Actifio, Inc.v.Delphix Corp.

Patent Trial and Appeal Board

Apr 29, 2016

13316263 (P.T.A.B. Apr. 29, 2016)

Trials@uspto.gov Paper 56
571-272-7822 Entered: April 29, 2016

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

ACTIFIO, INC.,
Petitioner,

v.

DELPHIX CORP.,
Patent Owner.
____________

Case IPR2015-00108
Patent 8,566,361 B2
____________

Before KARL D. EASTHOM, JENNIFER S. BISK, and
PATRICK R. SCANLON, Administrative Patent Judges.

SCANLON, Administrative Patent Judge.

FINAL WRITTEN DECISION
35 U.S.C. § 318(a) and 37 C.F.R. § 42.73
IPR2015-00108
Patent 8,566,361 B2

2
I. INTRODUCTION
Petitioner, Actifio, Inc., filed a Petition (Paper 1, “Pet.”) requesting an
inter partes review of claims 9–13 and 21–23 of U.S. Patent No. 8,566,361
(Ex. 1101, “the ’361 patent”) pursuant to 35 U.S.C. §§ 311–319. Patent
Owner, Delphix Corp., filed a Preliminary Response (Paper 6, “Prelim.
Resp.”). On May 1, 2015, we instituted an inter partes review as to all
challenged claims (Paper 7, “Inst. Dec.”).
After institution, Patent Owner filed a Patent Owner Response (Paper
18, “PO Resp.”), and Petitioner filed a Reply (Paper 25, “Reply”). Petitioner
relies on the Declaration of Erez Zadok (Ex. 1117) and the Supplemental
Declaration of Erez Zadok (Ex. 1137) in support of its contentions, and Patent
Owner relies on the Declaration of Prashant Shenoy, Ph.D. (Ex. 2115) in
support of its contentions.
An oral hearing was held on February 2, 2016. A transcript of the
hearing is included in the record. Paper 55 (“Tr.”).
We have jurisdiction under 35 U.S.C. § 6(b). This Final Written
Decision is issued pursuant to 35 U.S.C. § 318(a) and 37 C.F.R. § 42.73. For
the reasons that follow, we determine that Petitioner has shown by a
preponderance of the evidence that claims 9–13 and 21–23 of the ’361 patent
are unpatentable.
II. BACKGROUND
A. Related Matters
The parties indicate that the ’361 patent is at issue in Delphix Corp. v.
Actifio, Inc., Case No. 5:13-cv-04613-BLF (N.D. Cal.). Pet. 2; Paper 5, 2.
IPR2015-00108
Patent 8,566,361 B2

3
The ’361 patent is also the subject of another petition for inter partes review
filed by Petitioner—IPR2015-00100. Pet. 2; Paper 5, 2.
B. The ’361 Patent (Ex. 1101)
The ’361 patent describes systems and methods for managing databases
and lifecycle workflows based on databases. Ex. 1101, 1:12–14. More
specifically, the ’361 patent involves creating one or more virtual databases
based on a production database or another virtual database at a particular point
in time. Id. at 5:64–66. Virtual databases are created “using storage level
snapshots of production databases or clones of production databases instead of
a live production database.” Id. at 6:16–19. “A virtual database created for a
point in time is stored as a set of files that contain the information of the
database as available at that point in time.” Id. at 6:31–34.
Figure 1 of the ’361 patent is reproduced below.

Figure 1 illustrates how information may be copied from production database
systems 110 to database storage system 100. Id. at 7:12–15. To create a
IPR2015-00108
Patent 8,566,361 B2

4
virtual database, database storage system 100 creates files representing the
corresponding information from a production database system at a given point
in time. Id. at 7:50–52. More particularly, database storage system 100
retrieves and stores information from production database systems 110. Id. at
9:38–39. Database storage system 100 then exposes the files to virtual
database system 130 using file sharing system 120. Id. at 7:53–55. Virtual
database system 130 includes database server 230 and operating system 240.
Id. at 7:60–8:6, Fig 2(a).
C. Illustrative Claim
Of the challenged claims in the ’361 patent, claims 9 and 21 are
independent. Claim 9 is illustrative of the claims at issue:
9. A method for creating a data warehouse and a data
mart from a database, the method comprising:
linking a first database, wherein the linking comprises
receiving information identifying the first database;
loading the first database at multiple points in time,
wherein the loading comprises:
receiving database blocks for different point-in-time
copies of the first database, and
storing the database blocks in a storage system;
provisioning a first virtual database (VDB) to an
operational data store system, wherein provisioning the first VDB
comprises:
creating a set of files linked to the stored database
blocks associated with the first database on the storage
system, and
mounting the set of files to the operational data store
system;
performing extract, transform, and load (ETL) operations
on the data in the first virtual database and storing the output of
IPR2015-00108
Patent 8,566,361 B2

5
the ETL operations in a second database in a data warehouse
system; and
receiving database blocks for different point-in-time copies
of the second database.
Ex. 1101, 36:34–56.
D. The Prior Art
Petitioner relies on the following prior art:
1. JAWAHAR LAL, ROGER SANDERS & JEREMY BRUMER,
DB2: CLONING A DATABASE USING NETAPP FLEXCLONE™
TECHNOLOGY, TR-3460 (2006) (“Sanders”) (Ex. 1104);
2. John K. Edwards et al., FlexVol: Flexible, Efficient File
Volume Virtualization in WAFL, PROC. OF THE 2008 USENIX
ANNUAL TECHNICAL USENIX CONF. 129 (2008) (“Edwards”)
(Ex. 1105);
3. DIPESH PATEL & GENEROSA LITTON, NETAPP WHITE
PAPER, RAPID DATABASE DEVELOPMENT AND DEPLOYMENT, WP-
7014-0307 (2007) (“Patel”) (Ex. 1106); and
4. DARRIN CHAPMAN, MIKE FEDERWISCH, & CHUCK
DUFRESNE, SNAPMIRROR® BEST PRACTICES GUIDE, TR-3446
(2006) (“Chapman”) (Ex. 1107).
E. Instituted Grounds of Unpatentability
We instituted the instant inter partes review on the following grounds
of unpatentability:
References Basis Claims Challenged
Sanders, Edwards, and Patel § 103 9–12, 21, and 22
Sanders, Edwards, Patel, and
Chapman
§ 103 13 and 23

IPR2015-00108
Patent 8,566,361 B2

6
III. ANALYSIS
A. Printed Publication—Sanders, Chapman, and Patel
Patent Owner contests that Sanders, Chapman, and Patel are prior art
“printed publications” in accordance with 35 U.S.C. §§ 102 and 311(b). PO
Resp. 1‒5. We look to the underlying facts to make a legal determination as
to whether a document is a printed publication. Suffolk Techs., LLC v. AOL
Inc., 752 F.3d 1358, 1364 (Fed. Cir. 2014). The determination of whether a
document is a “printed publication” under 35 U.S.C. § 102(b) involves a case-
by-case inquiry into the facts and circumstances surrounding its disclosure to
members of the public. In re Klopfenstein, 380 F.3d 1345, 1350 (Fed. Cir.
2004). Public accessibility is a key question in determining whether a
document is a printed publication and is determined on a case-by-case basis.
Suffolk Techs., 752 F.3d at 1364. To qualify as a printed publication, a
document “must have been sufficiently accessible to the public interested in
the art.” In re Lister, 583 F.3d 1307, 1311 (Fed. Cir. 2009).
Initially, we note our disagreement with Patent Owner’s contention that
Petitioner cannot rely upon evidence not submitted with the Petition to show
that Sanders, Chapman, and Patel are prior art. PO Resp. 2‒4. In Patent
Owner’s view, Petitioner must make out a prima facie case of unpatentability
in its Petition, which includes the substantive element of Sanders, Chapman,
and Patel being publicly accessible and prior art. Id. at 3–4. That position,
however, is not informed by the difference between the threshold for
instituting a trial (35 U.S.C. § 314(a)) and proving unpatentability of a claim
in trial (35 U.S.C. § 316(e)). As noted by our reviewing court, “there is a
significant difference between a petitioner’s burden to establish a ‘reasonable
likelihood of success’ at institution, and actually proving invalidity by a
IPR2015-00108
Patent 8,566,361 B2

7
preponderance of the evidence at trial.” TriVascular, Inc. v. Samuels, 812
F.3d 1056, 1068 (Fed. Cir. 2016) (quoting 35 U.S.C. § 314(a) and comparing
§ 316(e)).
Based on the information presented in the Petition and Patent Owner’s
Preliminary Response, we determined there was a reasonable likelihood that
Petitioner would prevail in its challenge that included Sanders and Patel and
its challenge that included Sanders, Patel, and Chapman. Inst. Dec. 27; see 35
U.S.C. § 314(a) (threshold for instituting inter partes review); see also 37
C.F.R. § 42.108(c) (“The Board’s decision [on Institution] will take into
account a patent owner preliminary response where such a response is filed.”).
Patent Owner did not challenge the prior art status of any of the applied
patents or publications in its Preliminary Response. Patent Owner, in fact,
stated that it had “disclosed to the Patent Office every NetApp feature that
Petitioner now cites in the Petition” and that Edwards, Sanders, Chapman, and
Patel “were published years apart.” Prelim. Resp. 56, 58 (emphasis added).
We do not mean to suggest that a patent owner must raise any “printed
publication” issues in a preliminary response in order for the Board to
consider such issues in the preliminary proceeding phase. In this case,
however, based in part on the information in Patent Owner’s Preliminary
Response and in part on the printed dates and the lack of indicia of
confidentiality or internal, non-public distribution in Sanders, Chapman, and
Patel, we determined that Petitioner had met its burden for a threshold
showing to proceed to trial.
Patent Owner also argues that Petitioner cannot rely on declarations
filed after the Petition. PO Resp. 2–4. These declarations include two
declarations provided by Louis Hernandez (Ex. 1120) and Joseph Ortiz (Ex.
IPR2015-00108
Patent 8,566,361 B2

8
1128) in response to objections by Patent Owner1 and a Supplemental
Declaration by Mr. Hernandez (Ex. 1147) filed with its Reply. Our rules
authorize serving supplemental evidence in response to an objection. 37
C.F.R. § 42.64(b)(2). Patent Owner lacks a foundation to complain that
evidence has been produced in response to its objections. Petitioner also
relies, properly, on the supplemental evidence in its Reply, as evidence in
reply to Patent Owner’s arguments in its Response that Sanders, Chapman,
and Patel are not printed publications.
Turning to the substance of Exhibit 1120, Mr. Hernandez testifies that
he is currently employed by Petitioner, was employed by NetApp from 2004
to 2009, and was a NetApp customer from 2000 to 2004. Ex. 1120 ¶¶ 1, 2, 4.
Mr. Hernandez testifies further that for most of his time at NetApp, as a
Systems Engineer, he was responsible for marketing NetApp’s products and
services to numerous customers, prospective customers, business partners,
and/or alliances. Id. ¶ 3. “During the 2000-2009 time-frame, to support its
marketing efforts, it was NetApp’s standard practice to publish technical
reports, white papers, and product manuals or guides to customers, potential
customers, business partners, and alliances.” Id. ¶ 6. “These documents were
published, according to standard practice, as of the month and year that
appeared on the face of the documents.” Id. Mr. Hernandez testifies that he
has personal knowledge of and recognizes Sanders, Chapman, and Patel, and
that they were published during his tenure at NetApp. Id. ¶¶ 7, 9, 11, 12, 17.

1 Exhibit 1128 is expunged at Petitioner’s request. We do not further discuss
the Exhibit. We address Patent Owner’s motions to exclude these documents
in a separate section, below.
IPR2015-00108
Patent 8,566,361 B2

9
Patent Owner argues Mr. Hernandez does not declare that Sanders,
Chapman, or Patel was “publicly accessible.” PO Resp. 2. Patent Owner
submits:
Even if it was NetApp’s “standard practice” to provide its
documents to its “customers, potential customers, business
partners and alliances,” that does not establish that these
documents were available to the public, but instead shows
at most that they were only available to a subset of entities
affiliated with NetApp.
Id. at 3.
Petitioner replies with a Supplemental Declaration from Mr.
Hernandez. Reply 6‒7 (citing Ex. 1147). Mr. Hernandez testifies that he uses
the terms “publish” or “published” as referring to documents being publicly
distributed to customers, potential customers, business partners, and alliances
as of the month and year that appeared on the face of the documents, non-
confidentially. Ex. 1147 ¶ 5. According to Mr. Hernandez, NetApp had more
than two hundred Systems Engineers and other sales personnel during the
relevant timeframe (id. ¶ 4) and that technical reports, white papers, product
manuals, and product guides were freely distributed to support its marketing
efforts (id. ¶ 7). Mr. Hernandez testifies further that it was important for
NetApp to date the documents accurately so that customers and potential
customers could understand if a specific document accurately reflected
features for specific versions of NetApp’s products or if a document was
outdated or updated to reflect more current features. Id. ¶ 10. Further,
Petitioner provides evidence that by 2007 there were more than 94,000
IPR2015-00108
Patent 8,566,361 B2

10
NetApp systems deployed and the company had thousands of customers in
138 countries. Reply 5; Ex. 1156, 3.2
As part of routine discovery (37 C.F.R. § 42.51(b)(1)(ii)), Patent Owner
had the opportunity to cross-examine Mr. Hernandez during Patent Owner’s
first discovery period but elected not to. Patent Owner cross-examined Mr.
Hernandez in its second discovery period regarding the testimony in his
Supplemental Declaration. We have considered Patent Owner’s Motion for
Observation on Cross-Examination Testimony of Mr. Hernandez (Paper 38)
and Petitioner’s Response (Paper 44), insofar as they relate to public
accessibility of Sanders, Chapman, and Patel. We acknowledge the potential
for bias in Mr. Hernandez’s testimony as a present employee of Petitioner.
We find, however, the testimony in Mr. Hernandez’s Declarations as to public
accessibility of Sanders, Chapman, and Patel to be credible. As an earlier
panel of the Board has found, in a proceeding involving a different patent and
different parties, documents such as Sanders, Chapman, and Patel are dated
technical documents or whitepapers, having no indication of being mere drafts
or internal papers, each of which is “a type of document whose very purpose
is public disclosure.” Veeam Software Corp. v. Symantec Corp., Case
IPR2014-00089, slip op. at 14 (PTAB Apr. 25, 2014) (Paper 9).
Finally, Petitioner also submits a declaration it says was produced in
response to Patent Owner’s evidentiary objections. Reply 3–4. Petitioner
provided the declaration from the office manager of the Internet Archive
(Wayback Machine). Id. (citing Ex. 1123). Sanders is dated April 30, 2006

2 We find that Exhibit 1156, a NetApp Form 10-K SEC filing, was properly
submitted by Petitioner as evidence in rebuttal to Patent Owner’s public
accessibility challenge in its Response.
IPR2015-00108
Patent 8,566,361 B2

11
and is designated “TR-3460” (or Technical Report 3460). Ex. 1104, 1.
Chapman is dated April 2006 and is designated “TR-3446” (or Technical
Report 3446). Ex. 1107, 1. According to the testimony regarding how the
Internet Archive works (Ex. 1123 ¶¶ 3‒5), we find the evidence indicates that
Sanders and Chapman were available on NetApp’s commercial website on, or
at least as early as, November 22, 2006. Ex. 1123, 148, 231. Exhibit 1123
indicates that Sanders and Chapman were, thus, “sufficiently accessible to the
public interested in the art.” See In re Lister, 583 F.3d at 1311.
“A given reference is ‘publicly accessible’ upon a satisfactory showing
that such document has been disseminated or otherwise made available to the
extent that persons interested and ordinarily skilled in the subject matter or art
exercising reasonable diligence, can locate it.” SRI Int’l, Inc. v. Internet Sec.
Sys., Inc. 511 F.3d 1186, 1194 (Fed. Cir. 2008) (quoting Bruckelmyer v.
Ground Heaters, Inc., 445 F.3d 1374, 1378 (Fed. Cir. 2006)). Petitioner
asserts that the level of ordinary skill is not disputed in this case and that
“there is no difference between customers and potential customers of NetApp
and a [person of ordinary skill in the art].” Reply 5. Mr. Hernandez testifies
that NetApp’s customers and potential customers were entities or individuals
who were “interested in data storage technology, data management
technology, database storage and management technology, and related
technologies.” Ex. 1147 ¶ 3. As discussed above, Mr. Hernandez also
testifies NetApp had more than two hundred Systems Engineers and other
sales personnel during the period of 2004–2009 who distributed NetApp’s
technical documents (including Sanders, Chapman, and Patel) to “thousands”
of customers and potential customers. Id. ¶ 4. Hence, the record supports
Petitioner’s contention that Sanders, Chapman, and Patel were distributed to
IPR2015-00108
Patent 8,566,361 B2

12
persons interested and ordinarily skilled in the subject matter of database
technology at the time of its publication. See Reply 6–7.
In view of the foregoing considerations, we find that Petitioner has
established, by a preponderance of the evidence, that Sanders, Chapman, and
Patel were sufficiently disseminated to persons of ordinary skill interested in
database technology to be deemed “publicly accessible” at least more than one
year before October 21, 2009, the earliest possible priority date of the ʼ361
patent. See Ex. 1101, at (63). Therefore, on this record, we determine
Sanders, Chapman, and Patel qualify as prior art printed publications under 35
U.S.C. § 102(b).
B. Claim Construction
We interpret claims of an unexpired patent using the broadest
reasonable construction in light of the specification of the patent in which they
appear. See 37 C.F.R. § 42.100(b). On this record and for purposes of this
Decision, we determine that only the claim terms addressed below require
express construction.
1. “database block”
In the Institution Decision, we preliminarily construed the term
“database block,” as “a unit of data used by a database.” Inst. Dec. 11. In its
Response, Patent Owner asserts that this definition is impermissibly broad.
PO Resp. 16–25. Patent Owner’s proposed construction is “a unit of data
used by a database which comprises a specific number of bytes stored in the
storage, a portion of which stores metadata associated with the unit of data.”
Id. at 16 (emphasis added). Petitioner, on the other hand, agrees with our
conclusion in the Decision to Institute. Reply 8–12.
IPR2015-00108
Patent 8,566,361 B2

13
a) Metadata Associated with the Database Block
The main dispute between the parties centers on whether a database
block must necessarily include metadata. Patent Owner asserts that it does
(PO Resp. 16), and Petitioner disagrees (Pet. 8–9). We agree with Petitioner.
We begin our analysis by considering the language of the claims
themselves. Phillips v. AWH Corp., 415 F.3d 1303, 1314, 1315 (Fed. Cir.
2005) (en banc). The term “metadata” is not recited in any of the claims of
the ’361 patent. U.S. Patent No. 8,150,808 B2 (“the ’808 patent”),3 however,
includes two claims that recite “metadata”—dependent claims 32 and 33,
which are not challenged in any proceeding, of which we are aware, currently
before the Board. These claims depend indirectly from claim 1 of the ’808
patent and expressly recite “metadata of database blocks.” Thus, had the
patentees intended to limit “database blocks” recited in the claims of the ’361
patent to require metadata, they demonstratively could have done so by
explicitly modifying the disputed term with “metadata,” but did not.
Moving to the specification, Patent Owner asserts that the following
passage in the Summary section of the ’808 patent defines the term “database
block.”
A database block is a unit of data used by a database and
comprises a specific number of bytes stored in the storage.
A database block can also be referred to as a page. A
portion of the database block stores metadata associated
with the database block.
PO Resp. 16 (quoting Ex. 2102, 2:7–12).

3 Patent Owner describes the ’808 patent as being related to the ’361 patent.
Prelim. Resp. 28.
IPR2015-00108
Patent 8,566,361 B2

14
The first phrase in the cited passage above explicitly defines the term,
by stating “[a] database block is a unit of data used by a database.”4 Ex. 2102,
2:7–8. Although the sentence following shortly thereafter states that a
database block “stores metadata,” that sentence by itself is insufficient to limit
the disputed term by requiring the unclaimed “metadata” feature because it
does not state unambiguously that all “database blocks” must include
metadata. See Liebel-Flarsheim Co. v. Medrad, Inc., 358 F.3d 898, 906 (Fed.
Cir. 2004) (construing a claim term broadly because “[n]o statement in the
written description [ ] constitute[d] a limitation on the scope of the invention”)
(quoting Brookhill-Wilk 1, LLC. v. Intuitive Surgical, Inc., 334 F.3d 1294,
1301 (Fed. Cir. 2003)). Further, the cited passage also does not exclude the
possibility of some database blocks not having any metadata. See id. at 908
(passages in the Summary of the Invention section of a patent did not limit the
scope of the invention because the passages, “although focusing on the use of
the invention in conjunction with pressure jackets, do not disclaim the use of
the invention in the absence of a pressure jacket”). More importantly, the rest
of the ’361 patent is consistent with a definition of database blocks that does
not require metadata in all database blocks.
Patent Owner asserts, citing certain portions of the ’361 patent and the
Declaration of Prashant Shenoy, Ph.D. (Ex. 2115), that if a database block
does not include metadata, the disclosed system would not work as described.

4 The first sentence also states a database block comprises “a specific number
of bytes stored in the storage.” For the reasons discussed below, we find this
addition is not part of the explicit definition but, rather, represents
embodiments within the defined term. Patent Owner also fails to explain how
to interpret this particular phrase or how it presents a material issue related to
the prior art.
IPR2015-00108
Patent 8,566,361 B2

15
PO Resp. 17–18. For example, Patent Owner argues that a database block
must include metadata because the disclosed system analyzes the metadata of
each block to store only the incremental changes made to the production
database, which is “one of the main functions” of the claimed system (id. at 17
(citing Ex. 1101, 7:31–35, 41–44, 14:43–46; Ex. 2115 ¶¶ 86–87)) essential to
achieving “a main purpose” of the invention—“to efficiently provide virtual
databases . . . without proliferating redundant copies of database data” (id. at
17–18 (citing Ex. 2115 ¶ 87)). Patent Owner also asserts that metadata is
required in each database block in order to map the block to a database file
and a location within that file. Id. at 18 (citing Ex. 1101, 15:27–31).
Patent Owner’s argument is unpersuasive because the argued
advantages or purposes are not recited features of the claims. Moreover, a
claim is not required to encompass all of the advantages or purposes of the
invention. See Howmedica Osteonics Corp. v. Wright Med. Tech., Inc., 540
F.3d 1337, 1345 (Fed. Cir. 2008) (“An invention may possess a number of
advantages or purposes, and there is no requirement that every claim directed
to that invention be limited to encompass all of them.”) (citations omitted).
Moreover, Patent Owner focuses its arguments on the streaming
embodiment, however, Patent Owner does not show that the file sharing
embodiment requires this approach, where the database files to be copied have
defined boundaries and known structures, and the database blocks stored in
the files can be accessed directly. See, e.g., Ex. 1101, 14:10–14 (“the
production system library [] includes code to analyze the structures of the files
of the database stored in the data store [] and also includes code to process
metadata associated with database blocks stored in the data store”) (emphases
IPR2015-00108
Patent 8,566,361 B2

16
added),5 7:12–18 (“FIG. 1 illustrates one embodiment for how information
may be copied from a production database to a database storage system . . .
using a file sharing system. . . . In some embodiments information may be
copied from storage level snapshots of production databases.”) (emphases
added). When file sharing, as opposed to streaming, is used, the database file
on the production system can be accessed and copied directly by “mounting
the production DB data store” on the database storage system without packing
and unpacking the database blocks of the database file into and out of data
streams. Ex. 1101, 8:58–65.
Considering “the context of the surrounding words” to the term
“database block” in the claims, which “must be considered in determining the
ordinary and customary meaning” of the disputed term, ACTV, Inc. v. Walt
Disney Co., 346 F.3d 1082, 1088 (Fed. Cir. 2003), independent claims 9 and
21 recite receiving database blocks for different point-in-time copies of the
first database and storing the database blocks in a storage system. But the
claims do not say anything about a particular method of receiving point-in-
time copies, whether by streaming or by file sharing. Hence, the claims are
not limited to either embodiment, and, therefore, the streaming embodiment
cited by Patent Owner does not limit the claims.
Further, even looking at the streaming embodiment, we are not
persuaded that metadata is required by database blocks. Patent Owner points
only to a single streaming embodiment (PO Resp. 18), in which the data
stream may include unnecessary database blocks, such as the blocks that did
not change since the last point-in-time copy was transmitted, which are

5 The phrase “metadata associated with database blocks” implies any
metadata need not be in the database blocks.
IPR2015-00108
Patent 8,566,361 B2

17
eliminated after the data stream is received at the database storage system by
analyzing metadata for each database block. Ex. 1101, 14:43–60. In an
alternative embodiment, which is not addressed by Patent Owner, the
unchanged blocks are eliminated at the production system and never sent to
the database storage system. Id. at 14:64–15:3 (“In other [sic] embodiment,
some or all of the unnecessary blocks may be eliminated while the data
stream is built by the production system library []. In this embodiment, the
data stream . . . is reduced in size resulting in efficient communication
between” the production system and the database storage system (emphases
added)). Patent Owner does not explain why metadata must be included in
each database block to achieve the incremental update function in this second
embodiment. Hence, packing metadata within database blocks may be
involved in some streaming embodiments, but nothing in the specification
indicates it is required for the incremental update function. Therefore, nothing
in the specification indicates that copying database files by streaming data is
the essence of the claimed invention rather than a preferred embodiment,
which may not be read into the claims “absent clear disclaimer in the
specification.” In re Am. Acad. of Sci. Tech Ctr., 367 F.3d 1359, 1369 (Fed.
Cir. 2004).
Therefore, in view of the entire disclosure of the ’361 patent and the
plain language of the claims, we find that the intrinsic record before us does
not justify limiting the term “database block” by reading in the “metadata”
limitation not found in the claims. See Liebel-Flarsheim Co., 358 F.3d at 908.
Patent Owner also argues additional evidence supports its proposed
construction requiring data blocks to include metadata. For example, citing
the testimony of Dr. Shenoy, Patent Owner asserts that all database
IPR2015-00108
Patent 8,566,361 B2

18
management systems mentioned in the ’361 patent, such as Oracle and IBM
DB2 (Ex. 1101, 6:5–9), require metadata in database blocks. PO Resp. 19–20
(citing Ex. 2115 ¶¶ 41–46, 84–85). In the paragraphs cited by Patent Owner,
Dr. Shenoy discusses various documents describing the database systems
listed in the ’361 patent, including Oracle, Sybase, Microsoft SQL Server, and
IBM DB2, and testifies that these database systems all require metadata in
database blocks.6 Ex. 2115 ¶¶ 43–46, 84–85 (citing testimony from
Petitioner’s declarant, Dr. Zadok in another related case). Relying upon the
testimony of Dr. Shenoy, Patent Owner also asserts that “[i]t is well known to
one skilled in the art that a database block necessarily includes metadata.” PO
Resp. 19 (citing Ex. 2115 ¶¶ 84–85); see also id. at 17 (“[T]hat database
blocks include metadata is consistent with the ordinary meaning of the term in
the database field and comports with how the term is used by every major
database system provider.”) (citing Ex. 2115 ¶¶ 41–46, 83). In his
Declaration, Dr. Shenoy discusses a treatise on database systems (Ex. 2118,
“Database Systems”) and testifies that it is generally understood that a
database block or a page will include metadata. Ex. 2115 ¶ ¶ 40–42 (citing
Ex. 2118, 29, 31).
The evidence presented by Patent Owner in support of its argument—
i.e., testimony of experts and documents describing the commercially
available database systems listed in the ’361 patent—is properly characterized
as extrinsic evidence. Such extrinsic evidence is “less significant than the
intrinsic record in determining the legally operative meaning of claim

6 However, neither Patent Owner’s brief nor Dr. Shenoy’s testimony discuss
the MYSQL database system mentioned in the ’361 patent.
IPR2015-00108
Patent 8,566,361 B2

19
language.” Phillips, 415 F.3d at 1317 (citations omitted) (internal quotation
marks omitted).
Petitioner argues that Patent Owner does not provide any evidence that
database blocks require metadata in database systems other than relational
database systems. See Reply 10–11. Petitioner correctly points out that the
’361 patent specification states that the disclosed invention “can be used for
any database.” Id. at 10 (quoting Ex. 1101, 6:15–16). Further, the ’808
patent states that “[a] database may be implemented using a database model,
for example, a relational mode, object model, hierarchical mode or network
model” and goes on to state that “the techniques disclosed can be used for any
database.” Ex. 2102, 5:8–15 (emphasis added). Thus, we agree with
Petitioner that the ’361 patent does not limit its disclosed invention to
relational database technology.
Similarly, the portion of the Database Systems treatise relied upon by
Dr. Shenoy appears to describe features of a relational database system, not
characteristics common to all databases in general. See, e.g., Ex. 2118, 29
(“Collections such as relations are usually represented by placing the records
that represent their data elements in one or more blocks.”) (emphasis added),
31 (“Records representing tuples of a relation are stored in blocks of the disk .
. . there is a block header holding information such as: . . . Information about
which relation the tuples of this block belong to.”) (first and last emphases
added).
Relying on the testimony of Dr. Zadok submitted with its Reply
(Ex. 1137), Petitioner asserts that, in other type of database systems, such as
Google’s BigTable database, metadata is stored separately from the database
IPR2015-00108
Patent 8,566,361 B2

20
blocks.7 Reply 10–11 (citing Ex. 1148, 4; Ex. 1137 ¶ 24). Petitioner also
argues, citing the testimony of Dr. Zadok, that a “flat file database,” such as a
comma separated value (CSV) file used by spreadsheet applications, does not
have metadata in its blocks. Reply 10 (citing Ex. 1137 ¶ 24). Therefore,
Petitioner asserts that the ordinary meaning of database block does not require
including metadata. See id., Ex. 1137 ¶ 22.
We agree with Petitioner that all extrinsic evidence Patent Owner relies
upon to argue ordinary meaning appears to be directed to relational databases.
Because the challenged claims plainly recite “database,” and, therefore, are
not limited to relational databases, Patent Owner’s evidence, even assuming it
shows relational databases all have database blocks that require metadata,
does not establish an ordinary and customary meaning of “database blocks”
recited in the claims. Therefore, we find Patent Owner’s extrinsic evidence
regarding ordinary meaning does not overcome the intrinsic record of this
case. See Phillips, 415 F.3d at 1318 (“[A] court should discount any expert
testimony that is clearly at odds with the claim construction mandated by the
claims themselves, the written description, and the prosecution history, in

7 Petitioner argues the definition of a “database” proposed by Patent Owner’s
expert, “a collection of data that is organized so that it can be easily accessed,
managed or updated,” encompasses many types of databases other than the
relational databases. See Reply 10 (citing Ex. 2115 ¶ 33). In addition to the
disclosure that “the techniques disclosed can be used for any database,” (Ex.
1101, 6:14–15) the ’808 patent provides another broad description—“[a]
database comprises data stored in a computer for use by computer
implemented applications” (Ex. 2102, 4:67–5:1). Under either description or
Mr. Shenoy’s testimony, we agree with Petitioner that a “database”
encompasses many types of databases other than the relational databases,
including Google’s BigTable, flatfiles, and spreadsheet databases, and “any
database.”
IPR2015-00108
Patent 8,566,361 B2

21
other words, with the written record of the patent.”) (citations and internal
quotation marks omitted). Furthermore, we are persuaded by Petitioner’s
evidence that certain well-known databases, such as Google’s BigTable
database, do not require metadata in database blocks, and, therefore, find that
the evidence in this case does not establish a widely accepted meaning of
“database block” in the field of database systems that requires metadata to be
necessarily included in a database block.
b) “Application Level” v. “Storage Level”
Patent Owner also argues, citing the passages in the ’808 patent
describing the streaming embodiment, that a “database block” must include
metadata because “the ’361 patent describes an application level system that
understands database blocks and uses metadata in those database blocks to
determine whether to store database blocks and where to store them.” PO
Resp. 25 (emphasis added); see also id. at 14–16 (making the same argument)
(citing Ex. 1101, 14:43–60; Ex. 2115 ¶¶ 86, 87, 89, 90). Patent Owner
asserts, therefore, a “database block” cannot be a “storage level unit of data.”
Id. at 22 (emphasis added). The passages of the ’808 patent cited by Patent
Owner, however, describe solely the streaming embodiment and, therefore, do
not limit the challenged claims as discussed above.
Patent Owner also asserts that the system of the ’361 patent “operates at
the application level” because the system uses APIs (application program
interfaces) to copy database blocks. Id. at 14 (citing Ex. 1101, 8:48–53; Ex.
2115 ¶¶ 69, 88). Petitioner argues that the ’361 patent never mentions
“application level,” teaching instead that “the system ‘can create a virtual
database using storage level snapshots of production databases’” and
“retriev[ing] ‘database blocks from storage level snapshots.’” Reply 7 (citing
IPR2015-00108
Patent 8,566,361 B2

22
Ex. 1101, 6:17–18, 9:35) (emphases added by Petitioner). We agree with
Petitioner’s argument and further note that the ’361 patent also discloses
copying of database blocks from storage level snapshots of a database. See
Ex. 1101, 7:12–18 (“FIG. 1 illustrates one embodiment for how information
may be copied from a production database to a database storage system . . .
using a file sharing system. . . . In some embodiments information may be
copied from storage level snapshots of production databases.”) (emphases
added). More importantly, Patent Owner does not explain why the fact that
the system of the ’361 patent uses APIs to access database blocks necessarily
requires the database blocks to include metadata. For example, Patent Owner
does not identify, nor do we discern, anything in the written description that
explains why the APIs cannot access and process metadata that is stored
separately from database blocks.
Patent Owner further relies upon the testimony of Dr. Shenoy and
external documents to argue that a “database block” is an “application level”
entity, which is different from a “file system block,” an “operating system
block,” or a “storage level” unit of data. PO Resp. 22–25 (citing Ex. 2115
¶¶ 27–33). In particular, Patent Owner asserts that an Oracle document (Ex.
2104, “Oracle Database Concepts 11g Release 2 (11.2)”) explains this “well-
known distinction.” Id. at 22–23 (citing Ex. 2104, 250, Fig. 12-5). Patent
Owner further argues “neither are file system blocks accessible to a database
nor are database blocks accessible to the file system.” Id. at 24 (citing Ex.
2115 ¶¶ 95–99). Therefore, Patent Owner argues a construction of the term
“database block” that encompasses a storage unit block is incorrect. Id. at 25.
As a threshold matter, we note that the Oracle document is dated May,
2014, which is several years later than the filing date of the ’361 patent.
IPR2015-00108
Patent 8,566,361 B2

23
Hence, the Oracle document is not probative of what was known to one of
ordinary skill in the art at the time of the invention. More importantly, as
discussed above, we find the written description of the ’361 patent does not
support a limiting construction of the term “database block” based on the
purported “application level” versus “storage level” distinction. Further, we
do not find any disclosure in the specification that describes or discusses
whether a “database block” is an “application level” or a “storage level” entity
or construct. Hence, we find the testimony of Dr. Shenoy and the external
documents discussed by Patent Owner regarding the “application level”
versus “storage level” distinction to be largely divorced from the written
description, and, therefore, insufficient to overcome the plain language of the
claims. See Kara Tech. Inc. v. Stamps.com Inc., 582 F.3d 1341, 1348 (Fed.
Cir. 2009); Phillips, 415 F.3d at 1318.
c) A Specific Number of Bytes Stored in the Storage
In the Institution Decision, we found that database blocks of the ’361
patent need not have a specific number of bytes stored in storage because an
empty database block would not have metadata or “a specific number of bytes
stored in the storage.” Inst. Dec. 10–11. Patent Owner continues to argue that
this is a requirement of the proper construction of “database blocks.” PO
Resp. 16.
Patent Owner does not explain how to interpret “a specific number of
bytes stored in the storage.” For example, it is not clear if “a specific number”
is a constant. In another argument, Patent Owner states that “[d]atabase
blocks may be any number of different sizes independent of the size of the file
system data blocks which may ultimately store the database data.” PO Resp.
24–25 (citing Ex. 2115 ¶ 95). Patent Owner also states that an IBM DB2
IPR2015-00108
Patent 8,566,361 B2

24
database page can be up to 32KB in size, spanning 8 WAFL data blocks, most
of which will not contain any metadata for the database page. Id. at 25 (citing
Ex. 1108 at 28).
According to Patent Owner’s arguments, a database block may be “any
number of different sizes.” Therefore, the database blocks of the ’077 patent
need not have a constant specific number of bytes stored in storage.
d) “Database Block” v. “File”
Next, Patent Owner argues that the construction proposed by Petitioner
(Pet. 9) and preliminarily adopted in our Institution Decision (Inst. Dec. 11)—
“a unit of data used by a database”—is improper because it would equate a
“database block” with a “file.” PO Resp. 21–22. Citing dictionaries, Patent
Owner argues that the ordinary meaning of “file” is “a unit of data.” Id.
(citing Ex. 2125, 3; Ex. 2126, 3). Patent Owner further asserts that our
preliminary construction incorrectly encompasses a log file described in the
Specification because a log file is also “a unit of data used by database.” Id. at
22.
Petitioner argues, citing the deposition testimony of Patent Owner’s
expert, Dr. Shenoy, that Patent Owner is incorrect because a file has a name
associated with it but a database block, as construed by Petitioner, would not.
Reply 12 (citing Ex. 1131, 261:19–20, 262:8–9). We agree with Petitioner
that defining a database block as “a unit of data used by a database” does not
equate a database block with a file.
First, a file is not necessarily “a unit of data used by a database”
because not all files are used by a database. Conversely, the phrase “a unit of
data used by a database” is not sufficient to describe a file because a file has
additional properties or characteristics, such as a unique file name. In fact, the
IPR2015-00108
Patent 8,566,361 B2

25
full definitions of a “file” provided in the Patent Owner cited dictionaries,
including the portions Patent Owner had omitted when quoting the sources in
the Patent Owner Response, show that a file name is a required attribute of a
file. See Ex. 2125, 3 (defining “file” as “[a] document or other collection of
information stored on a disk and identified as a unit by a unique name”), Ex.
2126, 3 (defining “file” as “[a] collection of data or information that is stored
as a unit in the computer under a single name, called the file name”) (the
portions omitted by Patent Owner indicated with bold emphasis). Hence, the
fact that a definition of a “database block” reads on some “files,” such as a log
file used by a database, does not mean the definition equates the two terms
because a file has additional properties or characteristics, such as a file name.
e) Conclusion
In summary, we find nothing in the intrinsic record, including the
written description and the language of the claims, that justifies limiting the
term “database block” by reading in the “metadata” or “specific number of
bytes” limitations not found in the claims. Furthermore, no extrinsic
evidence, including the testimony of experts, presented in this case is
sufficient to overcome the plain claim language of the term “database block.”
Therefore, on this record, we construe “database block” to mean “a unit of
data used by a database.” See Phillips, 415 F.3d at 1316 (“The construction
that stays true to the claim language and most naturally aligns with the
patent’s description of the invention will be, in the end, the correct
construction.”) (citation omitted).
IPR2015-00108
Patent 8,566,361 B2

26
2. “virtual database”
In the Institution Decision, we preliminarily construed the term “virtual
database,” to mean “a set of database files capable of being read from and
written to, and capable of being mapped to physical addresses for stored
database blocks.” Inst. Dec. 16. Petitioner agrees with this construction.
Reply 12–17. Patent Owner, on the other hand, asserts that this definition
reads out the “virtual” requirement. PO Resp. 28. Patent Owner, instead
proposes that the term should be interpreted to mean “a set of files to which a
database server can read and write such that the physical implementation of
the database files is decoupled from the logical use of the database files by the
database server.” Id. at 25–26.
To analyze Patent Owner’s proposed construction, Patent Owner
defines a virtual database as “a set of files” or “database files” and modifies
the definition with two phrases, each expressing a concept involving a
database server: (1) “a database server can read and write” to the set of files,
and (2) “the physical implementation of the database files is decoupled from
the logical use of the database files by the database server.” We address each
proposed modifying phrase in turn.
a) Database Server Can Read and Write
Patent Owner argues that the point of the first proposed modifying
phrase is to “require something of the files . . . that they are of a form readable
and writable by a database server.” PO Resp. 29. In other words, the main
thrust of Patent Owner’s proposed phrase is that a database server can read
from and write to a set of files.
Patent Owner also argues that requiring a “database server” to be able
to read “a set of files” is necessary to clarify that “a set of files” in its
IPR2015-00108
Patent 8,566,361 B2

27
proposed definition must be “database files.” Id. Patent Owner’s argument is
unpersuasive because the same objective can be accomplished without
inserting the extraneous term by directly stating a virtual database must be “a
set of database files.” In fact, the ’361 patent describes that a virtual database
comprises “database files.” See, e.g., Ex. 1101, 6:19–22 (“The virtual
databases are ‘virtual’ in the sense that the physical implementation of the
database files is decoupled from the logical use of the database files by a
database server.”), 40–41 (“A virtual database may be created on a database
server by creating the database files . . . .”) (emphases added). Patent Owner
appears to argue this is insufficient because “[t]o be a database file, a database
server must be able to read, write and understand its contents.” PO Resp. 29
(citing Ex. 2115 ¶ 109). We are not persuaded by this argument because it
does not comport with the written description of the ’361 patent.
First, the ’361 patent provides a description of a “database” as follows:
“[a] database comprises data stored in a computer for use by computer
implemented applications.” Ex. 1101, 6:1–2 (emphasis added). Furthermore,
the ’361 patent announces that the disclosed invention “can be used for any
database.” Id. at 6:14–15. Hence, a database file may be used by an
application or a program but not necessarily a “database server,” which is a
server-side program in a client-server system or network. See id. at 6:2–4,
31:43–46.
Furthermore, the ’361 patent describes that various programs that are
not a database server can read and write to a database or database file and
understand its content. For example, the virtual database manager program at
the database storage system can create, read, and write to read/write files
representing a virtual database. See Ex. 1101, 7:3–5; Ex. 2102, 17:57–63,
IPR2015-00108
Patent 8,566,361 B2

28
18:27–31. Other programs of the database storage system, such as the point-
in-time copy manager and the storage allocation manager, can create, read,
and write to database files based on database blocks received from the
production system. See Ex. 1101, 10:46–63, 11:13–31.
Therefore, contrary to Patent Owner’s contention, the written
description of the ’361 patent shows that a database does not require a
database server reading and writing to the database. Patent Owner’s proposed
construction would improperly import a particular embodiment from the
specification by reading in the operation of a database server to impermissibly
narrow the meaning of the term “virtual database.” As discussed above,
features of particular embodiments may not be read into the claims “absent
clear disclaimer in the specification.” Am. Acad. of Sci. Tech Ctr., 367 F.3d at
1369.
On the other hand, the specification shows that the readable and
writable characteristics are inherent attributes of a virtual database. See, e.g.,
Ex. 1101, 5:64–6:1 (“the virtual databases can then be individually accessed
and modified as desired”); Ex. 2102, 18:27–28 (“FIG. 10 indicates how
storage efficient copies are made to create a read/write file structure
representing a VDB.”), 17:57–63 (“The virtual database manager [] identifies
[] the recent most PIT copy associated with time Tj . . . The read/write file
structure [] is created [] by making storage efficient copies of the database
blocks in the identified PIT copy.”), 18:12–15 (“The virtual database manager
[] sends . . . handles to the read/write file structure to the associated virtual
database system 130.”) (emphases added). Thus, a virtual database must
comprise readable and writable database files. However, the phrase “database
files” need not be further modified to indicate that the files are intended for
IPR2015-00108
Patent 8,566,361 B2

29
use by a computer program because the concept is already included in the
definition of “database” provided in the specification, as discussed above.
Therefore, the term “virtual database” is properly construed, to have as the
base phrase of its definition, “a set of readable and writable database files.”
b) The Physical Implementation of the Database Files
is Decoupled from the Logical Use of the Database Files
by the Database Server
Patent Owner relies on the following sentence in the ’361 patent to
support the second proposed modifying phrase—the “decoupled” aspect of its
proposed construction: “virtual databases are ‘virtual’ in the sense that the
physical implementation of the database files is decoupled from the logical
use of the database files by a database server.” PO Resp. 26 (quoting Ex.
1101, 6:19–22). Patent Owner asserts this sentence explicitly defines the use
of the word “virtual” in the term “virtual database.” Id. However, the ’361
patent does not describe expressly what it means for “the physical
implementation of the database files [to be] decoupled from the logical use of
the database files by a database server.” Citing the testimony of Dr. Shenoy,
Patent Owner asserts the cited sentence describes one of the key features of
the invention of the ’361 patent, which is applying the general concept of
virtualization—“the decoupling of a computing process from the platform on
which it operates”—to databases. Id. (emphases added) (citing Ex. 2115 ¶¶
23–26, 101). Petitioner asserts Patent Owner confuses “platform
virtualization” with “database virtualization” or “virtualizing data.” See Reply
14–15. We agree with Petitioner that “platform virtualization” is unrelated to
the concept of “virtual database” as claimed and described in the ’361 patent.
IPR2015-00108
Patent 8,566,361 B2

30
In his Declaration, Dr. Shenoy testifies that the Java programming
language running on a Java Virtual Machine as described in an Oracle
document explains the general concept of platform virtualization. See
Ex. 2115 ¶¶ 24–25 (citing Ex. 2103, 451–452, Fig. 24-7). Patent Owner’s
discussion of Java Virtual Machine illustrates the basic problem with Patent
Owner’s argument—that it is directed to the features or functions of software,
not a database. For example, Dr. Shenoy explains that the “decoupling [of] a
computing process from the platform on which it operates” is achieved by the
Java Virtual Machine, which “is software that emulates the operation and
interface of a physical processor.” See Ex. 2115 ¶¶ 23–24 (emphasis added).
The ’808 patent also describes that a “virtual machine” is “provided by
platform virtualization software” or “server virtualization software.” Ex.
2102, 7:16–19 (emphases added). Further, the ’361 patent describes a Java
Virtual Machine as part of production database system 110 (Ex. 1101, 9:13–
15), which is separate from virtual database 220 stored on database storage
system 100, as discussed above. Hence, Patent Owner’s discussion of the
Java Virtual Machine (PO Resp. 29 n.9) may relate to database servers or
other software of the disclosed system, but it is not directed to a virtual
database, which is a set of files as expressly recited in the claims and
described in the specification (see, e.g., Ex. 2102, 6:49–51; Fig. 2a) and
therefore is not persuasive even as an analogy.
Other examples of “decoupling” Dr. Shenoy discusses similarly relate
to the function or operation of various software or programs, not database
files. For instance, Dr. Shenoy testifies that, because the system of the ’361
patent operates at the “application level” using APIs, the database storage
system of the invention does not depend on and need not concern itself with
IPR2015-00108
Patent 8,566,361 B2

31
the underlying details of the physical storage system. Ex. 2115 ¶ 69 (quoting
Ex. 1101, 8:48–53). The cited portion of the ’361 patent describes, however,
the APIs are provided by “production system library 385” and “vendor
interface module 335,” which are programs or software code modules residing
at the production system. See, e.g., Ex. 1101, 8:53–56 (“An example of a
vendor interface module is the program code of a database server provided by
vendor ORACLE that implements RMAN APIs.”) (emphasis added), 14:10–
12 (“the production system library 385 includes code to analyze the structures
of the files of the database stored in the data store 350”) (emphasis added).
Dr. Shenoy similarly testifies that “each database (130c)” depicted in
Figure 1 of the ’361 patent is “virtual” (Ex. 2115 ¶ 102; see also PO Resp. 26
(citing Ex. 2115 ¶ 102)) when, in fact, Figure 1 shows the part identified by
Patent Owner is Virtual Database System 130(c), not the virtual database,
which is stored in a different system, namely, Database Storage System 100.
As described in Figure 2a of the ’361 patent, virtual database system 130 is
separate from and external to virtual database 220 stored on database storage
system 100. Further, the ’361 patent describes that virtual database system
130 includes no database or database files but, rather, comprises a database
server and a VDB system library, both of which are software modules or
programs. See Ex. 1101, 13:3–13 (“A virtual database system 130 includes a
database server 360 and a VDB system library 380. . . . The VDB system
library 380 contains program code for processing requests sent by the
database storage system 100.”) (emphasis added), Fig. 3. Hence, Patent
Owner appears to conflate the function of the software of a virtual database
system with the meaning of “a virtual database,” which is a set of files that
IPR2015-00108
Patent 8,566,361 B2

32
resides in a database storage system that is separate from the virtual database
system.
Petitioner argues that Patent Owner’s platform virtualization argument
is misguided because it conflates the meaning of a “virtual database” with the
disclosed embodiments of production database system and the virtual database
system. Reply 15–16. We agree with Petitioner, but our analysis underscores
a related salient point, which is that Patent Owner’s arguments are directed to
the software part of the disclosed system that is separate from and external to
a virtual database. As described in the ’361 patent, a database system
generally comprises a database, which is data stored in storage, and database
software, such as a database server or other program that accesses the
database. See Ex. 1101, 6:1–4; 8:36–65. Dr. Shenoy similarly describes a
database system as comprising a database and database software.
A database system is typically a computer that includes the
stored data itself, i.e., the database, as well as database
management software, often shortened to DBMS. A
DBMS includes software for accessing the data in the
database and modifying that data to add, remove or change
stored data. A DBMS may run as or include an
autonomous program which handles client requests to the
database and which is often called a database server.
Ex. 2115 ¶ 34 (emphases added). Hence, the platform virtualization or the
storage virtualization or abstraction argued by Patent Owner as demonstrating
what is meant by the phrase “the physical implementation of the database files
is decoupled from the logical use of the database files by a database server” is
a function or feature of the software part of a database system—be it Java
Virtual Machine, a database server, or various software modules providing
APIs—which is separate from and external to a database, i.e., the data part of
a database system. Therefore, contrary to Patent Owner’s argument,
IPR2015-00108
Patent 8,566,361 B2

33
virtualization of a computing platform or storage by software is unrelated to
and cannot be equated to virtualization of data, i.e., the subject matter claimed
by the term “virtual database.”
In effect, Patent Owner’s platform virtualization argument calls for
improperly importing the function or operation of the software of the
disclosed system into the meaning of a virtual database, which is a set of files
separate from and external to the software part. If patentee intended to claim
the function or operation of the software of the disclosed system, it could have
done so by explicitly claiming the subject matter. Not having done so, Patent
Owner may not import the array of functions and intelligence embodied in the
described software into a single term “virtual database” to impart a very
different meaning to the claims from what is indicated by the plain language
of the claims and the written description of the patent. Patent Owner asserts
that, by discounting Patent Owner’s platform virtualization argument, the
preliminary construction reads out the “virtual” requirement from the term
“virtual database.” PO Resp. 28 (citing Inst. Dec. 16). Patent Owner’s
argument is unpersuasive because our refusal to read the functions and
features of the software modules into the meaning of “virtual database” does
not equate reading out the word “virtual” from the term. On the contrary, as
discussed below, the preliminary construction is consistent with the language
of the claims and the written description of the ’361 patent.
Turning to the claim language and the written description, the
challenged independent claims recite, with emphasis added, provisioning “a
first virtual database” including creating “a set of files linked to the stored
database blocks.” Similarly, the Specification describes as follows:
IPR2015-00108
Patent 8,566,361 B2

34
A virtual database may be created on a database server by
creating the database files for the production database
corresponding to the state of the production database at a
previous point in time, as required for the database server. The
files corresponding to the virtual database are made available to
the database server using a file sharing mechanism, which links
the virtual database to the appropriate database blocks stored on
the storage system.
Ex. 1101, 6:40–47 (emphasis added); see also Ex. 2102, 2:24–27 (making
essentially the same statement). Hence, the database files of a virtual database
are linked or mapped to the database blocks associated with the database
being virtualized.
Looking to the Specification for further clarification, Figure 10 of the
’361 patent depicts “how . . . to create a read/write file structure representing a
VDB.” Ex. 1101, 19:27–28. When the VDB file structures are created, “the
blocks V11, V12, . . . , V25 may be implemented as pointers to the actual
database block that stores the data.” Id. at 19:47–49 (emphasis added). The
’361 patent further describes:
For example, V11 represents the information in block F1 and
since the block F1 was never updated during copies made at time
T1 and T2, V11 points at F11. V12 represents the information in
block F2 and since F2 was updated at time T1, V12 points at the
block F22. Similarly, V13 corresponds to block F3 that was
updated at time T2 and points at the block F33.
Id. at 19:49–55 (emphases added). Hence, blocks V11, V12, etc. of a virtual
database do not contain ordinary data, but, instead, point to the location or
address of the actual database blocks stored in a physical storage device.
Another example implies that writing to a pointing block in a database file,
such as V11, which is stored at one location, actually results in writing to a
copy of the original block of data F11 at another physical location:
IPR2015-00108
Patent 8,566,361 B2

35
For example, if the virtual database system 130 writes to the
block V11, space is allocated and block F11 copied to the
allocated block. Hence, the original copy of the block F11 is
maintained as a read only copy and the virtual database system
130 is allowed to write to a copy of the appropriate database
block created specifically for the virtual database system 130.

Id. at 20:51–57. The ’361 patent further describes “[s]ince the [virtual
database file] structure 1050 illustrated in FIG. 10, structure 1150 illustrated
in FIG. 11, or structure 1350 illustrated in FIG. 13 are read/write structures,
the virtual database system 130 is allowed to read from these structures as
well as write to them.” Id. at 20:44–48. In addition, “[a] database block that
is not written to by virtual database systems 130 may be shared by several
virtual database systems without being copied for a specific virtual database
systems 130.” Id. at 20:64–67.
Thus, the ’361 patent indicates that the database files are “virtual” in the
sense that they create the illusion of allowing data to be written to pointing
blocks in the virtual database files, but the system actually writes the data
elsewhere to another physical location specified by the pointing blocks such
as V11 and V12.
Figure 12 also shows how database blocks may be shared by file
structures created for several virtual databases. Ex. 1101, 19:63–65. File
structures 1050 and 1150 include blocks V11–V14 and U11–U14,
respectively. Id. at 19:65–20:8. Each of these blocks is implemented as a
pointer to the actual database block storing the relevant data. Id. Thus,
several blocks in the virtual database may share the same database blocks—
V11 and U11 both point to the same copy of F11. Thus, the ’361 patent
indicates that the database files are “virtual” in the sense that they create the
IPR2015-00108
Patent 8,566,361 B2

36
illusion of storing data, but the system actually stores the data elsewhere in
another physical location (F11) specified by the pointing blocks such as V11
and U11. The ’361 patent also describes the procedure for writing to a shared
database block, such as V11: “For example, if the virtual database system 130
writes to the block V11, space is allocated and block F11 copied to the
allocated block.” Id. at 20:51–53.
Based on the disclosure in the ’361 patent discussed above, “the logical
use of database files” can be understood as reading from or writing to
database blocks by reading from or writing to virtual database file structures
that point to the actual database blocks. Hence, as described in the ’361
patent, decoupling of physical implementation of the database files, i.e., actual
database blocks stored on a physical storage device, from the logical use of
the database files, i.e., accessing the database blocks through virtual database
files, is accomplished by using pointers to map blocks in virtual database files
to physical addresses for database blocks stored in physical storage devices.
In fact, although not discussed in the Patent Owner Response, Dr.
Shenoy acknowledges that Figure 10 illustrates the “decoupling of the
physical storage of the database from its logical use” in part, because
“[c]ommon database blocks are shared between different point-in-time copies.
This is imperceptible to the . . . user.” Ex. 2115 ¶ 63. Similarly, Dr Shenoy
testifies that “the physical storage for the virtual database is decoupled from
the logical view of the database . . . by mapping blocks captured at various
points in time to [virtual database files]” (id. ¶ 107) (emphasis added).
Nonetheless, Patent Owner argues that the preliminary construction is
erroneous because database files of any database, including a source database,
can be “mapped to physical addresses for stored database blocks.” PO Resp.
IPR2015-00108
Patent 8,566,361 B2

37
28 (citing Ex. 2115 ¶ 109–110). Apparent inconsistencies in Dr. Shenoy’s
testimony aside, Patent Owner’s argument is unpersuasive because the ’361
patent expressly describes that a source database can be a virtual database.
See Ex. 1101, 20:17–18. To the extent that Patent Owner is arguing the
preliminary construction does not distinguish a non-virtual source database
from a virtual database, the concern can be addressed by expressly indicating
that the stored database blocks are associated with another database since an
ordinary, non-virtual database consists of its own database blocks and would
not be created by mapping or pointing to stored blocks of a different database.
This tracks the claim language of the challenged claims, which define how to
create a VDB, which in turn, tracks Dr. Shenoy’s testimony. See Ex. 2115 ¶¶
63, 107.
Furthermore, although not argued by the parties, it is not necessary to
require that the mapped database blocks are associated with a point-in-time
copy of a database because the challenged independent claims expressly recite
virtual database files wherein “at least some of the database blocks are
associated with multiple point-in-time copies of the source database.” See
Digital-Vending Servs. Int’l, LLC v. Univ. of Phoenix, 672 F.3d 1270, 1274–
75 (Fed. Cir. 2012). Therefore, based on the foregoing discussion, the term
“virtual database” is construed as “a set of readable and writable database files
capable of being mapped to physical addresses for stored database blocks
associated with another database.”
C. Asserted Ground Based on Sanders, Edwards, and Patel
Petitioner challenges claims 9–12, 21, and 22 as obvious under
35 U.S.C. § 103 over Sanders, Edwards, and Patel. Pet. 4, 31–56. To support
IPR2015-00108
Patent 8,566,361 B2

38
this assertion, Petitioner relies on the Declaration of Dr. Erez Zadok (Ex.
1117).
1. Overview of Sanders
Sanders describes a system “to clone a DB2 database quickly and
easily.” Ex. 1104, 1. “Database cloning is [a] process by which you can
create an exact copy of a DB2 database.” Id. at 3. The disclosed NetApp
system uses FlexClone and SnapMirror technologies in a combined manner.
Id. at 8. This combined technology allows administrators to clone a
production FlexVol database system as a writable FlexClone database on
another storage system. Id. at 8–9. Specifically, FlexClone provides point-in-
time copies of the production database. Id. at 3. Further, “[a] FlexClone
volume is a writable point-in-time image of a FlexVol volume or another
FlexClone volume.” Id. Stated differently, “[t]he clone database is a frozen
image of the database file system at the time of the clone creation. If
necessary, the primary database can be restored from the snapshot created for
the clone; or applications can point directly to the clone database.” Id. at 6.
A FlexClone volume “uses space very efficiently, allowing both the
original FlexVol volume and the FlexClone volume to share common data,
storing only the data that changes between the original volume and the clone.”
Id. at 3. Clones can be created on the same or different storage systems. Id. at
6. The SnapMirror technology provides FlexClone volumes to be produced at
different destinations: “A SnapMirror source and its corresponding destination
can reside on the same storage system or on two separate storage systems that
are miles apart.” Id. at 3. Figure 3 of Sanders is reproduced below:
IPR2015-00108
Patent 8,566,361 B2

39

Figure 3 depicts using the combined SnapMirror and FlexClone technology to
transmit point-in-time clone copies of a production database to a destination
storage system. See id. at 8.
Sanders also describes mounting a clone database on a database server.
See id. at 14 (“In order to access the clone database, you need to mount the
clone volumes to a database server. . . . [Y]ou can mount the clone volume by
executing the following command on the database server: mount
[MountPoint].”).
2. Overview of Edwards
Edwards describes creating virtual instances of WAFL (write anywhere
file layout) file systems; read-only versions, called “FlexVol” volumes (Ex.
1105, 14)8 and writeable versions, called “FlexClone” volumes (id. at 014–
015). A FlexVol volume “break[s] this tight bond between volumes and their
underlying storage” by hiding a file system, which spans a pool of storage,
and creating externally visible volumes inside the files of the hidden file
system. Id. at 011. This “introduces a level of indirection, or virtualization,

8 When referencing Exhibit 1105, we refer to the pagination inserted by
Petitioner.
IPR2015-00108
Patent 8,566,361 B2

40
between the logical storage space used by a volume and the physical storage
space.” Id.
Edwards discloses that writable versions of the FlexVol volumes,
FlexClones, are particularly useful in database environments—“for example,
it is often desirable to make writable copies of a production database for
development or test purposes.” Id. at 014. To create a FlexClone volume,
WAFL creates a copy of vol_info block of the Snapshot copy on which the
clone is based. Id. at 015. The FlexClone volume, thus, inherits pointers to
the complete file system stored in the original Snapshot copy. Id.
3. Overview of Patel
Like Sanders and Edwards, Patel describes NetApp’s database cloning
technology. Ex. 1106, 1. Patel discloses performing ETL (extract, transform,
load) processes in the NetApp environment. Id. at 4, Fig. 2.
4. Independent claims 9 and 21
Regarding independent claim 9, Petitioner argues that “Sanders
discloses a method for creating a data warehouse and a data mart from a
database.” Pet. 31. In particular, Petitioner asserts that Sanders discloses
receiving information identifying a first database—and, thus, linking the first
database—because the reference discloses “creating a snapshot copy of a
source volume containing a first IBM DB2 database on a source storage
system and then mirroring the snapshot copy from the source volume to a
destination volume in a destination storage system.” Id. at 32 (citing
Ex. 1104, 22–23; Ex. 1117 ¶ 121).
Petitioner asserts that Sanders discloses loading the first database at
multiple points in time because execution of Sanders’s “snapmirror initialize”
IPR2015-00108
Patent 8,566,361 B2

41
command “results in the creation of a point-in-time (snapshot) copy of the
source FlexVol volume at the time the command is executed.” Id. at 33–34
(citing Ex. 1117 ¶ 123, Ex. 1104, 22). According to Petitioner, Edwards
discloses that the “data blocks” or “blocks” of Sanders are units of data used
by a database (i.e., “database blocks”). Id. at 34–35 (citing Ex. 1105, 21).
Petitioner then asserts that “Sanders in view of Edwards discloses loading the
first database at multiple points in time where the loading comprises receiving
database blocks for different point-in-time copies of the first database residing
in a FlexVol volume, and storing those database blocks in a destination
storage system.” Id. (citing Ex. 1117 ¶ 130). Petitioner concludes that one of
ordinary skill in the art would have known to combine the disclosures of
Sanders with the teachings of Edwards “in order to discern the units of data
used by a database in WAFL.” Id. (citing Ex. 1117 ¶ 129).
Petitioner asserts that Sanders discloses “provisioning” a first virtual
database by making it “available to a system through the use of FlexClone to
create a clone of the first database on FlexVol volumes and through mounting
the clone to a database server.” Id. at 36 (citing Ex. 1117 ¶ 131). Petitioner
acknowledges, however, that Sanders does not disclose explicitly creating a
set of files for the FlexClone volume. Id. at 37–38. Instead, Petitioner relies
on Edwards as disclosing this claimed subject matter. Id. at39. Petitioner
specifically argues that “Edwards explains that ‘[c]reating a clone volume is a
simple process. WAFL creates the files required for a new FlexVol volume’”
and “[a] container file for the new clone volume is created and seeded ‘with a
vol_info block that is a copy of the vol_info block of the snapshot copy on
which the clone is based.’” Id. at 38 (citing Ex. 1105, 15). Petitioner also
argues that Edwards discloses that “the cloned volume inherits all the pointers
IPR2015-00108
Patent 8,566,361 B2

42
or links of the snapshot copy that point to the underlying data blocks.” Id.
(citing Ex. 1105, 15, Ex. 1117 ¶ 135). Petitioner concludes that one of
ordinary skill in the art “would have looked to Edwards to understand the
specifics of the FlexClone creation process of how FlexClone creates files
linked to stored database blocks on the storage system.” Id. at 40 (citing
Ex. 1117 ¶ 139).
Petitioner argues that Sanders discloses “performing extract and
transform operations on the data in the clone database,” and “storing the
output of the extracted and transformed data in a database in a data warehouse
system.” Id. at 41–42 (citing Ex. 1104, 5; Ex. 1117 ¶¶ 144, 145). Petitioner
acknowledges, however, that “Sanders does not expressly disclose performing
‘load’ operations, or use the acronym ‘ETL’ to refer to extract, transform and
load operations performed on the data in the clone database.” Id. at 42 (citing
Ex. 1117 ¶ 146). To meet this limitation, Petitioner asserts that Patel discloses
the use of a clone database for ETL operations. Id. (citing Ex. 1106, 4, Fig. 2;
Ex. 1117 ¶ 146). Petitioner further contends that one of ordinary skill in the
art “would have looked to Patel to understand how to use NetApp
technologies to improve conventional data warehouse architectures by using
FlexClone technology.” Id. at 44 (citing Ex. 1117 ¶ 148).
Finally, Petitioner relies on Patel as disclosing receiving database
blocks for different point-in-time copies of a second database. Id. at 45 (citing
Ex. 1117 ¶ 149). According to Petitioner, Patel discloses using “SnapMirror
technology to replicate a database onto a ‘separate, often remote, storage
system’” and that “important data, such as the data in a data warehouse
database, is ‘routinely copied or migrated to multiple locations to protect
against all types of loss.’” Id. (citing Ex. 1106, 9; Ex. 1117 ¶ 149). Petitioner
IPR2015-00108
Patent 8,566,361 B2

43
also argues that “Sanders discloses that a ‘SnapMirror source and its
corresponding destination can reside on the same storage system or on two
separate storage systems that are miles apart’” (id. (citing Ex. 1104, 3)), such
that “both Patel and Sanders disclose that SnapMirror can be used to transmit
snapshots from a source volume to a destination volume (on the same storage
system) for storage on the destination volume” (id. (citing Ex. 1117 ¶ 149)).
Petitioner further asserts that Patel discloses that a snapshot is a point-in-time
image of data and “SnapMirror uses ‘Snapshot technology to send only
changed data blocks to the mirrored storage location.’” Id. at 45–46 (citing
Ex. 1106, 8, 9; Ex. 1117 ¶ 150) (emphasis by Petitioner removed). Based on
these arguments, Petitioner concludes that “Patel and Sanders disclose
receiving database blocks for different point-in-time copies of the second
database on a SnapMirror destination, such as a destination volume in the
storage system.” Id. at 46 (citing Ex. 1117 ¶ 150).
Independent claim 21 is directed to a “computer program product
having a non-transitory computer-readable storage medium storing computer-
executable code for creating a data warehouse and a data mart from a
database.” Ex. 1101, 38:33–36. Petitioner notes that the computer executable
code comprises instructions to “perform steps essentially identical to those
recited in Claim 9.” Pet. 53. Thus, Petitioner asserts that, “for the same
reasons that Sanders, in view of Edwards and Patel, discloses and renders
obvious Claim 9, Sanders, in view of Edwards and Patel, also discloses and
renders obvious the computer program product elements of Claim 21.” Id. at
55 (citing Ex. 1117 ¶ 172).
IPR2015-00108
Patent 8,566,361 B2

44
a) “linking a first database, wherein the linking
comprises receiving information identifying the first
database”
Patent Owner argues that Petitioner relies solely on Sanders’s
disclosure of a SnapMirror initialize command as disclosing this linking
limitation, but this command identifies a NetApp volume rather than a source
database. PO Resp. 37 (citing Ex. 2115 ¶ 170). According to Patent Owner,
“there is no necessary correspondence between a volume and a database,” and
“an identification of a volume is not an identification of a database.” Id. at 38
(citing Ex. 2115 ¶ 172). Patent Owner also argues that “in addition to
‘receiving information identifying the first database,’” the claim language
requires “actual ‘link[ing]’ of that first database.” Id. Patent Owner contends
that Sanders does not disclose such “link[ing].” Id. at 38–39.
In response, Petitioner argues that Sanders explicitly discloses a
correspondence between a volume and a database. Reply 19. In particular,
Petitioner asserts that Sanders’s SnapMirror initialize command identifies a
source volume named “dbdata,” and “dbdata” is “used to store the production
database table data.” Id. (citing Ex. 1104, 6, 22). Relying upon the testimony
of Dr. Zadok, Petitioner contends that “[t]here is no reason for a [person of
ordinary skill in the art] to create a NetApp volume with more than one
database.” Id. at 19–20 (citing Ex. 1137 ¶ 58). Regarding Patent Owner’s
“actual linking” argument, Petitioner argues that the claim language defines
“linking” as “receiving information identifying the source database” and does
not require “linking” to including anything more than receiving the
information. Id. at 18 (citing Ex. 1101, 35:42–43, Ex. 1137, ¶ 57.
We find a preponderance of the evidence demonstrates that Sanders
discloses the linking limitation. We are not persuaded that, as Patent Owner
IPR2015-00108
Patent 8,566,361 B2

45
argues, an identification of a volume cannot be an identification of a database.
Rather, we credit Dr. Zadok’s testimony that “there is no reason for a [person
of ordinary skill in the art] to think that a volume designated for and
containing a database does not identify a database.” Ex. 1137 ¶ 58. We also
are persuaded by Petitioner’s argument that the claim language does not
require “linking” to include anything more than “receiving information
identifying the first database.” Thus, we are not persuaded by Patent Owner’s
argument that Sanders fails to disclose “actual linking.”
b) “database blocks of the first database”
Patent Owner argues that the limitations of receiving “database blocks
for different point-in-time copies of the first database” and storing “the
database blocks in a storage system” are not rendered obvious by the
combination of references asserted by Petitioner “under any proposed
construction.” PO Resp. 39.
Patent Owner first argues that the proposed combination of references
does not render obvious the “database blocks” limitations even under
Petitioner’s proposed construction, which, as discussed above, we adopt. Id.
at 39–43. According to Patent Owner, “Petitioner contends that WAFL data
blocks [of Edwards] are the claimed ‘database blocks.’” Id. at 39 (citing Pet.
34–35). Patent Owner argues that Edwards’s WAFL data blocks are not
equivalent to database blocks because they are not “used by a database.” Id.
Instead, Patent Owner asserts that no database has knowledge of the WAFL
data blocks “existence, where they are located, or how to interact with them.”
Id. at 40 (citing Ex. 2115 ¶ 131).
Patent Owner’s argument, however, is premised on the previously
discussed “application level” versus “storage level” distinction—the
IPR2015-00108
Patent 8,566,361 B2

46
assumption that “[d]atabase servers only use database blocks which are data
structures that can be accessed or used at the application level” and “only the
database server,” but not the WAFL file system, “has knowledge of and
manipulates the database block as an independent unit of data.” Id. at 40–41
(citing Ex. 2115 ¶¶ 132, 133, 142). As discussed above, we reject this
premise as divorced from the written description of the ’361 patent.
Patent Owner next argues that the asserted references do not render
obvious the “database blocks” limitations under Patent Owner’s proposed
construction. Id. at 43–46. This argument is not persuasive, however,
because we do not adopt Patent Owner’s proposed construction, as discussed
above.
For the above reasons, we conclude that a preponderance of the
evidence demonstrates that a person of ordinary skill would have found the
claimed database blocks obvious over the data blocks disclosed by Edwards.
c) “point-in-time copies of the first database”
Patent Owner argues that Petitioner’s identification of the Snapshot
copies disclosed by Sanders as the claimed point-in-time copies is wrong. Id.
at 46. Patent Owner contends that “when a copy of a FlexVol volume is
created by SnapMirror, it is the exact replica of the current state of a FlexVol
volume” and “SnapMirror does not create or load a plurality of point-in-time
copies of the FlexVol volume it is cloning, it is merely cloning the current
state of the data that is within the volume.” Id. at 47 (citing Ex. 2115 ¶¶ 151,
158–160, 164). Patent Owner further argues that “merely cloning the current
state of the volume is fundamentally different than the invention of the ’361
patent,” and SnapMirror “cannot be used to accumulate any ‘database blocks’
IPR2015-00108
Patent 8,566,361 B2

47
reflecting the state of the source at different points-in-time.” Id. at 48–49
(citing Ex. 2115 ¶¶ 160–164).
In response, Petitioner argues that while Patent Owner “asserts that ‘a
Snapshot copy must exist in both the source and destination volume to create
a clone from the destination volume,’” Patent Owner does not explain how
this statement could mean that database blocks for different point-in-time
copies are not received and stored. Reply 18 (citing PO Resp. 48–49).
We are not persuaded by Patent Owner’s arguments. The crux of these
arguments appears to be that, because a Snapshot copy allegedly is an exact
copy of the current state of the volume, SnapMirror does not utilize database
blocks from different points in time to create a Snapshot copy. The claim
language, however, does not require database blocks for different point-in-
time copies to be utilized in this manner. Rather, the claim language merely
requires receiving and storing database blocks for different point-in-time
copies. We agree with Petitioner that the combined teachings of Sanders and
Edwards disclose receiving and storing database blocks for different point-in-
time copies. Sanders discloses that “[t]he SnapMirror initialize command
creates a Snapshot copy of the source volume and transfers data from the
source to the destination volume,” such that the transferred data is received
and stored at the destination volume. Ex. 1104, 22. We also credit Dr.
Zadok’s testimony that a Snapshot copy is a point-in-time copy. Ex. 1117
¶¶ 44, 123. Thus, each Snapshot copy would result in transferred data (i.e.,
database block) for different point-in-time copies being received and stored.
Lastly, even if Patent Owner is correct that a given Snapshot copy uses only
database blocks from the same point in time, we agree with Petitioner that this
IPR2015-00108
Patent 8,566,361 B2

48
does not mean that database blocks for different point-in-time copies are not
received and stored.
We conclude that a preponderance of the evidence demonstrates that a
person of ordinary skill would have found the limitation of loading the first
database at multiple points in time, including receiving and storing database
blocks for different point-in-time copies, obvious.
d) “provisioning a first virtual database”
Patent Owner first argues that the proposed combination of references
does not render obvious “provisioning a first virtual database” even under
Petitioner’s proposed construction, which, as discussed above, we adopt. PO
Resp. 49–52. Specifically, Patent Owner argues that Sanders’s FlexClone
volumes are not equivalent to virtual databases because they are not “a set of
database files.” Id. at 50. Patent Owner asserts that inodes are not database
files because “they are not capable of being mapped to physical addresses for
stored database blocks,” but instead are “collections of inodes which may
point to file system data blocks.” Id. at 51 (citing Ex. 2115 ¶ 149).
Patent Owner’s argument, however, is premised on the previously
discussed definition of “database block.” As discussed above, we reject
Patent Owner’s proposed construction.
Patent Owner next argues that the asserted references do not render
obvious the “provisioning a first virtual database” limitation under Patent
Owner’s proposed construction. Id. at 52–53. This argument is not
persuasive, however, because we do not adopt Patent Owner’s proposed
construction, as discussed above.
We conclude that a preponderance of the evidence demonstrates that a
person of ordinary skill would have found the claimed “provisioning a first
IPR2015-00108
Patent 8,566,361 B2

49
virtual database” obvious over FlexClone volumes disclosed by the Sanders
and Edwards.
e) “creating a set of files linked to the stored database
blocks”
Patent Owner argues that “Petitioner points to no disclosure in Edwards
or Sanders (or any other document) in which a set of files are created and then
linked to the stored database blocks of the source database.” PO Resp. 54. In
particular, Patent Owner argues that Petitioner relies on Edwards’s disclosure
of creating a container file, “[b]ut this ‘container file’ is not a newly created
set of files.” Id. (citing Ex. 2115 ¶ 167–169). According to Patent Owner, the
new “container file” of Edwards “is merely a copy of the ‘vol_info’ block of
the volume to be cloned with pointers to existing files,” and “[n]o new set of
files is created.” Id. (citing Ex. 2115 ¶ 167–169).
Petitioner argues that “Edwards teaches that FlexClone creates a new
‘FlexVol volume,’” and “a FlexVol volume is a file system created within a
file on an underlying file system.” Reply 20 (citing Ex. 1105, 11, 14).
According to Petitioner, “cloning a volume containing database files creates a
new FlexVol volume with a new file system comprising a new set of database
files.” Id. (citing Pet. 36–39).
We find Petitioner’s argument persuasive. Edwards discloses that
creating a clone volume, or FlexClone volume, “requires the writing of a
fixed, small number of blocks,” and the process includes creating the files
required for a new FlexVol volume by “seed[ing] the container file of the
clone with a vol_info block that is a copy of the vol_info block of the
Snapshot copy on which the clone is based.” Ex. 1105, 14–15. Edwards
discloses that a container file “contains all the blocks of the FlexVol volume”
IPR2015-00108
Patent 8,566,361 B2

50
and, in some respects, “serves as a virtual disk containing the FlexVol
volume.” Id. at 11. Moreover, although a container file is initially sparsely
populated, with “most of the logical offsets hav[ing] no underlying physical
storage,” “WAFL allocates physical storage to the container file as the
FlexVol volume writes data to new logical offsets.” Id. at 12. This disclosure
suggests that a container file is a physical file rather than a mere copy of a
block with pointers to actual files. And even if a container file is merely a
copy of a block with pointers to actual files, as Patent Owner argues, the claim
language does not distinguish between virtual and actual files. Rather, claims
9 and 21 simply recite “a set of files linked to the stored database blocks.” As
such, we find that creating container files creates a new set of files.
We conclude that a preponderance of the evidence demonstrates that a
person of ordinary skill would have found the limitation of creating a set of
files linked to the stored database blocks obvious.
f) “receiving database blocks of the second database”
Patent Owner argues that “Petitioner fails to demonstrate any disclosure
in the cited documents describing receiving database blocks for different
point-in-time copies of a second database where the second database is
created by applying ETL operations to a virtual database created with
database blocks from a first (production) database.” PO Resp. 57–58 (citing
Ex. 2115 ¶ 178.
This argument, however, is based on the flawed premise that the
asserted references do not disclose creating a second database by applying
ETL operations to a virtual database created with database blocks from a first
database. The Petition points to Sanders’s disclosure of “performing extract
and transform operations on the data in the clone database” and storing the
IPR2015-00108
Patent 8,566,361 B2

51
output of the extracted and transformed data in another database (Pet. 41–42
(citing Ex. 1104, 5; Ex. 1117 ¶¶ 144, 145)) and Patel’s disclosure of using a
clone database for ETL operations (id. at 42 (citing Ex. 1106, 4, Fig. 2;
Ex. 1117 ¶ 146)). Petitioner then argues that
[g]iven that Patel expressly discloses performing extract,
transform, and load operations on a clone database, and
that Sanders discloses that a cloned database used for
transformation can be loaded onto a second database (such
as a data warehouse database disclosed in Sanders),
Sanders in view of Patel discloses performing extract,
transform, and load (ETL) operations on the data in the
first virtual database and storing the output of the ETL
operations in a second database in a data warehouse
system.
Id. at 44 (citing Ex. 1117 ¶ 147) (emphasis added).
We find these arguments persuasive. Referencing its Figure 2, Patel
discloses an approach to database cloning and development in which
“[w]orking clones of very large databases can be created in just a few
minutes.” Ex. 1106, 6. Figure 2 depicts several clones of a production
database, including a clone labeled “PROD” and another clone, labeled
“ETL,” which appears to be a clone of the PROD clone. Id. at Fig. 2. The
acronym “ETL” refers to extract, transform, and load. See id. at 4. Given
these disclosures, and the testimony of Dr. Zadok, we are persuaded that a
person of ordinary skill in the art would understand Patel as teaching creating
a second database by performing extract, transform, and load operations on
the data in a first virtual database. We are also persuaded by Petitioner’s
argument that “Patel and Sanders disclose receiving database blocks for
different point-in-time copies of the second database on a SnapMirror
IPR2015-00108
Patent 8,566,361 B2

52
destination, such as a destination volume in the storage system.” See Pet. 46
(citing Ex. 1117 ¶ 150).
We conclude that a preponderance of the evidence demonstrates that a
person of ordinary skill would have found the receiving database blocks of the
second database limitation obvious over the combination of Sanders,
Edwards, and Patel.
g) Conclusion
In the Decision to Institute, we determined that Petitioner had shown a
reasonable likelihood of prevailing on this proposed ground of unpatentability.
Inst. Dec. 16–25. We have reviewed both parties’ arguments and supporting
evidence, including the disclosure of all three references and the testimony of
both Dr. Zadok and Dr. Shenoy. Despite the counter-arguments in Patent
Owner’s Response, and the evidence cited therein, for the reasons discussed
above, Petitioner has shown, by a preponderance of the evidence, that claims
9 and 21 are unpatentable under 35 U.S.C. § 103 as directed to subject matter
that would have been obvious to a person of ordinary skill in the art in light of
Sanders, Edwards, and Patel.
5. Claim 10
Claim 10 depends from claim 9 and further recites “storing the database
blocks received for different point-in-time copies of the second database in the
storage system.” Ex. 1101, 36:58–60. Relying upon the testimony of Dr.
Zadok, Petitioner contends that the combination of Sanders and Patel
discloses the additionally recited limitations of claim 10. Pet. 47 (citing
Ex. 1117 ¶¶ 153, 154).
IPR2015-00108
Patent 8,566,361 B2

53
Patent Owner argues that Sanders and Patel do not disclose “storing
database blocks of the second database back into the original storage system
from which they originated. PO Resp. 59 (citing Ex. 2115 ¶ 181). According
to Patent Owner
[r]eplicating a volume from a source to a destination, as
described in Patel and Sanders is not the same as
performing ETL operations on a specific set of database
blocks, using that output to create a second, non-virtual
database, in a data warehouse system, and then receive and
store database blocks from different-point-in-time copies
of the data warehouse database back into the original
storage system.
Id. (citing Ex. 2115 ¶ 181).
In its Reply, Petitioner argues that, as pointed out in the Petition,
“Sanders explicitly discloses that a ‘SnapMirror source and its corresponding
destination can reside on the same storage system.’” Reply 25 (citing Ex.
1104, 3). Petitioner argues that this disclosure means that “the snapshots of a
‘second database’ (Sander’s ‘data warehouse database’) may be mirrored from
one volume of the SnapMirror destination ‘storage system’ to a different
volume on the ‘same storage system,’ i.e. the original storage system from
which they originated.” Id. (citing Pet. 45).
After considering all evidence and arguments, we agree with Petitioner.
In particular, we find a preponderance of the evidence supports the assertion
in the Petition that Sanders discloses storing a SnapMirror source and its
corresponding destination on the same storage system. Moreover, we find
persuasive Dr. Zadok’s testimony that Patel and Sanders disclose storing the
database blocks for different point-in-time copies of a second database on the
original storage system. See Ex. 1117 ¶¶ 149–154. Accordingly, on the full
record before us, we determine that Petitioner has shown, by a preponderance
IPR2015-00108
Patent 8,566,361 B2

54
of the evidence, that claim 10 is unpatentable under 35 U.S.C. § 103 as
directed to subject matter that would have been obvious to a person of
ordinary skill in the art in light of Sanders, Edwards, and Patel.
6. Claims 11, 12, and 22
Claim 11 depends from claim 9 and further recites provisioning a
second virtual database (VDB) to a data mart system. Ex. 1101, 36:61–67.
Similarly, claim 12 depends from claim 9 and further recites provisioning a
plurality of VDBs to a plurality of data mart systems. Id. at 37:1–9. Claim 22
depends from claim 21 and adds that the computer executable code further
comprises instructions to provision a plurality of VDBs to a plurality of data
mart systems. Id. at 38:58–67.
Petitioner contends that the combination of Sanders, Edwards, and Patel
discloses the additionally recited limitations of claims 11, 12, and 22. Pet.
47–52, 55–56 (citing Ex. 1117 ¶¶ 155–162, 164, 166–168, 174, 175). For
example, relying upon the testimony of Dr. Zadok, Petitioner asserts that
Sanders discloses “provisioning” a first virtual database by making it
“available to a system by using FlexClone to create a clone of the first
database on FlexVol volumes and by mounting the clone to a database
server.” Id. at 47 (citing Ex. 1117 ¶ 155). Petitioner also asserts that “Patel
discloses creating multiple clones from a single database for various
purposes,” and concludes that “Sanders in view of Patel discloses
provisioning of a second VDB to a data mart system.” Id. at 48 (citing
Ex. 1117 ¶ 157). In addition, Petitioner argues, in the same manner as
discussed above in connection with claim 9, that “Sanders in view of Edwards
discloses creating a set of files linked to the stored database blocks on the
IPR2015-00108
Patent 8,566,361 B2

55
storage system, wherein the stored database blocks are associated with the
second database.” Id. at 50 (citing Ex. 1117 ¶ 159).
Patent Owner makes two arguments with respect to claims 11, 12, and
22. First, Patent Owner argues that Petitioner has failed to demonstrate any
disclosure of “receiving database blocks for different point-in-time copies of
the second database” or storing these database blocks in “the storage system.”
PO Resp. 60 (citing Ex. 2115 ¶ 182). This argument, however, is essentially
the same argument made in connection with claim 10, and is not persuasive
for the same reasons discussed above.
Second, Patent Owner argues that “Petitioner has also failed to identify
any disclosure in any document that describes the additional step of
provisioning one or more VDBs to one or more data mart systems by creating
sets of files linked to the required stored database blocks from the second
database.” Id. (citing Ex. 2115 ¶ 182). We disagree. As described above, the
Petition sets forth an explanation of how the combination of Sanders and Patel
purportedly discloses provisioning multiple virtual databases to one or more
data mart systems. We find Petitioner’s argument and Dr. Zadok’s testimony
persuasive.
Thus, we find that Petitioner has shown, by a preponderance of the
evidence, that claims 11, 12, and 22 are unpatentable under 35 U.S.C. § 103 as
directed to subject matter that would have been obvious to a person of
ordinary skill in the art in light of Sanders, Edwards, and Patel.
D. Asserted Ground Based on Sanders, Edwards, Patel, and
Chapman
Petitioner challenges claims 13 and 23 as obvious under 35 U.S.C.
§ 103 over Sanders, Edwards, Patel, and Chapman. Pet. 4, 56–59. Claim 13
IPR2015-00108
Patent 8,566,361 B2

56
depends from claim 9 and further recites “performing backup of the database
blocks associated with the first database to a second storage system, wherein
performing backup comprises transmitting database blocks associated with the
first database from the first storage system to a second storage system for
storage in the second storage system.” Ex. 1101, 37:9–16. Claim 23 depends
from claim 21 and adds that the computer executable code further comprises
instructions to perform the same backup. Ex. 1101, 39:1–9.
Petitioner relies on Chapman for disclosing “the use of data cascading,
which is a ‘variation on the basic SnapMirror deployment and function
involves mirroring from established mirrors to more SnapMirror
destinations’” (Pet. 57 (quoting Ex. 1107, 22) (emphasis by Petitioner)) and
“SnapMirror volume replication is a block-for-block replication; it transfers
the file system verbatim” (id. (quoting Ex. 1107, 9)). Petitioner also argues
that Sanders discloses using SnapMirror to transmit database blocks and notes
that SnapMirror provides an efficient way to transfer data from one storage
system to another. Id. at 58 (citing Ex. 1104, 3). Based on these teachings,
Petitioner asserts that the combination of Sanders and Chapman discloses the
additional limitations recited in claims 13 and 23. Id. at 58–59. Patent Owner
does not separately argue these claims, but, although not expressly stating as
much, apparently relies on the same arguments as those advanced with respect
to the independent claims. PO Resp. 30–60.
In the Decision to Institute, we determined that Petitioner had shown a
reasonable likelihood of prevailing on this proposed ground of unpatentability.
Inst. Dec. 25–26. We now determine that a preponderance of evidence
demonstrates that a person of ordinary skill would have found the additional
limitations recited by dependent claims 13 and 23 obvious. In particular, we
IPR2015-00108
Patent 8,566,361 B2

57
are persuaded by Petitioner’s argument and Dr. Zadok’s testimony that the
combination of Sanders and Chapman discloses the limitations of claims 13
and 23. See Pet. 56–59; Ex. 1117 ¶¶ 177–181.
Therefore, on the full record before us, we determine that Petitioner has
shown, by a preponderance of the evidence, that claims 13 and 23 are
unpatentable under 35 U.S.C. § 103 as directed to subject matter that would
have been obvious to a person of ordinary skill in the art in light of Sanders,
Edwards, Patel, and Chapman.
IV. MOTION TO EXCLUDE
Patent Owner moves to exclude Petitioner’s Exhibits 1106, 1120, 1128,
1132, 1133, 1135, 1136, 1138, 1141–1145, 1147, 1148, 1150, 1154, 1155,
1157, 1159, 1160, 1164, 1166, and 1167. Paper 36 (“Mot. to Exclude”), 1. In
inter partes reviews, documents are admitted into evidence subject to an
opposing party asserting objections to the evidence and moving to exclude the
evidence. 37 C.F.R. § 42.64. As movant, Patent Owner has the burden of
showing that an Exhibit is not admissible. 37 C.F.R. § 42.20(c).
Patent Owner contends that Petitioner does not rely on Exhibits 1132,
1133, 1135, 1138, 1144, 1148, 1150, 1154, 1157, 1159, 1164, 1166, and 1167.
Mot. To Exclude 1 n.1. Petitioner does not dispute this contention. See Paper
46 (“Opp. Mot. to Exclude”). We, therefore, grant Patent Owner’s motion as
it pertains to Exhibits 1132, 1133, 1135, 1138, 1144, 1150, 1154, 1157, 1159,
1164, 1166, and 1167. We note, however, that Petitioner does rely on Exhibit
1148. See Reply 11. For reasons discussed below, we deny Patent Owner’s
motion as it pertains to Exhibit 1148. Further, Petitioner has moved,
unopposed, to expunge Exhibit 1128 (see Paper 34), which motion we hereby
grant.
IPR2015-00108
Patent 8,566,361 B2

58
Of the remaining objected-to Exhibits, our Final Written Decision
discusses and relies on only Exhibits 1106, 1120, 1147, and 1148. We
determine, for reasons set forth above, that Petitioner has demonstrated by a
preponderance of the evidence that the challenged claims are unpatentable,
without need for Petitioner’s additional arguments or evidence in relation to
the remaining additional Exhibits. Accordingly, Patent Owner’s motion to
exclude Exhibits 1136, 1141–1143, 1145, 1155, and 1160 is dismissed as
moot.
According to Patent Owner, “Exhibit 1106 is allegedly a white paper
for a NetApp product which Petitioner relies upon for its prima facie case of
obviousness.” Mot. to Exclude, 2. Patent Owner moves to exclude the
Exhibit (Patel) as not being authenticated pursuant to Federal Rule of
Evidence 901. Id.
Petitioner contends testimony by Louis Hernandez is sufficient to
authenticate the document pursuant to Rule 901(b)(1). Opp. Mot. to Exclude,
14. Patent Owner, however, argues that Mr. Hernandez lacks personal
knowledge to testify regarding this issue, rendering his statements hearsay.
Mot. to Exclude, 8–11 (referring to Exs. 1120, 1147); Paper 47, 4–5. We note
that Mr. Hernandez’s personal knowledge of Patel comes from his review
recognition of that document as being published during his tenure as an
employee of NetApp. Ex. 1120 ¶¶ 7, 12, 17–18; Ex. 1147 ¶ 3, 6, 11. This is
ample basis for him to testify with personal knowledge of the facts under FRE
602. We are not persuaded otherwise by Patent Owner’s characterization of
Mr. Hernandez’s cross-examination testimony. See Mot. to Exclude, 9–10
(citing Ex. 2131); Paper 38 (citing Ex. 2131)). As to the hearsay objection,
Mr. Hernandez’s testimony consists of statements made by the Declarant
IPR2015-00108
Patent 8,566,361 B2

59
while testifying in this proceeding — not “hearsay” (Fed. R. Evid. 801(c)) —
but sworn testimony that is subject to cross-examination. Indeed, as noted
above, Patent Owner cross-examined Mr. Hernandez with respect to that
testimony. We, therefore, deny Patent Owner’s motion to exclude Exhibits
1120 and 1147.
Petitioner also contends Patel is self-authenticating. Opp. Mot. to
Exclude, 13–14 (citing Rules 901(b)(1), 901(b)(4), 902(7)). Specifically,
Petitioner refers to the “NetApp trade logo, copyright (©) notices and trade
inscriptions such as trademark (® and TM) symbols throughout the
document,” as well “NetApp’s distinctive logo” appearing on the cover and
last pages. Id. Further, Petitioner points to Patel’s contents, which “include
repeated references to NetApp and various NetApp technologies.” Id.at 14.
We are persuaded that Patel is authenticated at least under Federal
Rules of Evidence 901(b)(1) and 901(b)(4). On this record, we accept Exhibit
1006 for what it purports and is alleged to be: A “White Paper” describing a
NetApp product, published by NetApp, for the purpose of providing detailed
instructions to its customers and potential customers. Ex. 1006.
The only reason Patent Owner gives for excluding Exhibit 1148 is that
it is “untimely under 37 C.F.R. § 42.23(b) and 37 C.F.R. § 42.123.” Mot. to
Exclude, 1 n.1. We find that Exhibit 1148, an article titled “Bigtable: A
Distributed Storage System for Structured Data,” was properly submitted by
Petitioner as evidence in rebuttal to Patent Owner’s claim construction
argument for the term “database block” asserting that it must necessarily
include metadata. In particular, Petitioner argues this article is relevant to
Patent Owner’s assertion that “it is well known to one skilled in the art that a
IPR2015-00108
Patent 8,566,361 B2

60
database block necessarily includes metadata.” See PO Resp. 17–21; Reply
10–11 (citing Ex. 1148).
V. ORDER
In consideration of the foregoing, it is hereby:
ORDERED that claims 9–13 and 21–23 of the ’361 patent have been
shown to be unpatentable;
FURTHER ORDERED that Petitioner’s motion to expunge Exhibit
1128 is granted;
FURTHER ORDERED that Patent Owner’s motion to exclude
evidence is granted-in-part, dismissed-in-part and denied-in-part;
FURTHER ORDERED that Exhibits 1132, 1133, 1135, 1138, 1144,
1150, 1154, 1157, 1159, 1164, 1166, and 1167 shall be expunged; and
FURTHER ORDERED that, because this is a Final Written Decision,
parties to the proceeding seeking judicial review of the decision must comply
with the notice and service requirements of 37 C.F.R. § 90.2.
IPR2015-00108
Patent 8,566,361 B2

61

PETITIONER:

Robert Steinberg
Jonathan Link
Latham & Watkins LLP
Giri Pathmanaban
bob.steinberg@lw.com
jonathan.link@lw.com
giri.pathmanaban@lw.com

PATENT OWNER:

David Hadden
Saina Shamilov
FENWICK & WEST LLP
dhadden-ptab@fenwick.com
sshamilov-ptab@fenwick.com