figure 2, makes a new file available on one or more
`
`file servers. (Browne at 4; Ex. 1002.) The publisher also informs the mirror sites
`
`of the existence of the new file, and registers the new file the search servers,
`
`including the LIFN server that clients can use to later access the file. (Browne at 4;
`
`Ex. 1002.) Browne discloses that copies of files are stored on a set of multiple
`
`servers in the network, distinct from the publisher’s file server. For example,
`
`Browne states that “copies of popular software packages may be mirrored by a
`
`number of sites to increase availability (e.g., if one site is unreachable, the software
`
`may be retrieved from a different site) and to prevent bottlenecks.” (Browne at 1–
`
`2; Ex. 1002.) Browne also discloses that “mirror copies of a file cannot be located
`
`from a URL reference, since each copy has a different URL.” (Browne at 2; Ex.
`
`1002.) Finally, as disclosed in Browne, a publisher “inform[s] mirror sites of the
`
`new or updated file.” (Browne at 4; Ex. 1002.)
`
`20. Browne also discloses claim portions [36c] and [38c]. For example,
`
`Browne discloses a Bulk File Distribution (“BFD”) package which a person of
`
`ordinary skill in the art would know is a software program running on a computer.
`
`The BFD package includes logic that determines an identifier for a corresponding
`
`data item, e.g., a file. The identifier is referred to as a <signature> and is
`
`calculated as an MD5 hash of the contents and only the contents of the data file.
`
`
`
`- 10 -
`
`
`
`(Browne at 6; Ex. 1002.)5. The entire contents of the file, and only the contents of
`
`the file, are input into the MD5 hash function, which returns a 128-bit fingerprint.
`
`(Browne at 3-4; Ex. 1002.) (“This fingerprint is a 128-bit quantity resulting from
`
`applying the MD5 function to the contents of the file”). The MD5 hash is thus a
`
`value determined as a given function of the contents of the file. Because the MD5
`
`hash is content-based, files with identical content will have the same MD5
`
`fingerprint, even if they have different names or are located on different servers.
`
`(Browne at 4; Ex. 1002.) (“The function is designed to make it computationally
`
`infeasible to find a different sequence of bytes that produces the same
`
`fingerprint….”) Browne teaches that the signature is contained within a Location
`
`Independent File Name, or LIFN, for files and sets of files in a distributed network
`
`of physical software repositories residing on different file servers. (Browne at 6;
`
`Ex. 1002.).
`
`21. Browne further discloses using a LIFN to determine whether a file is
`
`present on a cache site, or mirror site, and to access the file. (Browne at 4-5; Ex.
`
`1002.) As shown in Figure 3 below, a client program can request access to a file
`
`using the file’s LIFN. The LIFN server then uses the LIFN, which may include an
`
`5 Browne discloses that each LIFN is of the form “lifn:netlib:<signature>”,
`referencing the file access protocol (“lifn,” similar to the “http” protocol identifier
`in a URL), the server handling the request (“netlib”), and the unique MD5 hash
`used to identify a file (“<signature>”). (Id. at 4, 6; Ex. 1002.)
`
`
`
`- 11 -
`
`
`
`MD5 fingerprint, to identify the requested file, and provide a list of locations
`
`(URLs) for servers on which the file is stored. (Browne at 4-5; Ex. 1002.) The
`
`requested file is then accessed from one of the URLs using the file’s LIFN.
`
`(Browne at 5; Ex. 1002.) Alternatively, if a cache site is present in the system, it
`
`can use the LIFN to determine if it has a copy of the file and, if so, can directly
`
`provide the file to the client. (Browne at 5; Ex. 1002.) A file server maps a LIFN
`
`to a file by aliasing the ASCII form of the MD5 signatures of each of the files to
`
`actual file locations. The mapping of a LIFN to an actual file location requires
`
`examining the MD5 signatures of the plurality of files stored on the server, to
`
`match them to the LIFN in the request from the client. (Browne at 6; Ex. 1002.)
`
`
`
`
`
`
`
`- 12 -
`
`
`
`Browne also discloses that the client can determine whether a particular file is
`
`present in the system by sending a request for the LIFN to each mirror site returned
`
`by the LIFN server. (Browne at 4, 6; Ex. 1002.) In response to the request, the
`
`mirror site either returns a copy of the requested file, or, if the requested file is not
`
`present at the site, an error indicating that the file was not found. (Browne at 6;
`
`Ex. 1002.)
`
`Ground 2: Challenged Claims 36 and 38 are Unpatentable as Obvious in view of
`Browne in combination with Langer
`
`22.
`
`It would be well within routine creativity of a person of ordinary skill
`
`in the art to combine the teachings of Browne and Langer. Both references are
`
`concerned with uniquely identifying files, mirroring of files across servers, and
`
`efficient access to files. (Browne at 1–2; Ex. 1002; Langer at 3-4; Ex. 1003.)
`
`Indeed, Langer provides an express motivation to do so, by disclosing that its MD5
`
`signature provides a simple method of defining a unique identifier that can be
`
`generated locally, without requiring a registration system, and with negligible
`
`chances of collision. (Langer at 4; Ex. 1003.) A person of ordinary skill in the art,
`
`exercising ordinary creativity, would have been motivated to modify Browne to
`
`use the MD5 hash disclosed in Langer. Further, the modified BFD package that
`
`would result from using the MD5 hash function in Browne’s system would clearly
`
`meet the limitations of claims 36 and 38 of the ‘280 patent. Such combination of
`
`Browne with Langer would have been the application of Langer’s known
`
`
`
`- 13 -
`
`
`
`technique to the known device of Browne, ready for improvement, to yield the
`
`predictable result of improving file access by using a unique identifier, with
`
`negligible chances of collision, which can be generated locally.
`
`Grounds of Invalidity for Challenged Claims 36 and 38 Based on Woodhill as
`a Primary Reference
`
`23.
`
`I have reviewed and understand the description of the Woodhill set
`
`forth in Section V(B) of the Petition for Inter Partes Review and think it accurately
`
`summarizes its disclosure. In my opinion, a person of ordinary skill in the art
`
`would find Woodhill to be an enabling disclosure of the subject matter it discusses.
`
`24.
`
`I have reviewed and understand the claim chart in Exhibit 1032. In
`
`my opinion, a person of ordinary skill in the art would agree that the chart
`
`identifies and discusses representative subject matter from Woodhill that teaches
`
`each and every claim limitation of claims 36 and 38 of the‘280 patent.
`
`Ground 3: Woodhill Anticipates Challenged Claims 36 and 38
`
`25. Woodhill discloses claim portions [36b] and [38b]. For example,
`
`Woodhill discloses a network of processors in Figure 1:
`
`
`
`- 14 -
`
`
`
`
`
`(Woodhill at Fig. 1; Ex. 1005). Woodhill further discloses, for the files stored on
`
`each local computer 20, “perform[ing] two concurrent backup operations” during
`
`which it “stores a compressed copy of every binary object it would need to restore
`
`every disk drive 19 on every local computer 20 somewhere on the local area
`
`network 16 other than on the local computer 20 on which it normally resides. At
`
`the same time, the Distributed Storage Manager program 24 transmits every new or
`
`
`
`- 15 -
`
`
`
`changed binary object to the remote backup file server 12.” 6 A person of ordinary
`
`skill in the art would understand that this concurrent backup procedure operates
`
`just like the “Mirror True File” mechanism of the ‘280 patent, which is “is used to
`
`ensure that files are available in alternate locations in mirror groups or archived on
`
`archival servers.” (‘280 patent at col. 26, ll. 21-23; Ex. 1001.) Woodhill’s remote
`
`backup file is analogous to the archival servers discussed in the ‘280 patent. The
`
`archival servers have the same purpose as the remote backup file server of
`
`Woodhill, which “ensure[ed] that at least one copy of every binary object is stored
`
`and that a disaster that destroys an entire site would not destroy all copies of that
`
`site’s data.” (Woodhill at col. 9, ll. 42-44; Ex. 1005.)
`
`26. Woodhill also discloses claim portions [36c] and [38c]. For example,
`
`Woodhill discloses a local computer “initiat[ing] a restore of a randomly selected
`
`6 Woodhill describes files “as a collection of data streams” where each data stream
`
`is “a distinct collection of data within the file that may be changed independently
`
`from other distinct collections of data within the file.” (Woodhill at col. 4, ll. 14-
`
`18; Ex. 1005.) Each data stream is then divided into one or more binary objects,
`
`which are of one megabyte or less in size in Woodhill’s preferred system. (Id. at
`
`col. 4, ll. 22-26; Ex. 1005.) Accordingly, a person of ordinary skill in the art
`
`would understand that a file of less than one megabyte, with only one
`
`corresponding data stream, could thus be represented as a single binary object.
`
`
`
`- 16 -
`
`
`
`binary object identified by a Binary Object Identification Record 58.” A Binary
`
`Object Identification Record includes a “Binary Object Identifier 74 portion . . .
`
`comprised of the Binary Object Size field 64, the Binary Object CRC32 field 66,
`
`the Binary Object LRC field 68 and the Binary Object Hash field 70.” (Woodhill
`
`at col. 7, line 60 – col. 8, line 65; Ex.1005). In particular, the Binary Object Hash
`
`field 70 “is calculated against the contents of the binary object.” (Id.; Ex. 1005).
`
`The Binary Object Hash field 70 is thus a value determined as a given function of
`
`the contents of the file. A person of ordinary skill in the art would understand that
`
`a local computer initiating a restore of a binary object is a client request.
`
`27.
`
`Indeed, a person of ordinary skill in the art would find many
`
`similarities between Woodhill and the ‘280 patent. Furthermore, both Woodhill
`
`and the ‘280 patent are directed to file backup applications. The ‘280 patent notes
`
`that “[b]acking up data items in a DP system employing the present invention can
`
`be done based on the True Names of the data items. By tracking backups using
`
`True Names, duplication in the backups is prevented.” (‘280 patent at col. 36, ll.
`
`57-59; Ex. 1001.) Woodhill is similarly directed to a distributed backup system,
`
`and reduces duplicate copies of binary objects by tracking Binary Object
`
`Identifiers of previously backed up binary objects. (Woodhill at col. 1, ll. 11-17
`
`and col. 8, l. 66 – col. 9, l. 27; Ex. 1005.) As well, both Woodhill and the ‘280
`
`patent perform self-checks to ensure that backed up data remains secure and
`
`
`
`- 17 -
`
`
`
`unchanged. The ‘280 patent describes that “the system might store the True
`
`Names of all executable applications on the system and then periodically
`
`redetermine the True Names of each of these applications to ensure that they match
`
`the stored True Names.” (‘280 patent at col. 34, ll. 50-54; Ex. 1001.) Like the
`
`‘280 patent, Woodhill “is able to perform self-audits on a periodic basis to ensure
`
`that the binary objects that have been backed up can be restored.” (Woodhill at
`
`col. 18, ll. 11-13; Ex. 1005.) Woodhill does so by restoring a randomly selected
`
`binary object and ensuring that a new Binary Object Identifier calculated for that
`
`restored binary object matches the Binary Object Identifier that was previously
`
`stored. (Id. at col. 18, ll. 13-38; Ex. 1005.)
`
`Ground 4: Challenged Claims 36 and 38 are Unpatentable as Obvious in view of
`Woodhill
`
`28. As discussed above, Woodhill discloses a distributed storage
`
`management system for storing multiple backup copies of files and ultimately
`
`restoring those files. (Woodhill at col. 2, ll. 39-49; Ex. 1005.) A person of
`
`ordinary skill in the art at the time of the filing of the ‘280 patent would understand
`
`a distributed storage management system to disclose multiple servers.
`
`29. Furthermore, to the extent multiple servers are not already disclosed
`
`in Woodhill, it would be well within the routine creativity of a person of ordinary
`
`skill in the art to add an additional remote backup file server, or servers, to the
`
`system disclosed in Woodhill. Indeed, Woodhill provides an express motivation
`
`
`
`- 18 -
`
`
`
`to do so, for example, Woodhill discloses “[T] he remote backup file server 12
`
`ensures that at least one copy of every binary object is stored and that a disaster
`
`that destroys an entire site would not destroy all copies of that site’s data.” (Id. at
`
`col. 9, ll. 40-44; Ex. 1005.) Adding an additional remote backup server to
`
`Woodhill would obtain the predictable result of providing additional data security
`
`through extra redundancy, and would have been obvious to a person of ordinary
`
`skill in the art.
`
`Grounds of Invalidity for Challenged Claims 36 and 38 Based on the ESM
`Manual as a Primary Reference
`
`30.
`
`I have reviewed and understand the description of the Enterprise
`
`Storage Manager manual set forth in Section V(C) of the Petition for Inter Partes
`
`Review and think it accurately summarizes its disclosure. In my opinion, a person
`
`of ordinary skill in the art would find Woodhill to be an enabling disclosure of the
`
`subject matter it discusses.
`
`31.
`
`I have reviewed and understand the claim chart in Exhibit 1033. In
`
`my opinion, a person of ordinary skill in the art would agree that the chart
`
`identifies and discusses representative subject matter from the Enterprise Storage
`
`Manager manual that teaches each and every claim limitation of claims 36 and 38
`
`of the‘280 patent.
`
`
`
`
`
`- 19 -
`
`
`
`Ground 5: The ESM Manual Anticipates Challenged Claims 36 and 38
`
`32. The ESM Manual discloses claim portions [36b] and [38b]. For
`
`example, the ESM Manual describes a dual backup process that “protects data by
`
`completely automating the backup process, by saving the data in two locations, on
`
`the LAN and at the corporate data center, and by automatically auditing the data as
`
`it is saved.” (ESM Manual at 2-1; Ex 1026.)
`
`33. This process is depicted in Figure 2-1of the ESM Manual:
`
`
`
`34. The above figure explicitly shows three copies of the a user’s files
`
`being stored in different locations. For example, the user’s files on Server 1(first
`
`copy) are stored in the local backup of Server 2 (second copy ) and on backup
`
`tapes at the corporate data center (third copy). Alternatively, the user’s files on
`
`Server 2 (first copy) are stored in the local backup of Server 1 (second copy) and
`
`on backup tapes at the corporate data center (third copy). (ESM Manual at Fig. 2-
`
`
`
`- 20 -
`
`
`
`1; Ex 1026.) “[T]he corporate data center, which is unmatched in its ability to
`
`assimilate large volumes of data over telecommunications lines, process and verify
`
`that data, and store it on tape for later retrieval.” (ESM Manual at 2-1; Ex 1026.)
`
`A person of ordinary skill in the art at the time of the filing of the ‘280 patent
`
`would understand that the corporate data center would include a server or servers
`
`in order to process, verify, and store the data on tape for later retrieval. For
`
`example, the ESM manual discloses a “Central Site Process” that “runs on your
`
`IBM mainframe” and “provides data storage, retrieval, and verification services to
`
`all of the servers in the enterprise running ESM server” while using an “optimum
`
`combination of mainframe disk and tape.” (ESM Manual at 4-4; Ex 1026.)
`
`35. As depicted in Figure 4-1, the ESM Manual further discloses a
`
`multiple-server configuration in which a user’s files on one server are backed up to
`
`the other servers in a distributed fashion.
`
`
`
`- 21 -
`
`
`
`
`
`(ESM Manual at Fig. 4-1; Ex 1026.)
`
`36. The ESM Manual also discloses claim portions [36c] and [38c]. For
`
`example, the ESM Manual discloses “decompos[ing] files into binary objects, or
`
`BOBs.” (ESM Manual at 3-3 – 3-4; Ex 1026.) “For large files, the data portion is
`
`broken up into a number of BOBs of 1 megabyte” (Id.) Similarly, a person of
`
`ordinary skill in the art at the time of the filing of the ‘280 patent would understand
`
`the above disclosure to mean that small files of less than 1 megabyte would be
`
`processed as one BOB containing the entire contents of the file. The ESM Manual
`
`further discloses that “BOBs are not assigned names based upon the name of the
`
`
`
`- 22 -
`
`
`
`original file, an arbitrary numbering scheme, or other random criteria. Instead, the
`
`name of a BOB, the BOBID, is calculated from the contents of the BOB itself.”
`
`(ESM Manual at 3-4; Ex 1026.) The ESM Manual also discloses “BOBIDs have
`
`exactly the same length, 128 bits (16 bytes)” and each contains “… a 32-bit hash
`
`value, followed by 32 bits containing the length of the BOB in bytes.” (ESM
`
`Manual at 3-4; Ex 1026). The BOBID is thus a value determined as a given
`
`function of the contents of the file.
`
`37. The ESM Manual also discloses that “ESM's architecture enables all
`
`restore requests to arrive as BOBID requests.” (ESM Manual at 3-9; Ex 1026.)
`
`This provides an element of security because, “[t]o retrieve any data file, the user
`
`must already know the identity of the data, an identity only available to the ESM
`
`backup process.” (ESM Manual at 3-9; Ex 1026.) For example, following a
`
`request for retrieval of prior version of a file, “[o]nce the user has identified the
`
`file name and desired point in time, ESM restores the file automatically [using the
`
`file’s BOBID(s)] and notifies the user when and where it is available.” (ESM
`
`Manual at 5-3; Ex 1026.)
`
`38.
`
`Indeed, a person of ordinary skill in the art would find many
`
`similarities between the ESM Manual and the ‘280 patent. For example, both the
`
`ESM Manual and the ‘280 patent calculate content-based identifiers for data items.
`
`The ‘280 patent calculates True Names for data items using a combination of the
`
`
`
`- 23 -
`
`
`
`MD5 function (a particular hash function) and the length of the data item. (‘280
`
`patent at Figure 10(a); Ex 1001.) Similarly, the ESM Manual calculates “ the
`
`name of a BOB, the BOBID, [] from the contents of the BOB itself.” (ESM
`
`Manual at 3-4; Ex 1026.) Each BOBID is equal to “32-bit cyclical redundancy
`
`check (CRC) of the BOB, followed by a 32-bit LRC (longitudinal redundancy
`
`check, or XOR), followed by a 32-bit hash value, followed by 32 bits containing
`
`the length of the BOB in bytes.” (ESM Manual at 3-4; Ex 1026.) Furthermore,
`
`both the ESM Manual and the ‘280 patent are directed to file backup applications.
`
`The ‘280 patent notes that “[b]acking up data items in a DP system employing the
`
`present invention can be done based on the True Names of the data items. By
`
`tracking backups using True Names, duplication in the backups is prevented.”
`
`(‘280 patent at col. 36, ll. 57-60; Ex. 1001.) The ESM Manual is similarly directed
`
`to a distributed backup system that “protects data by completely automating the
`
`backup process, by saving the data in two locations, on the LAN and at the
`
`corporate data center, and by automatically auditing the data as it is saved.” (ESM
`
`Manual at 2-1; Ex 1026.) As well, both the ESM Manual and the ‘280 patent
`
`perform self-checks to ensure that backed up data remains secure and unchanged.
`
`The ‘280 patent describes that “the system might store the True Names of all
`
`executable applications on the system and then periodically redetermine the True
`
`Names of each of these applications to ensure that they match the stored True
`
`
`
`- 24 -
`
`
`
`Names.” (‘280 patent at col. 34, ll. 50-54; Ex. 1001.) Like the ‘280 patent, the
`
`ESM Manual describes that “ESM ensures restorability by performing random
`
`audits of the data as it is backed up, comparing backup data to the original file.”
`
`(ESM Manual at 2-2; Ex 1026.)
`
`Grounds of Invalidity for Challenged Claims 36 and 38 Based on
`Satyanarayanan as a Primary Reference
`
`39.
`
`I have reviewed and understand the description of Satyanarayanan set
`
`forth in Section V(D)of the Petition for Inter Partes Review and think it accurately
`
`summarizes its disclosure. In my opinion, a person of ordinary skill in the art
`
`would find Satyanarayanan to be an enabling disclosure of the subject matter it
`
`discusses.
`
`40.
`
`I have reviewed and understand the claim chart in Exhibit 1035. In
`
`my opinion, a person of ordinary skill in the art would agree that the chart
`
`identifies and discusses representative subject matter from Satyanarayanan that in
`
`combination with Langer or Kantor renders obvious each of claims 36 and 38 of
`
`the‘280 patent.
`
`41. Satyanarayanan discloses claim portions [36b] and [38b]. For
`
`example, Figure 6 of Satyanarayanan depicts a Coda file system, which maintains
`
`files on a preferred server and a group of other servers. (Satyanarayanan at 16; Ex.
`
`1029.)
`
`
`
`- 25 -
`
`
`
`
`
`(Satyanarayanan at Fig. 6; Ex. 1029.)
`
`42. Satyanarayanan also discloses most aspects of claim portions [3