Ex Parte Cao et al

Patent Trial and Appeal Board

Dec 20, 2018

14308993 (P.T.A.B. Dec. 20, 2018)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
14/308,993 06/19/2014 Bin Cao
46296 7590 12/25/2018
MARTIN & ASSOCIATES, LLC
P.O. BOX548
CARTHAGE, MO 64836-0548
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.
ROC920140020US1 1078
EXAMINER
NGUYEN, TRUONG
ART UNIT PAPER NUMBER
2449
NOTIFICATION DATE DELIVERY MODE
12/25/2018 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):
derekm@ideaprotect.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte BIN CAO, BRIAN R. MURAS, and JINGDONG SUN
Appeal2018-005740
Application 14/308,993 1
Technology Center 2400
Before ROBERT E. NAPPI, ERIC S. FRAHM, and MICHAEL T. CYGAN,
Administrative Patent Judges.
CYGAN, Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF CASE
Introduction
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner's
Final Rejection of claims 1-9 and 20. We have jurisdiction under 35 U.S.C.
§ 6(b ).
We affirm.
1 According to Appellants, International Business Machines Corporation is
the real party in interest. App. Br. 1.
Appeal2018-005740
Application 14/308,993
Disclosed Invention and Exemplary Claim
The disclosed invention is a streaming application and a streams
manager residing in memory coupled to a processor, in which the streams
manager automatically deploys and adjusts a helper operator to relieve a
bottleneck detected at an operator in the streaming application (Spec. ,r 9,
claim 1 ).
Independent claim 1 is exemplary of the disclosed invention, and
reads as follows (with key limitations of the claim emphasized):
1. An apparatus comprising:
at least one processor;
a memory coupled to the at least one processor;
a streaming application residing in the memory and
executed by the at least one processor, the streaming application
comprising a flow graph that includes a plurality of operators that
process a plurality of data tuples; and
a streams manager residing in the memory and executed
by the at least one processor, the streams manager deploying the
streaming application to the memory and deploying at least one
helper operator that includes logic for multiple operators in
the flow graph and has an input and an output that initially are
disconnected, the streams manager monitoring performance of at
least one of the plurality of operators in the streaming
application, and when one of the at least one operators in the
streaming application becomes a bottleneck, the streams
manager adjusts the at least one helper operator by connecting
the input and the output of the helper operator to the flow graph
to alleviate the bottleneck in the one operator.
Examiner's Rejections
(1) The Examiner rejected claims 1-9 and 20 under 35 U.S.C.
§ 112(b) as being indefinite for failing to particularly point out and distinctly
2
Appeal2018-005740
Application 14/308,993
claim the subject matter which the inventor or a joint inventor regards as the
invention. Final Act. 2-3; Ans. 4--5.
(2) The Examiner rejected claims 1-5, 7-8, and 20 under 35 U.S.C.
§ I03(a) as being obvious over the combination of Raspl et al. (US
2007 /0046664 Al; published March 1, 2007) (hereinafter "Raspl") and
Andrade et al. (US 2011/0083046 Al; published April 7, 2011) (hereinafter
"Andrade" or "Andrade '046"). Final Act. 4--8; Ans. 5-8.
(3) The Examiner rejected claim 9 under 35 U.S.C. § I03(a) as
being obvious over the combination of Rasp 1, Andrade, and Imaki et al. (US
2013/0173587 Al; published July 4, 2013) (hereinafter "Imaki"). Final Act.
8-9; Ans. 8-9.
( 4) The Examiner rejected claim 6 under 3 5 U.S. C. § 10 3 (a) as
being obvious over the combination of Rasp 1, Andrade, and Doherty et al.
(US 8, 189,479 B 1; patented May 29, 2012) (hereinafter "Doherty"). Final
Act. 9-10.
Appellants ' Contentions in the Appeal Brief
Appellants contend, inter alia (see generally App. Br. 8-14) the
following:
( 1) the language of claims 1-9 and 20 meets the definiteness
requirement of 35 U.S.C. § 112(b) (App. Br. 3-5);
(2) the combination of Raspl and Andrade does not teach or suggest
deploying at least one helper operator that includes logic for multiple
operators in the flow graph (App. Br. 5-9);
(3) the combination of Raspl and Andrade does not teach or suggest
the claim 2 limitation directed to one operator becoming a bottleneck by
3
Appeal2018-005740
Application 14/308,993
processing incoming data tuples at a rate less than a rate of receiving the
incoming data tuples (App. Br. 9);
(4) the combination ofRaspl and Andrade does not teach or suggest
the claim 8, limitation that the helper operator processes data tuples in
parallel (App. Br. 10-11);
( 5) the combination of Raspl and Andrade does not teach or suggest
the claim 20 limitation that the helper operator includes logic for all of the
plurality of operators in the flow graph (App. Br. 11 );
( 6) the combination of Rasp 1, Andrade, and Imaki lacks any teaching
of dynamically creating and destroying helper operators as required by claim
9 (App. Br. 12-13); and
(7) the combination of Rasp 1, Andrade, and Doherty nowhere teaches
or suggests than an individual operator detects when it becomes a bottleneck
and sends a notification to the streams manager as required by claim 6 (App.
Br. 13).
ANALYSIS
We have reviewed the Examiner's rejections (Final Act. 4--14) in light
of Appellants' contentions that the Examiner has erred (App. Br. 3-13;
Reply Br. 2-8). Further, we have reviewed the Examiner's response to
Appellants' arguments (Ans. 4--9). We agree with Appellants' contentions
that the Examiner erred in rejecting claims 1-9 and 20 under 35 U.S.C.
§ 112(b ). However, we disagree with Appellant's contentions that the
Examiner erred in rejecting claims 1-9 and 20 under 35 U.S.C. § 103.
4
Appeal2018-005740
Application 14/308,993
Rejections Under 35 US.C. § l l 2(b)
The Examiner has rejected claims 1-9 and 20 under 35 U.S.C.
§ 112(b) because the Examiner found the claim term "helper operator" in
independent claim 1 to be unclear. Ans. 4. The Examiner identifies it is
unclear whether the term "helper operator" is referring to "an operator" that
is "nothing but 'logic for multiple operators,"' and further it is unclear,
"what exactly the 'logic or helper operator' does in the flow graph." Id.
( emphasis omitted).
Appellants' responses have persuaded us of error in the Examiner's
rejection. The Appellants explain the "helper operator" is an operator that
includes logic for other operators. In the Appeal Brief, Appellants refer to
the Specification's exemplary disclosure of a "helper operator K [which]
includes logic for operator F." App. Br. 3--4; Spec. ,r 73. Appellants
contend that the "logic for operator F" is "logic that causes operator F to
operate on tuples in some defined way." App. Br. 3.
Both the Examiner and Appellants thus appear to agree that a "helper
operator" is an operator that contains logic for other operators. We note that
the Specification discloses an operator as having logic to process tuples it
receives. App. Br. 3; Spec. ,r 73. The Examiner has not explained
persuasively why the term "helper operator" would have an indefinite claim
scope.
Nor has the Examiner explained persuasively why the claimed
"streams manager adjusts the at least one helper operator by connecting the
input and the output of the helper operator to the flow graph to alleviate the
bottleneck" would not be considered by one having ordinary skill in the art
to have a clearly defined scope. We are persuaded by Appellants'
5
Appeal2018-005740
Application 14/308,993
explanation that when an operator becomes a bottleneck, a helper operator
can be placed into the flow graph in parallel with the operator to process
data tuples in parallel with the operator, thereby relieving the bottleneck at
the operator. App. Br. 4. Appellants point to Specification and Figures to
illustrate how this accomplished in the flow graph. Id.; Spec. ,r,r 7 6-77,
Figs. 9--10.
The Examiner has not explained persuasively why one having
ordinary skill in the art would not understand what is claimed when claims
1-9 and 20 are read in light of the Specification. Accordingly, we do not
sustain the Examiner's rejection of claims 1-9 and 20 under§ 112.
Rejections Under 35 US.C. § 103
All of the rejections of claims 1-9 and 20 under 35 U.S.C. § 103 are
based on the combination of the teachings or suggestions of the
Andrade '046 reference in combination with one or more other prior art
references. The Appeal Brief did not address the Andrade '046 reference
that was applied in the rejections that are on appeal. Reply Br. 4--5. Instead,
the Appeal Brief mistakenly referred to content from a different publication
by the same named inventor, U.S. Patent Application Publication
2010/0293535 Al (Nov. 18, 2010) to Andrade et al. (hereinafter
"Andrade '535"). 2 This confusion appears to be rooted in the prior use of
the Andrade '535 reference in the prosecution of the patent application. 3
2 See particularly the quotation attributed to paragraph 22 of "Andrade," and
the reliance on a text search of "Andrade" in the Appeal Brief at pages 7-8.
3 The Andrade '535 reference was applied in rejections made in Office
Actions issued on January 26, 2016, and on August 2, 2016. The Examiner
withdrew these rejections by reopening prosecution on April 14, 2017, with
a Non-Final Action rejecting claims over the combination of Raspl and
6
Appeal2018-005740
Application 14/308,993
Appellants admit the error in quoting from the incorrect reference, but do not
allege any defect in the Non-Final Action, the Final Action, or the
Examiner's Answer that caused any confusion as to the identity of the
references applied in the rejections. Reply Br. 4--5. To the extent that
Appellants instead have argued in the Appeal Brief against features of the
Andrade '535 reference, which was not applied in any claim rejection, such
arguments are not persuasive to overcome the rejections based upon the
Andrade '046. To the extent that in the Reply Brief, Appellants for the first
time present arguments against the teachings and suggestions of the Andrade
'046 references, such arguments are waived. These arguments could have
been presented in the Appeal Brief, were not prompted by the Examiner's
Answer, and are not based on any new arguments or grounds of rejection in
the Examiner's Answer. As a result, Appellants have waived such untimely
argument because Appellants have not shown good cause for belatedly
raising the new argument, in accordance with 37 C.F.R. § 4I.41(b)(2).
Claims 1, 3-5, and 7
With respect to claim 1, Appellants contend that the combination of
applied references do not teach or suggest the limitation "at least one helper
operator that includes logic for multiple operators in the flow graph." App.
Br. 6; Reply Br. 6. The Examiner finds that Andrade teaches or suggests
this limitation. We concur with the conclusions reached by the Examiner,
and highlight and amplify certain teachings and suggestions of the Andrade
reference.
Andrade '046. The Final Action repeated the rejection made in the Non-
Final Action.
7
Appeal2018-005740
Application 14/308,993
As explained by the Examiner, Andrade teaches a stream processing
application having multiple operators for processing streaming data. Ans. 6;
Andrade ,r,r 18, 24. The stream processing application organizes these
operators into groups, described as "high availability groups." Ans. 6;
Andrade ,r,r 21-22. The Examiner explains that Andrade describes, at
Figure 1 and ,r 22, a policy defining how the operators in the high
availability groups permit the application to recover from failure. Ans. 6.
We understand this to refer to the teaching or suggestion in Andrade that for
each high availability group of operators, the application creates a replica of
an identical number of processing elements to contain the logic of all of the
operators within the high availability group. Andrade ,r 22, Fig. 1 as
described at ,r,r 23-26. These replicas are outside the flow graph until a
bottleneck occurs at an operator within the flow graph. Id.; see also Figs.
2-3, ,r,r 28-30. Andrade's teaching or suggestion of a "replica" of a high
availability group of operators, containing the logic of all of those operators
therein, therefore, teaches or suggests the claimed "helper operator that
includes logic for multiple operators in the flow graph." Accordingly, we
sustain the Examiner's obviousness rejection of claim 1.
Dependent claims 3-5 and 7 were not argued separately from claim 1,
and the rejections of those claims are sustained for the same reasons as for
claim 1.
Claim 2
With respect to claim 2, Appellants contend the Raspl reference does
not teach or suggest "the one operator becomes a bottleneck by processing
incoming data tuples at a rate less than a rate of receiving the incoming data
tuples," because Raspl discusses bottlenecks at network connections and not
8
Appeal2018-005740
Application 14/308,993
streaming applications. App. Br. 9; Reply Br. 6. In the rejection of claim 1,
from which claim 2 depends, the Examiner found that the combination of
Raspl and Andrade teaches or suggests determining when at least one of the
operators in the streaming application becomes a bottleneck. Appellants
have not contended otherwise in either the Appeal Brief or the Reply Brief.
With respect to claim 2, the Examiner points to Raspl for the teaching of a
bottleneck being identified by decreased data throughput at a point in the
data flow. Ans. 7; Raspl ,r 10. We concur with the Examiner's reasoning,
and the Examiner conclusion that the limitation "the one operator becomes a
bottleneck by processing incoming data tuples at a rate less than a rate of
receiving the incoming data tuples" is taught by the combination of Raspl
and Andrade as set forth in the rejection of claim 2.
To the extent that Appellants contend that Raspl alone does not teach
the entirety of the limitation, such a contention is unpersuasive as one cannot
show nonobviousness by attacking references individually where the
rejections are based on combinations of references. In re Keller, 642 F .2d
413 (CCPA 1981); In re Merck & Co., 800 F.2d 1091 (Fed. Cir. 1986).
Here, the Examiner has explained persuasively how the teaching or
suggestion of Rasp 1, in view of the combined teachings or suggestions of
Raspl and Andrade, teaches or suggests the limitations of claim 2.
Accordingly, we sustain the Examiner's obviousness rejection of claim 2.
Claim 8
With respect to claim 8, Appellants contend "the [E]xaminer has not
identified which features of F[ig]. 7 of Andrade allegedly read on the one
operator and the helper operator in claim 8." App. Br. 10; Reply Br. 7. As
discussed for claim 1, supra, the replica of the group of operators having a
9
Appeal2018-005740
Application 14/308,993
common "high availability policy," teaches or suggests the claimed helper
operator. Furthermore, Andrade teaches or suggests applying such groups of
operators in parallel with operators in the flow chart to process data tuples in
the event of failure at an operator in the flow chart. Final Act. 7; Andrade
Fig. 7; see also Andrade Figs. 2A, 2B, 4A, 4B, 5A, 5B. As explained by the
Examiner, this functionality allows the application to recover from failure at
an operator and to be highly available. Ans. 8. The Examiner has therefore
persuasively shown how the combination of Raspl and Andrade teaches the
disputed limitations of claim 8.
Appellants contend "F[ig.] 7 of Andrade would only support the
[E]xaminer's rejection if one of the operators in F[ig]. 7 were identified as a
helper operator that implements logic for the one operator and processes data
tuples in parallel with the one operator." App. Br. 10. Appellants' argument
is not persuasive because Appellants do not refer to the Andrade reference
that was applied in the rejection. Accordingly, we sustain the Examiner's
obviousness rejection of claim 8.
Claim 20
With respect to claim 20, Appellants contend that Raspl does not
teach or suggest a helper operator that includes logic for all of the plurality
of operators in the flow graph. The Examiner's rejection of claim 20 states
"Raspl-Andrade teaches the apparatus as recited in claim 1, wherein the at
least one helper operator includes logic for all of the plurality of operators in
the flow graph." Final Act. 8. As discussed for claim 1, supra, the Andrade
reference teaches or suggests one helper operator including logic for
multiple operators. The number of operators in the flow graph of Andrade
may be "one or more," for example, the number of operators is illustrated in
10
Appeal2018-005740
Application 14/308,993
numerous flow graphs as two. Andrade ,r 18, Figs. 2A, 2B, 4A, 4B, 5A, 5B.
In those examples, the helper operator includes a replica of the logic for both
of the operators, i.e., all of the operators in the flow graph, thereby resulting
in a helper operator having logic for all of the operators in the flow graph.
To the extent that Appellants contend that Raspl alone does not teach
the entirety of the limitation, such a contention is unpersuasive as one cannot
show nonobviousness by attacking references individually where the
rejections are based on combinations of references. Keller, 642 F.2d 413;
Merck, 800 F .2d 1091. Here, the combined teachings or suggestions of
Raspl and Andrade teach or suggest the limitations of claim 20.
Furthermore, Appellants' Appeal Brief did not address the combined
teachings of Raspl and the Andrade reference that was applied in the
rejection; the Examiner's Answer pointed out that the Appellant's
contentions regarding claims 1-5, 7-8, and 20 were addressing the wrong
Andrade reference. Ans. 5---6. Accordingly, we sustain the Examiner's
obviousness rejection of claim 20.
Claim 9
With respect to claim 9, Appellants contend that Imaki provides no
teaching of dynamically creating helper operators as claimed. App. Br. 8.
Appellants do not contest that Imaki teaches or suggests dynamically
deleting helper operators. App. Br. 8. Citing Imaki at ,r,r 40, 134, 140, and
160, the Examiner states "Imaki disclose[ s] the process of deleting operators
and constructing operators (equivalent to helper operator) in operator tree as
needed during execution." Ans. 8-9. As described at ,r,r 134, 140, and 160,
Imaki teaches arranging an order of operators via an operator execution
control unit used at the time of execution, wherein operators may be
11
Appeal2018-005740
Application 14/308,993
arranged and prior operator trees may be "reconstructed." We concur with
this reasoning and the Examiner's conclusion that Imaki teaches
dynamically creating operators as needed. We also agree with the Examiner
that the combination of Imaki, Andrade, and Raspl teaches or suggests
dynamically creating helper operators as claimed.
Appellant further contends a lack of a teaching or suggestion of a
"helper operator" in claim 1 from which claim 9 depends. Reply Br. 7. As
we have discussed, supra, we agree with the Examiner's conclusion that the
combination of Raspl and Andrade teaches the helper operator as claimed.
Accordingly, we sustain the Examiner's obviousness rejection of claim 9.
Claim 6
With respect to claim 6, Appellants contend that although the
combination of Rasp 1, Andrade, and Doherty references teaches or suggests
a node containing operators reporting a bottleneck, the combination does not
teach or suggest an operator reporting the bottleneck to the streams manager.
App. Br. 13; Reply Br. 8.
We are not persuaded that the combination of Rasp 1, Andrade, and
Doherty references fails to teach reporting the bottleneck to the streams
manager. As explained by the Examiner in the rejection of claim 1, Raspl
teaches or suggests the streams manager monitoring performance of the
operators. Final Act. 5. Doherty further teaches or suggests an operator
sending a bottleneck notification. Final Act. 9. The Examiner concludes,
and we concur, that it "would have been obvious ... to modify the apparatus
of Raspl-Andrade to incorporate the one operator detecting when the one
operator becomes a bottleneck and sending a notification to the streams
manager." Final Act. 10.
12
Appeal2018-005740
Application 14/308,993
To the extent that Appellants limit the combined teaching merely to
the node-level reporting of Doherty, without considering the combination of
the teachings or suggestions of Doherty and Raspl to teach or suggest an
operator reporting the bottleneck to the streams manager, Appellants'
contention is not persuasive. The test for obviousness is "what the combined
teachings of the references would have suggested to those of ordinary skill
in the art," Keller, 642 F.2d at 425. Furthermore, as noted supra,
Appellants' Appeal Brief did not address the Andrade reference that was
applied in the rejection. Accordingly, we sustain the Examiner's
obviousness rejection of claim 6.
DECISION
We reverse the Examiner's rejection of claims 1-9 and 20 under
35 U.S.C. § 112(b ).
We affirm the Examiner's rejections of claims 1-9 and 20 under
35 U.S.C. § 103.
Because we affirm at least one ground of rejection with respect to
each claim on appeal, the Examiner's decision rejecting claims 1-9 and 20 is
affirmed. See 37 C.F.R. § 41.50(a)(l ).
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