Ex Parte Hertz et al

Board of Patent Appeals and Interferences

Jan 22, 2009

10303125 (B.P.A.I. Jan. 22, 2009)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte MARK D. HERTZ, STEPHEN R. CORNABY,
TRAVIS D. FOX, and EDWIN S. OLDS
____________

Appeal 2008-3120
Application 10/303,125
Technology Center 2100
____________

Decided: January 22, 2009
____________

Before JAMES D. THOMAS, JEAN R. HOMERE,
and STEPHEN C. SIU, Administrative Patent Judges.
THOMAS, Administrative Patent Judge.

DECISION ON APPEAL
I. STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s
final rejection of claims 21 through 40. We have jurisdiction under
Appeal 2008-3120
Application 10/303,125

2
35 U.S.C. § 6(b). An Oral Hearing was conducted in this appeal on
November 18, 2008.
We affirm.
A. THE INVENTION
Appellants’ invention is a data retention prioritization technique in a
cache memory device. In use with a cache memory system, read data is
characterized as host data and speculative data. Host data is that data
actually requested and used by the host processor. Speculative data includes
data preceding the actual host data that is characterized as read on arrival
(ROA) data. Correspondingly, speculative data also includes data following
the actual host data that is characterized as read look ahead (RLA) data.
Once the actual host data is utilized by the processor, Appellants’ invention
prioritizes the remaining speculative data, that is, prioritizes the persistence
of the remaining data to be retained. (Spec. 3 and 23).
The claimed version of this disclosed approach to data management
characterizes the host data actually read by the host computer as requested
data and the speculative data as non-requested data. The point of contention
revolves around the persistence of the non-requested data greater in time
than the persistence of the requested data in the cache memory. (Appeal Br.
6 and Answer 3-4).

B. ILLUSTRATIVE CLAIM
Representative claim 21 is reproduced below:
Appeal 2008-3120
Application 10/303,125

3
21. An apparatus comprising a cache memory which receives data
comprising requested data and non-requested data, and a controller coupled
to the cache memory which respectively prioritizes the requested data and
the non-requested data to facilitate a persistence of the non-requested data in
said cache memory for a period of time greater than a persistence of the
requested data in said cache memory.

C. PRIOR ART AND REJECTIONS
The prior art relied by the Examiner in rejecting the claims on appeal
is:
Rubin US 5,812,996 Sep. 22, 1998
Applicants’ Admitted Prior Art (AAPA); see Specification, page 1, l. 11 –
page 2, l. 29.
Claims 21-40 stand rejected under 35 U.S.C. § 103. As evidence of
obviousness, the Examiner relies on Appellants’ admitted prior art in view
of Rubin.

II. ISSUES
The first issue is whether the Appellants have shown that the
Examiner has erred in finding that the admitted prior art is properly
combinable with Rubin within 35 U.S.C. § 103. The second issue is
whether the combination of references teaches that the persistence of the
non-requested data is greater than the persistence of the requested data.

Appeal 2008-3120
Application 10/303,125

4
III. PRINCIPLES OF LAW
As to the obviousness rejections, we note the following principles:
“To reject claims in an application under section 103, an
examiner must show an unrebutted prima facie case of
obviousness…. On appeal to the Board, an applicant can
overcome a rejection [under § 103] by showing insufficient
evidence of prima facie obviousness or by rebutting the prima
facie case with evidence of secondary indicia of
nonobviousness.”

In re Kahn, 441 F.3d 977, 985-86 (Fed. Cir. 2006)(quoting In re Rouffet,
149 F.3d 1350, 1355 (Fed. Cir. 1998)).
Section 103 forbids issuance of a patent when “the differences
between the subject matter sought to be patented and the prior art are
such that the subject matter as a whole would have been obvious at
the time the invention was made to a person having ordinary skill in
the art to which said subject matter pertains.”

KSR Int'l Co. v. Teleflex Inc., 127 S. Ct. 1727, 1734 (2007).

The Supreme Court reaffirmed principles based on its precedent that
“[t]he combination of familiar elements according to known methods is
likely to be obvious when it does no more than yield predictable results.”
KSR, 127 S. Ct. at 1739. The operative question in this “functional
approach” is thus “whether the improvement is more than the predictable use
of prior art elements according to their established functions.” Id. at 1740.
The Court noted that “[c]ommon sense teaches . . . that familiar items may
have obvious uses beyond their primary purposes, and in many cases a
Appeal 2008-3120
Application 10/303,125

5
person of ordinary skill will be able to fit the teachings of multiple patents
together like pieces of a puzzle.” Id. at 1742.
The Federal Circuit recently concluded that it would have been
obvious to combine (1) a device for actuating a phonograph to play back
sounds associated with a letter in a word on a puzzle piece with (2) a
processor-driven device capable of playing the sound associated with a first
letter of a word in a book. Leapfrog Enters., Inc. v. Fisher-Price, Inc.,
485 F.3d 1157, 1161 (Fed. Cir. 2007). In reaching that conclusion, the
Federal Circuit recognized that “[a]n obviousness determination is not the
result of a rigid formula disassociated from the consideration of the facts of a
case. Indeed, the common sense of those skilled in the art demonstrates why
some combinations would have been obvious where others would not.” Id.
at 1161 (citing KSR, 127 S. Ct. 1727, 1739 (2007)). The Federal Circuit
relied in part on the fact that Leapfrog had presented no evidence that the
inclusion of a reader in the combined device was “uniquely challenging or
difficult for one of ordinary skill in the art” or “represented an unobvious
step over the prior art.” Id. (citing KSR, 127 S. Ct. at 1740-41).
In the absence of separate arguments with respect to claims subject to
the same rejection, those claims stand or fall with the claim for which an
argument was made. See In re Young, 927 F.2d 588, 590 (Fed. Cir. 1991).
See also 37 C.F.R. § 41.37(c)(1)(vii)(2004).
Dovetailing with this precedent, we note further that the test for
obviousness has been further characterized as not whether the features of a
Appeal 2008-3120
Application 10/303,125

6
secondary reference may be bodily incorporated into the structure of a
primary reference. It is also not that the claimed invention must be
expressly suggested in any one or all of the references. Rather, the test is
what the combined teachings of the references would have suggested to
those of ordinary skill in the art. In re Keller, 642 F.2d 414, 425 (CCPA
1981); In re Young, 927 F.2d 588, 591 (Fed. Cir. 1991).
The prior art relied on to prove obviousness must be analogous art.
As explained in Kahn,
the ‘analogous-art’ test . . . has long been part of the primary Graham
analysis articulated by the Supreme Court. See Dann [v. Johnston,]
425 U.S. [219,] 227-29 (1976), Graham, 383 U.S. at 35. The
analogous-art test requires that the Board show that a reference is
either in the field of the applicant’s endeavor or is reasonably
pertinent to the problem with which the inventor was concerned in
order to rely on that reference as a basis for rejection. In re Oetiker, at
1447. References are selected as being reasonably pertinent to the
problem based on the judgment of a person having ordinary skill in
the art. Id. (“[I]t is necessary to consider ‘the reality of the
circumstances,’- in other words, common sense-in deciding in which
fields a person of ordinary skill would reasonably be expected to look
for a solution to the problem facing the invention.” (quoting In re
Wood, 599 F.2d 1032 (C.C.P.A. 1979)).

Kahn, 441 F.3d at 986-87. See also In re Clay, 966 F.2d 656, 659 (Fed. Cir.
1992) (“[a] reference is reasonably pertinent if, even though it may be in a
different field from that of the inventor’s endeavor, it is one which, because
of the matter with which it deals, logically would have commended itself to
an inventor’s attention in considering his problem.”).
Appeal 2008-3120
Application 10/303,125

7
IV. FACTS AND ANALYSIS
To the extent Appellants urge at page 2 of the Reply Brief that the
positions of the Examiner in the Answer constitute a new ground of
rejection, this argument is unavailing. Even if we were to agree with
Appellants’ views that the Answer presented such a radical new thrust of
analysis so as to constitute a new ground of rejection, Appellants filed a
Reply Brief, which was entered by the Examiner, and had the opportunity to
argue the positions at Oral Hearing. To the extent there was an actual
administrative due process issue presented by the Examiner, these latter
events have obviated it.
Moreover, we essentially agree with Appellants’ basic urging in the
Reply Brief, particularly at page 6 of it, that the artisan may understand the
admitted prior art relied upon by the Examiner as indicating that the term
“read data” constitutes both the claimed requested data and the non-
requested data, which are handled as a single unit to the extent indicated at
Specification page 2, lines 12-14. This appears to be consistent with the
approach of the prior art characterized as Least Recently Used (LRU) data
management principles which are briefly discussed in Rubin as well.
Although we do not agree with certain portions of the Examiner’s
reliance upon the admitted prior art at Specification page 2, the Examiner’s
reasoning of combinability of this admitted prior art as a whole with Rubin
appears to comply with the requirements of the above-noted case law. To
the extent pages 10-12 of the Reply Brief attempt to distinguish over the
Appeal 2008-3120
Application 10/303,125

8
United States Supreme Court’s analysis in KSR, the position is equally
unavailing. Both the admitted prior art and Rubin relate to optimizing use of
buffer caches in a data processing system. Therefore, the teachings of the
applied prior art are properly combinable within 35 U.S.C. § 103 as urged by
the Examiner.
What we find significant initially is that there are additional teachings
of the admitted prior art at Specification page 2 that have been mentioned
but unappreciated by both the Appellants and the Examiner. At lines 3
through 6, it is recognized that there is a continuing need in the art to
improve throughput performance for data storage devices particularly based
upon industry standard metrics characterized as benchmarks. As discussed
at lines 16 through 20, with respect to benchmark commands relative to file
caching, it is recognized that host read data is distinguished from other kinds
of data characterized as speculative data which is also recognized as being
often requested. Thus, the art recognizes that the actual data read by the host
processor is rarely re-requested while, at the same time, it is recognized that
the speculative data or the actual non-requested data of the claimed
invention and of the read data as a whole is actually often requested. Lines
21 through 23 also recognize that the artisan is well aware of the fact that
requests for such speculative data portion of the read data as a whole often
occurs after the read data as a whole has been jettisoned from the cache
memory.
Appeal 2008-3120
Application 10/303,125

9
We recognize that a thrust of the Examiner’s reasoning is the relative
persistence of the type of data characterized as non-requested data and
requested data of the claimed invention. It appears to us that the artisan
would well recognize that there is a relative persistence of the just discussed
admitted prior art to the extent recited broadly in the claims on appeal.
Moreover, this position is buttressed by the actual teachings of Rubin as the
reader will now appreciate.
The data base management system of Rubin employs an Optimizer
and a Buffer Manager to determine if the actual data sought is in the buffer
cache memory at all or not. The formulated search query strategy or plan
utilizes so-called hints according to the summary of the invention beginning
at column 2 through column 4. Which of two hints are actually processed or
utilized is based upon interactions between the Optimizer and the Buffer
Manager.
According to the Specification at page 2, lines 7-8, “additional non-
requested read data [is] spatially adjacent to [that is, directly before and
directly after] the host-requested read data.” This is consistent with what the
artisan would recognize is taught at column 3, lines 22-29, as we reproduce
them here:
In an exemplary embodiment, two types of hints are employed:
prefetch size and replacement strategy. The “prefetch” size, the first
hint, is employed for I/O operations. “Prefetch” is the amount that the
system retrieves or fetches in addition to the amount requested. If, for
instance, a request is made for one data page, the system might
Appeal 2008-3120
Application 10/303,125

10
additionally fetch or “prefetch” the next page into cache, on the
likelihood that it might soon be needed.

Because so-called prefetch data includes data greater than the actual
requested and used read data, the prefetch size is recognized in Rubin to be
larger than the amount actually requested. The actual example given
includes the capability of prefetching data actually subsequent to the actual
data requested, which is consistent with the data which would be spatially
adjacent, that is, data that follows the actual host-requested read data of the
admitted prior art. The artisan would therefore well appreciate that both the
admitted prior art and Rubin’s teachings are consistent with respect to
managing the data in cache memory. Subsequent teachings in Rubin such as
those beginning at column 7 relate to the variability of the actual requested
size being determined by the interaction between Rubin’s Optimizer and
Buffer Manager.
With respect to the stated second hint, the following discussion in
Rubin beginning at column 3, line 50 through column 4, line 5 is equally
significant:
Replacement strategy, the second hint, is the strategy employed
for "buffer reuse." Here, the system adopts a "most-recently used"
(MRU) replacement strategy, as opposed to the conventionally-
employed LRU strategy. MRU replacement is the strategy of linking a
page which has been used back to the LRU (least-recently used) side
of the LRU chain, instead of the MRU side which is typically
employed. Thus, this hint indicates that the system suspects that it
might not need this particular page again, so the page is marked for
indicating that it should be preferentially "bumped" (paged out) from
Appeal 2008-3120
Application 10/303,125

11
the cache.

This "fetch and discard" or reverse strategy is useful in
situations where the system suspects that particular buffers will not be
reused. In such instances, the system knows that it would be
inefficient to "wipe out" cache--filling cache memory with data pages
which are not likely to be re-used (anytime soon). Accordingly,
buffers used for the query are linked to the end of the reuse chain such
that their buffers are available for immediate reuse. The MRU's
strategy is generated both by the particular syntax of a query and by
empirical observations of the Optimizer. The Optimizer's automatic
recognition can be disabled from the SQL query, if desired.

The reader will thus appreciate that the so-called Most Recently Used
(MRU) teachings in Rubin are consistent with those actually utilized in
Appellants’ disclosed invention and are distinguished from the admitted
prior art approach utilizing the above-noted LRU strategy. The significant
teaching of this quoted material is the actual use of a given page is treated as
an LRU PAGE. As such, the actual page used by the processor is essentially
“bumped” out of the cache memory, leaving the remaining portions directly
before and directly after the portion used when the prefetch size is large
enough to encompass more than the actual used or requested data to remain.
Thus, the artisan would have appreciated that the actual teachings here are
consistent with the extended persistence of the non-requested data feature of
representative independent claim 21 on appeal. Correspondingly, the artisan
would have appreciated that the combined teachings of the admitted prior art
and Rubin make it clear that, in the context of cache memory management,
Appeal 2008-3120
Application 10/303,125

12
the combined teachings of the references do teach or suggest that non-
requested data maybe made to persist longer in cache memory than the
actual requested or used data as recited in the claims on appeal.
Since, as indicated beginning at the bottom of page 3 of the principal
Brief on appeal, the subject matter of independent claims 21, 31, and 32 are
considered to be argued together, we have chosen independent claim 21 as
representative thereof. No arguments are presented in the principal Brief or
Reply Brief as to any other dependent claim except dependent claims 29, 30,
39, and 40 at page 14 of the principal Brief.
The Examiner respectively addresses the features of claims 29 and 39
in the discussion at the bottom of page 6 of the Answer in the initial
statement of the rejection. Likewise, the subject matter of dependent claims
30 and 40 are addressed by the Examiner in the statement of rejection at the
top of page 7 of the Answer. Our previous analysis of both the admitted
prior art and Rubin indicates that data characterized as RLA data occurring
after the requested data and data preceding the actual requested data and
characterized as ROA data were known in the art. Patentability will not be
predicated based upon the recitation of the term RLA or ROA in the noted
claims. Page 14 of the principal Brief does not actually contest what
evidence the Examiner relies upon in the statement of rejection noted earlier
in this paragraph, and no remarks are directed to all these claims in the
Reply Brief.

Appeal 2008-3120
Application 10/303,125

13
V. CONCLUSION
We conclude that Appellants have not shown that the Examiner has
erred in rejecting claims 21-40 under 35 U.S.C. § 103. We therefore affirm
the Examiner’s rejection of these claims on this statutory basis and conclude
that these claims are not patentable.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. §1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).

AFFIRMED

rwk

Fellers, Snider, Blankenship, Bailey & Tippens
Bank One Tower
Suite 1700
100 North Broadway
Oklahoma City, OK 73102-8820