Ex Parte Barsness et al

Patent Trial and Appeal Board

Apr 11, 2014

11861343 (P.T.A.B. Apr. 11, 2014)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte ERIC L. BARSNESS, DAVID L. DARRINGTON,
AMANDA PETERS, and JOHN MATTHEW SANTOSUOSSO
____________

Appeal 2011-009517
Application 11/861,343
Technology Center 2100
____________

Before DONALD E. ADAMS, DEMETRA J. MILLS, and
LORA M. GREEN, Administrative Patent Judges.

ADAMS, Administrative Patent Judge.

DECISION ON APPEAL1
This appeal under 35 U.S.C. § 134 involves claims 1, 2, 4-11, 13, 14,
and 16-20 (App. Br. 2). Examiner entered a rejection under 35 U.S.C.
§ 103(a). We have jurisdiction under 35 U.S.C. § 6(b).
We reverse.

1 The Real Party in Interest is International Business Machines Corporation
(App. Br. 1).
Appeal 2011-009517
Application 11/861,343

2
STATEMENT OF THE CASE
The claims are directed to a parallel computer system, a computer
implemented method for pre-loading an in-memory database into memory of
a plurality of compute nodes of a parallel computer system, and a computer-
readable article of manufacture. Claims 1, 9, and 14 are representative and
are reproduced in the Claims Appendix of Appellants’ Brief.
Claims 1, 2, 4-11, 13, 14, and 16-20 stand rejected under 35 U.S.C.
§ 103(a) as unpatentable over the combination of Baru,2 Muzaffar,3 and
Pandit.4
ISSUE
Does the preponderance of evidence relied upon by Examiner support
a conclusion of obviousness?
FACTUAL FINDINGS (FF)
FF 1. Baru suggests
[A]lternative paths . . . for determining and arranging for the
redistribution of resource loading either based on the volume of
data present at each node or the workload for instance
transaction activity of the nodes. Depending on the potential
advantages the most optimal distribution of data may be
selected from balancing either workload activity or data volume
storage.
(Baru, col. 6, ll. 49-55; see generally Ans. 4.)
FF 2. Baru suggests that “the software of the invention obtains transaction
activity information for all table partitions by reading the database logs
associated with them . . . and generates the current workload distribution file

2 Baru et al., US 5,970,495, issued October 19, 1999.
3 Muzaffar et al., US 7,263,695 B1, issued August 28, 2007.
4 Pandit, US 5,937,402, issued August 10, 1999.
Appeal 2011-009517
Application 11/861,343

3
which depicts the current distribution of workload among the nodes” (Baru,
col. 7, ll. 31-36; see generally Ans. 4).
FF 3. Examiner finds that Baru does not suggest “an ‘in-memory
database’” and relies on Muzaffar to make up for this deficiency in Baru
(Ans. 5).
FF 4. Examiner finds that the combination of Baru and Muzaffar fails to
suggest “flag[s] in an SQL statement” and relies on Pandit to make up for
this deficiency in the combination of Baru and Muzaffar (Ans. 5 (alteration
original)).
FF 5. Pandit suggests that a
[P]rocess prepares a number of SQL statements that are
required by the runtime system. These SQL statements contain
placeholder flags which are filled by the runtime system for
each message required to be sent to the database. This is a
performance measure to save time at runtime by doing the work
of generating the SQL statements beforehand.
(Pandit, col. 5, ll. 30-34.)
ANALYSIS
Based on the combination of Baru, Muzaffar, and Pandit, Examiner
concludes that, at the time Appellants’ invention was made, it would have
been prima facie obvious “to incorporate the teaching of Pandit into the
combination of Muzaffar and Baru’s system” (Ans. 5). According to
Examiner, the “combination of Pandit and Baru . . . [suggests] using SQL
statements with flags for clustering data[] by allowing SQL statements in
Baru to contain flags in place of workload activities or data volume to
indicate when and where a partition or re-distribution of data should take
place by acting as a signaling mechanism when a threshold is met” (Ans. 5-
6). We are not persuaded.
Appeal 2011-009517
Application 11/861,343

4
As Appellants explain, Baru suggests “pre-loading [a] database based
on workload activity” (App. Br. 6; FF 1-2). “Baru does not address the
concept of loading a database based on a flag in an SQL statement,” as
required by Appellants’ claimed invention (id.; see Appellants’ claims 1, 9,
and 145). Further, while Pandit suggests “the general concept of placing a
flag in the SQL that can later be used by the runtime system . . .[,] Pandit
does not teach or suggest ‘wherein the database loader clusters the database
attributes based on a flag in an SQL statement,” as is required by
Appellants’ claimed invention (App. Br. 7; see FF 5; see also Appellants’
claims 1, 9, and 14). In combination, the cited art, at best, suggests “pre-
loading a database (Baru) and that flags can be used as placeholders to
generate a message to the database (Pandit)” (App. Br. 7; FF 1, 2, and 5).
Muzaffar’s suggestion of an “in-memory database” fails to make up for the
foregoing deficiencies in the combination of Baru and Pandit (FF 3).
We recognize Examiner’s assertion that “Baru must use SQL
statements to query the database to get workload activities and other
database related queries,” because “Baru teaches pre-loading [a] database
based on workload activities” (Ans. 9-10). Examiner, however, fails to
establish an evidentiary basis on this record to support this conclusion.
Therefore, Examiner’s rationale that it would have been prima facie obvious
to a person of ordinary skill in this art to “allow[] SQL statements in Baru to
contain flags in place of workload activities or data volume to indicate when
and where a partition or re-distribution of data should take place by acting as

5 Examiner asserts that Appellants’ “[i]ndependent [c]laims 1, 9, and 14
consist[] of similar limitations . . . [and] hence [are] rejected similarly” (Ans.
8 (emphasis removed)).

Appeal 2011-009517
Application 11/861,343

5
a signaling mechanism when a threshold is met” is not supported by an
evidentiary basis on this record (emphasis added, see Ans. 6 and 10). See In
re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006) (“rejections on obviousness
grounds cannot be sustained by mere conclusory statements; instead, there
must be some articulated reasoning with some rational underpinning to
support the legal conclusion of obviousness.”).
CONCLUSION OF LAW
The preponderance of evidence relied upon by Examiner fails to
support a conclusion of obviousness. The rejection of claims 1, 2, 4-11, 13,
14, and 16-20 under 35 U.S.C. § 103(a) as unpatentable over the
combination of Baru, Muzaffar, and Pandit is reversed.

REVERSED

lp