Ex Parte Bonola

Board of Patent Appeals and Interferences

Sep 8, 2011

10980538 (B.P.A.I. Sep. 8, 2011)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________________

Ex parte THOMAS J. BONOLA
____________________

Appeal 2009-012907
Application 10/980,538
Technology Center 2100
____________________

Before LANCE LEONARD BARRY, JEAN R. HOMERE,
and THU A. DANG, Administrative Patent Judges.

DANG, Administrative Patent Judge.

DECISION ON APPEAL

Appeal 2009-012907
Application 10/980,538

2
I. STATEMENT OF THE CASE
Appellant appeals under 35 U.S.C. § 134(a) from a Final Rejection of
claims 1, 2, and 4-19. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.

A. INVENTION
Appellant’s invention relates to a system and method to coordinate
access to a sharable data structure using deferred cycles; wherein, control
logic within the system implements a deferred cycle mechanism that issues a
deferred-reply message to a requesting agent (e.g., a central processing unit)
which causes the spin lock loop software of the requesting agent to stop
execution until the control logic responds again that the requested data
structure (lock variable) is available (Abstract; Spec. ¶ [0015]).

B. ILLUSTRATIVE CLAIM
Claim 1 is exemplary:
1. A system, comprising:

a plurality of requestors, each capable of accessing a
sharable data structure; and

control logic coupled to said requestors, said control
logic causes a request for access to the sharable data structure to
be deferred to coordinate access to said sharable data structure
to permit only one requestor at a time to access said sharable
data structure;

wherein said control logic issues a deferred reply if a
lock variable indicates that a requestor has exclusive access to
said sharable data structure, said deferred reply causes another
requestor to cease requesting access to said sharable data
Appeal 2009-012907
Application 10/980,538

3
structure and to wait for said control logic to indicate that the
sharable data structure is available to the another requestor.

C. REJECTION
The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Anderson, “The Performance of Spin Lock Alternatives for
Shared-Memory Multiprocessors,” 1 IEEE Transactions on
Parallel and Distributed Systems 6-26 (January 1990).

Claims 1-2 and 4-19 stands rejected under 35 U.S.C. § 102(b) as
being anticipated by Anderson.

II. ISSUE
The dispositive issue before us is whether the Examiner has erred in
finding that Anderson teaches a system having control logic that “issues a
deferred reply if a lock variable indicates that a requestor has exclusive
access to said sharable data structure, said deferred reply causes another
requestor to cease requesting access to said sharable data structure and to
wait for said control logic to indicate that the sharable data structure is
available to the another requestor” (claim 1, emphasis added).

Appeal 2009-012907
Application 10/980,538

4
III. FINDINGS OF FACT
The following Findings of Fact (FF) are shown by a preponderance of
the evidence.
Anderson
1. Anderson is directed to spin lock alternatives for shared-
memory multiprocessors; wherein, instead of leaving the processors
spinning (or “busy-waiting”), the system uses software algorithms similar to
arbitration schemes employed in a distributed system for control of a
network to enhance the performance and reduce the cost of spin-waiting
(Abstract).
2. Atomic instructions are used to arbitrate between simultaneous
attempts to acquire the lock; but, when the lock is busy, a waiting process is
implemented in software that can either block the processor from attempting
to acquire the lock, relinquishing the processor to do other work until the
lock is released or to spin (‘busy-wait’) until the lock is released (p.6, right
col., third para.).
3. Another embodiment includes a method for explicitly queuing
spinning processors; wherein, when the processors arrive at a lock, each
acquires a unique sequence number specifying the order in which they have
to wait for access to the data and, when the lock is released, control is
directly passed to the next processor in line (Abstract; p. 7, left col., second
para.).
Appeal 2009-012907
Application 10/980,538

5

IV. ANALYSIS
Claims 1, 2, 4-19
Appellant does not provide a separate argument with respect to
independent claims 1, 10, 12 and 14 (App. Br. 11-14). Appellant does not
provide arguments with respect to dependent claims 2-9, 11, 13, and 15-19.
Accordingly, we select claim 1 as being representative of the claims. See 37
C.F.R. § 41.37(c)(1)(vii).
Appellant contends that “Anderson does not explicitly or inherently
state … that the processor performs other work until it receives a signal
from a control logic indicating that the lock is available” (App. Br. 11).
In particular, Appellant asserts that “Anderson does not teach or suggest
such a deferred reply”; rather, “Anderson appears to teach that the
requesting entity (controller or processor) ceases requesting access to the
lock and performs other activities until the requesting entity ‘decides’ to
resume the lock-acquisition process” (App. Br. 12). According to Appellant,
“Anderson does not wait for an indication that the sharable data
structure is available”; rather, “Anderson only discloses that the processor
is ‘relinquished to do other work’ - no mention is made of waiting for a
signal indicating that the lock is available (i.e., no explicit teaching), nor is
such a signal even necessary (i.e., no inherent teaching)” (App. Br. 13).
Appellant contends that “the processor [of Anderson] ‘spins’ by
continuously trying to acquire the lock, or by spin-reading (continuously
reading) a cached lock value” (id.).
However, the Examiner finds that when Appellant argues that
Anderson must disclose “a signal from a control logic indicating that the
Appeal 2009-012907
Application 10/980,538

6
lock is available. . . . Appellant is reading limitations of the [S]pecification
into the claims to thereby narrow the scope of the claims by implicitly
adding disclosed limitations which have no express basis in the claims”
which “is impermissible” (Ans.10). The Examiner finds that “Anderson
clearly discloses ‘arbitrating between simultaneous attempts to acquire a
lock[’] wherein [‘]if the lock is busy, waiting is done in software[’] and
the waiting process block[s] the processor’s request; at a later time, the
processor may resume its attempts to acquire the lock” (Ans. 12 (quoting
Anderson, p. 6, col. 2)). The Examiner notes that “Anderson unequivocally
discloses ‘[w]hen a lock is busy, the waiting process can either block,
relinquishing the processor to do other work … (which has the same effect
as issuing a deferred reply) until the lock is released’” and “‘[w]hen the lock
is released, control can be directly passed to the next processor in line’”
(Ans. 14 (quoting Anderson, pp. 6,7).
In the Reply Brief, Appellant merely repeats the same arguments and
contends further that Anderson’s processors “tries again of its own
volition” and “does not appear to wait for any indication from the control
logic” (Reply Br. 3).
Appellant’s argument that “Anderson does not explicitly or inherently
state … that the processor performs other work until it receives a signal
from a control logic indicating that the lock is available” (App. Br. 11;
emphasis added) is not commensurate in scope with the specific language of
claim 1. Similarly, Appellant’s argument that “no mention is made [in
Anderson] of waiting for a signal indicating that the lock is available” (App.
Br. 13; emphasis added) is not commensurate in scope with the specific
language of claim 1. In particular, claim 1 does not recite such “a signal” as
Appeal 2009-012907
Application 10/980,538

7
Appellant argues. Thus, we find in this Appeal whether Anderson discloses
control logic that issues a deferred reply which causes another requestor to
cease requesting access to the sharable data structure and wait for the control
logic to indicate that the sharable data structure is available as specifically
required by claim 1.
To determine whether Anderson teaches a system and method to
coordinate access to a sharable data structure having control logic, wherein
“said control logic issues a deferred reply if a lock variable indicates that a
requestor has exclusive access to said sharable data structure, said deferred
reply causes another requestor to cease requesting access to said sharable
data structure and to wait for said control logic to indicate that the sharable
data structure is available to the another requestor” as recited in claim 1, we
give the claim its broadest reasonable interpretation consistent with the
Specification. See In re Morris, 127 F.3d 1048, 1054 (Fed. Cir. 1997).
However, we will not read limitations from the Specification into the claims.
In re Van Geuns, 988 F.2d 1181, 1184 (Fed. Cir. 1993).
Claim 1 does not place any limitation on what “deferred reply” means,
includes, or represents other than it “causes another requestor to cease
requesting access to said sharable data structure and to wait for said control
logic to indicate that the sharable data structure is available.” Thus, we give
“deferred reply” its broadest reasonable interpretation as any instruction that
instructs a requestor to stop requesting access and to wait until the data is
accessible, as specifically defined in claim 1.
Anderson is directed to spin lock alternatives for shared-memory
multiprocessors; wherein, instead of leaving the processors spinning (or
“busy-waiting”), the system uses software algorithms similar to arbitration
Appeal 2009-012907
Application 10/980,538

8
schemes employed in a distributed system for control of a network (FF 1).
Atomic instructions are used to arbitrate between simultaneous attempts to
acquire the lock; but, when the lock is busy, a waiting process is
implemented in software that can block the processor from attempting to
acquire the lock, relinquishing the processor to do other work until the lock
is released (FF 2). We find that the software code that implements the
waiting process represents the control logic (FF2). We find further that, by
blocking the processor until the lock is released, the waiting process not only
denies the processor access to the shared data until the lock is released but
also instructs the processor to stop attempting to acquire access since it
relinquishes the processor to do other work (FF 2). Accordingly, we find
that this feature of the waiting process represents a deferred reply that causes
processor to stop requesting access to the sharable data structure and to wait
for the waiting process to indicate that data is available.
Anderson further includes a method for explicitly queuing spinning
processors; wherein, when the processors arrive at a lock, each acquires a
unique sequence number specifying the order in which they have to wait for
access to the data and, when the lock is released, control is directly passed to
the next processor in line (FF 3). We find that the queuing software that
implements this method of queuing spinning processors also represents
control logic that makes the processors wait (without spinning); wherein, the
queuing software issues control to the next processor when the data becomes
available (FF 3).
Therefore, in view of our claim construction above, we find that
Anderson discloses “deferred reply causes another requestor to cease
requesting access to said sharable data structure and to wait for said control
Appeal 2009-012907
Application 10/980,538

9
logic to indicate that the sharable data structure is available,” as specifically
required by claim 1.
Accordingly, we find that Appellant has not shown that the Examiner
erred in rejecting claims 1, 10, 12 and 14 and claims 2, 4-9, 11, 13, and 15-
19 depending from claims 1, 10, 12 and 14 under 35 U.S.C. § 102(b) over
Anderson.
V. CONCLUSION AND DECISION
The Examiner’s rejection of claims 1, 2, and 4-19 under 35 U.S.C.
§ 102(b) is affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED

peb