Ex Parte SRIVASTAVA et al

Patent Trials and Appeals Board

Jan 31, 2019

13237865 - (D) (P.T.A.B. Jan. 31, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE
13/237,865 09/20/2011
55498 7590 02/04/2019
Vista IP Law Group, LLP (Oracle)
2160 Lundy Avenue
Suite 230
San Jose, CA 95131
FIRST NAMED INVENTOR
Manish SRIVASTAVA
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
ORAl 10487-US-NP 1294
EXAMINER
EYERS, DUSTIN D
ART UNIT PAPER NUMBER
2164
NOTIFICATION DATE DELIVERY MODE
02/04/2019 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
docketing@viplawgroup.com
ev@viplawgroup.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte MANISH SRIVASTAVA, GERMAN BERTOT, YU LUNG NG,
SUMAN GURA, LEE RIAN QUEK, and JEFF GLANVILLE
Appeal2018-003930
Application 13/237,865
Technology Center 2100
Before CARL W. WHITEHEAD JR., IRVINE. BRANCH, and
JOSEPH P. LENTIVECH, Administrative Patent Judges.
LENTIVECH, Administrative Patent Judge.
DECISION ON APPEAL
Pursuant to 35 U.S.C. § 134(a), Appellants appeal from the
Examiner's decision to reject claims 1-20, the only claims pending in the
application on appeal. 1 We have jurisdiction over the pending claims under
35 U.S.C. § 6(b).
We AFFIRM.
1 According to Appellants, the real party in interest is Oracle International
Corporation. App. Br. 2.
Appeal2018-003930
Application 13/237,865
STATEMENT OF THE CASE
Appellants' Invention
Appellants' invention generally relates to "data management in an
electronic business intelligence system and more particularly to systems for
integration of historical data with online transactional data." Spec. ,r 1.
Claim 1, which is illustrative, reads as follows:
1. A computer-implemented method for performing
integrated transactional and data warehouse business intelligence
analysis, the method comprising:
selecting, based at least in part on user input, a business
intelligence attribute to analyze;
determining, for the business intelligence attribute,
whether to retrieve data records only from an online transactional
processing (OLTP) system, only from a data warehouse, or from
both the OL TP and the data warehouse;
extracting and pre-computing at least some data stored in
the data warehouse into precomputed data and storing the pre-
computed data in a memory for subsequent retrievals based in
part or in whole upon likelihood of subsequent repeated
retrievals of the at least some data;
when it is determined to retrieve the data records from
both the OL TP and the data warehouse:
retrieving a transactional data record from the
OLTP system when it is determined to retrieve the data
records for the business intelligence attribute from the
OLTP system, the transactional data record comprising at
least one OLTP business intelligence value of the selected
business intelligence attribute;
receiving, from the data warehouse that is different
than the OLTP system, a result set of a query to the data
warehouse when it is determined to retrieve the data
records for the business intelligence attribute from the data
warehouse, the result set comprising a historical business
2
Appeal2018-003930
Application 13/237,865
intelligence value of the selected business intelligence
attribute;
analyzing the selected business intelligence
attribute using the at least one OLTP business intelligence
value and the historical business intelligence value at least
by comparing the at least one OL TP business intelligence
value with the historical business intelligence value in a
context of each other when the data records are retrieved
from both the OLTP system and the data warehouse; and
processing, in an application server, a business
intelligence application, the business intelligence
application combining the at least one OL TP business
intelligence value from the OLTP system with the
historical business intelligence value from the data
warehouse and at least some of the pre-computed data
from the memory to form a correlated report when the data
records are retrieved from both the OLTP system and the
data warehouse, wherein the correlated report relates the
at least one OLTP business intelligence value to the
historical business intelligence value; and
displaying the correlated report on a single surface.
References and Rejections
Claims 1, 5, 6, 8, 12, 13, 15, and 19 stand rejected under 35 U.S.C.
§ I03(a) as being unpatentable over the combination of Hossfeld et al. (US
2007/0156718 Al; published July 5, 2007) ("Hossfeld") and Zadorozhny
(US 2008/0250058 Al; published Oct. 9, 2008). Final Act. 2-8.
Claims 2, 9, and 16 stand rejected under 35 U.S.C. § I03(a) as being
unpatentable over the combination of Hossfeld, Zadorozhny, and Kalenius
(US 2005/0216846 Al; published Sept. 29, 2005). Final Act. 8-10.
Claims 3, 4, 10, 11, 17, and 18 stand rejected under 35 U.S.C.
§ I03(a) as being unpatentable over the combination of Hossfeld,
3
Appeal2018-003930
Application 13/237,865
Zadorozhny, and Delurgio et al. (US 2008/0221949 Al; published Sept. 11,
2008) ("Delurgio"). Final Act. 10-15.
Claim 7 stands rejected under 35 U.S.C. § I03(a) as being
unpatentable over the combination of Hossfeld, Zadorozhny, Delurgio, and
Kalenius. Final Act. 15-19.
Claim 14 and 20 stands rejected under 35 U.S.C. § I03(a) as being
unpatentable over the combination of Hossfeld, Zadorozhny, and Kalenius.
Final Act. 19-23.
ANALYSIS
Claim 1
Appellants contend the combination of Hossfeld and Zadorozhny fails
to teach or suggest "determining, for the business intelligence attribute,
whether to retrieve data records only from an online transactional processing
(OLTP) system, only from a data warehouse, or from both the OLTP and the
data warehouse," as recited in claim 1. App. Br. 21-28; Reply Br. 2-9.
Appellants argue "Hossfeld merely describes a single source of data -
database 1 7 - and thus never needs to perform a 'determining' step to
determine whether to retrieve data records from among multiple choices of
sources ( either an OL TP system or a data warehouse as claimed where ' ...
the data warehouse . .. is different than the OLTP system . .. ')." App.
Br. 21.
Appellants further contend:
[B]ecause Hossfeld only describes orgamzmg database 17
(whether singular or plural) to process real-time transactions and
to support a business intelligence analysis, Hossfeld thus not only
does not and need not "determin[e] ... whether to retrieve data
4
Appeal2018-003930
Application 13/237,865
records only from an online transactional processing (OLTP)
system, only from a data warehouse, or from both the OLTP and
the data warehouse" but also fails to disclose or suggest
"retrieving a transactional data record from the OLTP system"
and "receiving, from the data warehouse . . . , a result set of a
query" "when it is determined to retrieve the data records from
both the OLTP and the data warehouse" "that is different than
the OLTP system".
App. Br. 23. Appellants argue Hossfeld fails to teach the disputed
limitations because Hossfeld does not teach or suggest determining a source
for data retrieval from three choices-an OLTP system, a data warehouse,
both the OL TP system and the data warehouse, as required by claim 1. App.
Br. 24--28 ( citing Hossfeld ,r,r 20, 25, 55). Appellants argue Hossfeld,
instead, teaches retrieving data from a single data source-Hossfeld's multi-
dimensional database. Id.
Appellants' arguments are not persuasive. The Examiner finds, and
we agree, that "Hossfeld discloses that business data is stored in a central
repository containing a multi-dimensional database and that multiple
business data sets are stored together as multi-dimensional data cubes within
the database." Ans. 3 (citing Hossfeld ,r 24). We also agree with the
Examiner that Hossfeld teaches the multi-dimensional database "may be
organized both to process real-time transactions as in online transaction
processing systems ('OLTP'), and to support business intelligence analysis."
Ans 3 ( citing Hossfeld ,r 25). Hossfeld further teaches that "[ d]ata cubes
may be used to associate and store any number of discrete data sets" and
"discrete data sets may be stored separately within database 17 and
associated with other data sets thereafter by various reporting software
applications." Hossfeld ,r 24. As such, we find Hossfeld teaches, or at least
suggests, that the database is organized into a first section composed of data
5
Appeal2018-003930
Application 13/237,865
cubes storing discrete data sets for processing real-time transactions (e.g., an
OLTP system) and a second, different section composed of data cubes
storing discrete data sets for supporting business intelligence analysis ( e.g., a
data warehouse).
Appellants further contend the combination of Hossfeld and
Zadorozhny fails to teach or suggest "extracting and pre-computing at least
some data stored in the data warehouse into pre-computed data and storing
the pre-computed data in a memory for subsequent retrievals based in part or
in whole upon likelihood of subsequent repeated retrievals of the at least
some data," as recited in claim 1. App. Br. 28-30; Reply Br. 9-11.
Regarding these disputed limitations, the Examiner finds:
Zadorozhny discloses "a forward table that can contain
information representing pre-computed combined probabilities"
and "the combined probability of transitioning from state 81 to
state 82" (Paragraph [0027]). This shows that Zadorozhny
discloses pre-computing information and storing the information
inside of a forward table. The probability of transitioning from
one state to another is considered the likelihood of subsequent
repeated retrievals. Appellant emphasizes that there is no
bearing on any "likelihood" however, a probability of occurrence
as disclosed in Zadorozhny is an obvious variant of a
"likelihood". Zadorozhny further discloses that "summarization
component can employ mapping rules to map received data into
specified states" (Abstract). This shows that transitioning from
one state to another as described in Paragraph [0027] would be a
form of data retrieval as each state corresponds to a set of data.
Furthermore, the claim limitation only requires that "at least
some data stored" is pre-computed, this would mean that any
data stored in the respective source that is precomputed would
meet this limitation.
Ans. 7.
6
Appeal2018-003930
Application 13/237,865
Appellants argue Zadorozhny teaches that the summarization
component "can generate a forward table that contains computed combined
probabilities of transitioning from state S 1 to state S2" and "utilizing the
information in the forward table to generate tables that contain information
of combined probabilities and determining the probabilities and de-
reification of observations" but these teachings "ha[ ve] no bearing on any
'likelihood of subsequent repeated retrievals of the at least some data,'" as
required by claim 1. App. Br. 30-31 ( citing Zadorozhny ,r 27). Appellants
further argue Zadorozhny does not teach or suggest "that the precomputed
'probabilities' have any bearing on subsequent repeated retrievals," as
required by claim 1. Reply Br. 11.
We find Appellants' arguments unpersuasive. Zadorozhny teaches
"using information for the PADB [process analytical database], the
summarization component 102 can generate a forward table that can contain
information representing pre-computed combined probabilities."
Zadorozhny ,r 27. As such, we find Zadorozhny teaches "extracting and pre-
computing at least some data stored in the data warehouse into pre-computed
data and storing the pre-computed data in a memory for subsequent
retrievals." Zadorozhny further teaches "[i]t is desirable to be able to
quickly and efficiently derive pertinent information from large collections of
data" and that "it is desirable to discover trends in dynamic data associated
with complex processes." Zadorozhny ,r 19. As such, we find Zadorozhny
teaches or suggests that the pre-computed combined probabilities are stored
in the forward table based in part or in whole upon likelihood of subsequent
repeated retrievals of the pre-computed combined probabilities ( e.g., "at
least some data"), as required by claim 1.
7
Appeal2018-003930
Application 13/237,865
Appellants' arguments (see e.g., App. Br. 29-30) regarding
Zadorozhny failing to teach or suggest the disputed limitations because
Zadorozhny does not teach "the claimed 'extracting and pre-computing at
least some data ... based in part or in whole upon likelihood of subsequent
repeated retrievals of the at least some data"' are not persuasive because
they are not commensurate with the scope of the claim. Claim 1 does not
require that the data be extracted and pre-computed based in part or in whole
upon likelihood of subsequent repeated retrievals of the at least some data.
Instead, claim 1 requires storing the pre-computed data in a memory for
subsequent retrievals based in part or in whole upon likelihood of
subsequent repeated retrievals of the at least some data.
For the foregoing reasons, we are not persuaded the Examiner erred in
rejecting claim 1; and claims 5, 6, 8, 12, 13, 15, and 19, which are not
separately argued with particularity.
Claims 2, 9-11, 14, 16--18, and 20
Claims 2, 9-11, 14, 16-18, and 20 stand rejected under 35 U.S.C.
§ 103(a) based on Hossfeld, Zadorozhny, and various additional references.
Final Act. 8-19. Appellants argue the additional references cited in the
rejection of these claims do not cure the deficiencies in the teachings of
Hossfeld and Zadorozhny discussed above with respect to claim 1. App.
Br. 32. Accordingly, we are not persuaded the Examiner erred in rejecting
claims 2, 4, 7, 9-11, 14, 16-18, and 20 for the reasons discussed above
regarding claim 1.
8
Appeal2018-003930
Application 13/237,865
Claim 3
Claim 3 depends from claim 1 and recites that the method further
comprises:
selecting a performance metric for analysis;
extracting the data records only from the OLTP system for
the analysis in the first time period;
identifying a second time period for the analysis;
extracting the data records only from the data warehouse
for the analysis in the second time period;
identifying a third time period that represents an overlap
between the first time period and the second time period;
extracting the data records from only the data warehouse,
from only the OLTP system, or from both the data warehouse
and the OLTP system based in part or in whole upon correlated
nature of the data records in the data ware house and the OLTP
system during the third time period; and
performing the analysis for the first time period, the
second period, and the third time period with the data records.
App. Br. 40-41 (Claims App'x).
Appellants contend the combination of Hossfeld, Zadorozhny, and
Delurgio fails to teach or suggest the limitations recited in claim 3. App.
Br. 33-37; Reply Br. 13-18. Appellants argue the cited references fail to
teach the limitations recited in claim 3 because "Delurgio never even
remotely suggests retrieving data from different sources in the three claimed
'time periods'" and "this deficiency is especially pronounced given the fact
that Hossfeld merely describes a single database (17) that processes real-
time transactions as in OLTP systems and supports a business intelligence
analysis." App. Br. 35. Appellants further argue the combination of
Hossfeld, Zadorozhny, and Delrugio "simply fails to disclose any
9
Appeal2018-003930
Application 13/237,865
mechanisms store data records on different systems (the claimed OLTP
system and data warehouse) for two different time periods (the claimed "first
time period" and the "second time period"). Id.
Appellants also argue:
[ A ]lthough Delrugio recites that "actual historical dependent
data 102 and historical causal data 104 can be for any past time
period and of varying length, such as a days, weeks, months,
years, etc." in ,r [0025], Delrugio NEVER even remotely
suggests retrieving any specific data in three different time
periods with some overlapping relationship ( the claimed "first
time period", "second time period", and "third time period that
represents an overlap ... "), much less extracting data records
from different sources for different time periods ( e.g., only from
OLTP for the first time period and only from the data warehouse
for the second time period).
App. Br. 35-36. Additionally, Appellants argue "to the [extent] that
Delrugio' s comparison between its predicted historical dependent data 110
and actual historical dependent data 102 is considered as allegedly
disclosing the claimed 'analysis', Delrugio nevertheless fails to disclose
extracting the alleged 'data records' from different sources for different time
periods." App. Br. 36.
Appellants further argue:
Delrugio' s "predicted historical dependent data 11 O" is
derived from the "historical causal data 104[.]" Delrugio then
compares the derived "predicted historical dependent data 11 O"
to the "actual historical dependent data 102" to determine the
reliability of its forecast model. Although Delrugio does
describe that its "actual historical dependent data 102" and
historical causal data 104 may be for any time period and of
varying length and may or may not end at a specified time ( 108),
Appellant respectfully asserts that in order for Delrugio' s
comparison between the actual historical dependent data 102 and
10
Appeal2018-003930
Application 13/237,865
predicted historical dependent data 110 to be meaningful, the
same time period must be selected for the following reason.
That is, if different time periods were to be selected, the
driver in the historical causal data 104 would NOT "affect" the
actual historical dependent data because the driver events do not
even occur in the same time period. The predicted historical
dependent data 110 derived from such historical causal data 104
will also involve the driver event information in a time period
and thus does not "affect" the "actual historical dependent data
11 O" in a different time period. As a result, the comparison
between the predicted historical dependent data 110 and the
actual historical dependent data 102 will not provide meaningful
results, not to mention determining the reliability of Delrugio' s
forecast models. Therefore, the only way that such comparisons
may yield meaningful results is to have Delrugio' s actual
historical dependent data 102 and historical causal data 104 in
the same time period, instead of the claimed "first time period"
and "second time period".
App. Br. 36-37.
Appellants' arguments are not persuasive. As discussed supra with
respect to claim 1, the Examiner finds, and we agree, Hossfeld teaches or
suggests the claimed OL TP system and data warehouse. Ans. 3 ( citing
Hossfeld ,r 24). The Examiner finds Delrugio teaches extracting data
associated with first, second and third time periods. Ans. 10-11 ( citing
Delrugio ,r,r 27, 40, and Fig. 4). As such, the Examiner relies on the
combined teachings of Hossfeld and Delrugio for teaching or suggesting the
disputed limitations. Appellants' arguments do not persuasively address the
combined teachings of the references and, therefore, are unpersuasive of
error.
Delrugio describes actual dependent data 102 and historical causal
data 104 as being associated with a time period ending at time 402.
Delrugio ,r 40, Fig. 4. Delrugio describes predicted dependent data 306 and
11
Appeal2018-003930
Application 13/237,865
predicted causal data 302 as being associated with a time period beginning at
time 402 and ending at time (future) 406. Id. Delrugio further describes
incremental actual dependent data 408 and incremental historical causal data
410 as being associated with a time period beginning at time 402 and ending
at updated time 404, which is prior to time (future) 406. Delrugio ,r 40,
Fig. 4. The Examiner finds, and we agree, Delrugio teaches that as time
evolves from 402 to updated time 404, the actual/historical data will overlap
with the predicted data. Ans. 9 ( citing Delrugio ,r 40). As such, we agree
with the Examiner (Ans. 9-10) that Delrugio teaches or suggests a first time
period ending at the updated time 404, a second time period beginning at
time 402 and ending at time (future) 406, and a third time period beginning
at time 402 and ending at updated time 404 that overlaps a portion of the
first time period and the second time period.
Delrugio teaches using the historical dependent and causal data to
estimate a forecast model. Delrugio ,r 28. Delrugio further teaches
"[i]ncremental actual dependent data 408 is compared with the same portion
of the predicted dependent data 306 to determine if forecast model 308 is
reliable." Delrugio ,r 40. Because the incremental actual dependent data
408 is associated with the first and third time periods and the predicted
dependent data is associated with the second and third time periods, we
agree with the Examiner that Delrugio teaches or suggests extracting data in
the first, second and third time periods. Ans. 10-11 ( citing Delrugio ,r,r 27,
40, and Fig. 4). As discussed supra with respect to claim 1, the Examiner
finds, and we agree, Hossfeld teaches or suggests the claimed OLTP system
and data warehouse (Ans. 3 (citing Hossfeld ,r 24)) and concludes the
12
Appeal2018-003930
Application 13/237,865
combination of Hossfeld and Delrugio teaches or suggests the limitations
recited in claim 3 (Final Act. 10-13).
Accordingly, we are not persuaded the Examiner erred in rejecting
claim 3, and claims 4 and 7, which depend from claim 3 and are not
separately argued with particularity. See App. Br. 33-37.
DECISION
We affirm the Examiner's rejections of claims 1-20 under 35 U.S.C.
§ 103(a).
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l )(iv).
AFFIRMED
13