Ex Parte Graefe et al

Patent Trial and Appeal Board

Aug 4, 2016

13562906 (P.T.A.B. Aug. 4, 2016)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/562,906 07 /31/2012 Goetz Graefe
56436 7590 08/08/2016
Hewlett Packard Enterprise
3404 E. Harmony Road
Mail Stop 79
Fort Collins, CO 80528
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.
82962608 4803
EXAMINER
PHAM, KHANH B
ART UNIT PAPER NUMBER
2166
NOTIFICATION DATE DELIVERY MODE
08/08/2016 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):
hpe.ip.mail@hpe.com
mkraft@hpe.com
chris.mania@hpe.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte GOETZ GRAEFE, HIDEAKI KIMURA,
and HARUMI KUNO
Appeal2014-009363
Application 13/562,906
Technology Center 2100
Before CAROLYN D. THOMAS, JOSEPH P. LENTIVECH, and
KARA L. SZPONDOWSKI, Administrative Patent Judges.
SZPONDOWSKI, Administrative Patent Judge.
DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner's
Final Rejection of claims 1---6 and 8-21. Claim 7 has been cancelled. (App.
Br. 20). We have jurisdiction under 35 U.S.C. § 6(b).
We AFFIRM.
Appeal2014-009363
Application 13/562,906
STATEMENT OF THE CASE
Appellants' invention is directed to an early release of transaction
locks based on tags. Claims 1, 5, and 8, reproduced below with the disputed
limitations in italics, are representative of the claimed subject matter:
1. A computing system comprising:
a processor; and
a memory resource storing instructions that, when
executed by the processor, cause the processor to:
execute a first transaction associated with an update
to data and a release of at least one lock on the data prior to the
first transaction being durable;
transaction;
replace the at least one lock with at least one tag;
identify a read request corresponding to a second
associate the at least one tag with the second
trans action;
perform a comparison between one or more
characteristics of the at least one tag and a recovery log
associated with the data to determine whether the first transaction
is durable; and
delay a transaction commit for the second
transaction until the first transaction is durable.
5. The computing system of claim 1, wherein, in response to
delaying the transaction commit, the processor wakes a log
flusher to increment a flushed tail position of the recovery log.
8. The computing system of claim 1, wherein the processor
delays the transaction commit based on performing an integer
comparison between a first LSN corresponding to the at least one
tag and a second LSN corresponding to a flushed tail of the
recovery log upon acquiring or releasing the at least one lock.
2
Appeal2014-009363
Application 13/562,906
REJECTIONS
Claims 1---6 and 8-21 stand rejected under 35 U.S.C. § 102(b) as being
anticipated by Vishal Kathuria, et al., Transaction Isolation and Lazy
Commit, IEEE 23rd International Conference (2007) ("Kathuria").
ANALYSIS
After consideration of each of Appellants' arguments, we agree with
the Examiner. We refer to and adopt the Examiner's findings and
conclusions as set forth in the Examiner's Answer and in the action from
which this appeal was taken. (Ans. 2-7; Final Act. 7-13). Our discussions
here will be limited to the following points of emphasis.
Claim 1
Issue 1: Did the Examiner err in finding Kathuria discloses "replace
the at least one lock with at least one tag," as recited in independent claim 1?
Appellants contend "Kathuria [] dislose[s] marking 'a data page [that]
is modified by a lazy commit transaction,"' but "Kathuria makes no mention
of replacing a lock, such as the 'X-Lock' or 'S-Lock,' on the modified data
with the described page marker[.]" (App. Br. 7-9). Thus, according to
Appellants, the Examiner's findings do not show replacing the at least one
lock with at least one tag. (Reply Br. 3). Rather, Appellants argue the lock
remains on the page, and is therefore not replaced by the marking. (Reply
Br. 4). Appellants further argue the second transaction (T2 Durable Read
3
Appeal2014-009363
Application 13/562,906
Transaction) must still acquire a lock to perform the read transaction. (App.
Br. 8; Reply Br. 5-6).
We are not persuaded by Appellants' arguments and agree with the
Examiner's findings. (Final Act. 2-3; Ans. 7-10). Kathuria describes "lazy
commit" transactions, where the commit log record is logged in a log buffer,
but not synchronously written to stable storage. (Kathura § 5.1, p. 1206).
The locks held by these transactions are released immediately after the
commit log record has been written to the log buffer. (Id.). To ensure a
transaction has been durably committed before a reader requiring durability
can consume it, changes made by lazily committed transactions are tracked
by marking the page DIRTIED_BY_LC_XACT and maintaining a counter
called LazyCommitLSN. (Id. at 1207).
The Examiner finds the marking and the counter correspond to the
claimed "tag." (Ans. 7). As shown in Figure 1 of Kathuria, with respect to
the first transaction (T 1 ), the page is marked, the counter is set, and the lock
(X-Lock) is released. (Kathuria p. 1207). Appellants have not persuasively
explained why such disclosed marking does not replace the released lock.
Moreover, Appellants' arguments that the T2 Durable Read Transaction
must still acquire a lock to perform the read transaction are not persuasive
because the claim language does not prevent the second transaction from
acquiring a lock. We agree with the Examiner that the lock acquired by the
second transaction (S-Lock) is different than the lock acquired by the first
transaction (X-Lock). (See Ans. 8). As described above, the lock acquired
by the first transaction (X-lock) is released after the lazy commit.
Accordingly, we are not persuaded the Examiner erred in finding
Kathuria discloses the disputed limitation.
4
Appeal2014-009363
Application 13/562,906
Issue 2: Did the Examiner err in finding Kathuria discloses "associate
the at least one tag with the second transaction," as recited in independent
claim 1?
Appellants contend "no such association takes place in Kathuria."
(App. Br. 9, emphasis omitted). Specifically, Appellants argue "[ n ]owhere
does Kathuria disclose a processor or transaction manager making an
association between the marker, 'DIRTIED_BY_LC_XACT,' and 'T2
Durable Read Transaction."' (App. Br. 10; see also Reply Br. 7).
We are not persuaded by Appellants' arguments and agree with the
Examiner's findings. (Final Act. 2-3; Ans. 7-10). Initially, we note
Appellants have not explicitly defined the term "associate" in the
Specification or briefing. Paragraph 4 7 of Appellants' Specification states
In block 730, a second transaction is identified, the second
transaction associated with acquiring the at least one tag based
on a read of the data. For example, the transaction manager
identifies that the second transaction reads data that is
uncommitted or otherwise associated with a not-yet-durable
transaction, including identifying one or more dependencies
upon such a transaction.
Webster's Universal College Dictionary ( 1997) defines "associate" as "to
connect, bring into relation in thought, feeling, memory, etc." Figure 1 of
Kathuria describes if the second transaction (T2 Durable Read Transaction)
"finds that [page] Pis marked with the DIRTIED_BY_LC_XACT" it
"triggers a log flush up to LazyCommitLSN." (Kathuria p. 1207). The
second transaction will only read the page after this occurs, in order to
ensure that any lazily committed transactions have all their effects
committed durably. (Kathuria, p. 1207). Based on the described
relationship between the second transaction, DIRTIED_BY_LC_XACT, and
5
Appeal2014-009363
Application 13/562,906
the LazyCommitLSN counter, as well as the plain and ordinary meaning of
the term "associate," we are not persuaded DIRTIED_BY_LC_XACT and
LazyCommitLSN counter (the tag) are not associated with the second
transaction. See In re Bond, 910 F.2d 831, 832-33 (Fed. Cir. 1990)
(interpretation of references "is not an 'ipsissimis verbis' test").
Accordingly, we are not persuaded the Examiner erred in finding
Kathuria discloses the disputed limitation. We, therefore, sustain the
Examiner's rejection of independent claim 1, as well as independent claims
11 and 15, for which Appellants present substantially the same arguments,
(App. Br. 14--15), and which recite substantially similar limitations. For the
same reasons, we sustain the Examiner's rejection of dependent claims 2--4,
6, 9, 10, 12-14, 17, and 19-21, which were not argued separately.
Claim 5
Issue 3: Did the Examiner err in finding Kathuria discloses "in
response to delaying the transaction commit, the processor wakes a log
flusher to increment a flushed tail position of the recovery log," as recited in
dependent claim 5?
Appellants argue "Kathuria merely discloses that 'the
DurableReadLSN must be updated after the log flush,' without providing a
specific time or nature for the update." (App. Br. 11 ). Appellants argue
"Kathuria is silent regarding waking up a log flusher to increment 'the
DurableReadLSN,' much less waking up the log flusher in response to
delaying the transaction commit." (App. Br. 11 ). Appellants also argue
Kathuria does not disclose that the read transaction is delayed. (Reply Br.
9).
6
Appeal2014-009363
Application 13/562,906
We are not persuaded by Appellants' arguments and agree with the
Examiner's findings. (Final Act. 4; Ans. 11 ). Kathuria describes a counter
called DurableReadLSN, where any page with PageLSN less than or equal
to DurableReadLSN is guaranteed to have all of its changes durably
committed. (Kathuria, §§ 5.3, 5.4, p. 1208-1209). Therefore, to ensure it is
reading durable data, when a durable read transaction (such as T2 Durable
Read Transaction as discussed above) visits a page where
DIRTIED_BY_LC_XACT is set, it flushes the log ifthe pageLSN is greater
than the DurableReadLSN and then sets the DurableReadLSN to the
PageLSN. (Id.). In other words, if the DurableReadLSN is less than
PageLSN, the durable read transaction flushes the log and then sets the
DurableReadLSN to the PageLSN, thus incrementing the DurableReadLSN
to the PageLSN. (Id.). This will only happen in the scenario where the
contents of the page being read are not durable, as described above. (See
id.).
Accordingly, we are not persuaded the Examiner erred in finding
Kathuria discloses "in response to delaying the transaction commit" (as
described in claim 1 ), "the processor wakes a log flusher" (forcing the flush
if the DurableReadLSN is less than PageLSN) "to increment a flushed tail
position of the recovery log" (setting DurableReadLSN to PageLSN). We,
therefore, sustain the Examiner's rejection of claim 5, and for the same
reasons, sustain the Examiner's rejection of claim 16, for which Appellants
present the same arguments (App. Br. 15-16) and which recites substantially
similar limitations.
7
Appeal2014-009363
Application 13/562,906
Claim 8
Issue 4: Did the Examiner err in finding Kathuria discloses "wherein
the processor delays the transaction commit based on performing an integer
comparison between a first LSN corresponding to the at least one tag and a
second LSN corresponding to a flushed tail of the recovery log upon
acquiring or releasing the at least one lock," as recited in dependent claim 8?
Appellants argue Kathuria does not perform an integer comparison
upon acquiring or releasing at least one lock. (App. Br. 13).
We are not persuaded by Appellants' arguments and agree with the
Examiner's findings. (Final Act. 5; Ans. 12). Specifically, the Examiner
finds comparing PageLSN (a first LSN) with DurableReadLSN (a second
LSN), where both are integers, discloses the disputed limitation. (Ans. 12).
Appellants argue such comparison is performed during the read transaction,
so "Kathuria cannot disclose basing a delay on the transaction on such LSN
comparisons." (Reply Br. 10). Appellants' argument is not commensurate
with the scope of the claim, which recites the "transaction commit" is
delayed, not the "transaction." Further, as described in Section 5.1 and in
the Durable Read Transaction pseudo code in Section 5.4 of Kathuria, such
comparison is performed after transaction T 1 releases the lock.
We, therefore, sustain the Examiner's rejection of claim 8, and for the
same reasons, sustain the Examiner's rejection of claim 18, for which
Appellants present the same arguments (App. Br. 16-17) and which recites
substantially similar limitations.
8
Appeal2014-009363
Application 13/562,906
DECISION
For the above reasons, the Examiner's rejection of claims 1-6 and 8-
21 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)(l )(iv).
AFFIRMED
9