Ex Parte Gaither et al

Patent Trial and Appeal Board

Mar 13, 2015

11831541 (P.T.A.B. Mar. 13, 2015)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________________

Ex parte BLAINE D. GAITHER and RUSS W. HERRELL
____________________

Appeal 2012-012113
Application 11/831,541
Technology Center 2400
____________________

Before: JOSEPH L. DIXON, JAMES R. HUGHES, and ERIC S. FRAHM,
Administrative Patent Judges.

DIXON, Administrative Patent Judge.

DECISION ON APPEAL

Appeal 2012-012113
Application 11/831,541

2
STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134 from a rejection of claims 1–
21. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.

The invention relates to datacenter workload migration using power
cycling awareness (See Spec. ¶¶ 9–11). Claim 1 (App. Br. 24), reproduced
below, is illustrative of the claimed subject matter:
1. A method comprising:

tracking at least one power cycle occurring for a plurality
of computers located within a datacenter and generating power
cycling information as a result of the tracking of the at least one
power cycle;

initiating a workload migration from a first computer of
the plurality of computers to a second computer of the plurality
of computers; and

determining whether to power down the first computer
after completion of the workload migration based at least in part
on the power cycling information.

REFERENCES
The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Nickels
Bradley
Moore
US 2002/0083378 A1
US 2005/0060590 A1
US 2005/0273642 A1
June 27, 2002
Mar. 17, 2005
Dec. 8, 2005
Appeal 2012-012113
Application 11/831,541

3
REJECTIONS
The Examiner made the following rejections:
Claims 1–5, 7–16, and 19–21 stand rejected under 35 U.S.C § 103(a)
as being unpatentable over Bradley and Nickels.
Claims 6, 17, and 18 stand rejected under 35 U.S.C § 103(a) as being
unpatentable over Bradley, Nickels, and Moore.

ANALYSIS
Claims 1–5 and 7–9
Appellants contend that Nickels’ replacing of a faulty sensor device
does not teach and would not have suggested determining whether to power
down a computer after a workload migration based on power cycling
information because Nickels replaces the device after power down has
already occurred (see App. Br. 8–12; Reply Br. 2–3). We are not persuaded
by Appellants’ argument.
The Examiner relies on Bradley for teaching the claim 1 limitation
“determining whether to power down the first computer after completion of
the workload migration” (Ans. 5, 26). Specifically, Bradley discloses
The management entity tracks the utilization of the server
resources and determines the optimal distribution of virtual
machines across the servers to minimize power consumption
while maintaining efficient execution of the workload. Based
on its analysis, the management entity may direct one or more
of the virtual machines to migrate to a different one of the
managed servers while directing other managed servers to
power down.

(Bradley, ¶ 18). Those servers from which data will be migrated “can be
instructed to lower their power consumption, either to a reduced state or to
Appeal 2012-012113
Application 11/831,541

4
an ‘off’ state, resulting in a 50% reduction overall in power consumption”
(Bradley, ¶ 19). The Examiner relies on Nickels for teaching the use of
power cycling information for maintenance decisions, for example, whether
to replace a device in a network (Ans. 5–6). Specifically, Nickels discloses
Together with a host control program stored on the computer
103, [power cycle] information enables diagnostic monitoring
of the power conditions in each device as . . . well as for the
entire network of devices 120n. Problems related to
malfunctioning device connections can be isolated to a specific
device by allowing the computer 103 to obtain and compare the
number of power cycles between each device 120n resulting in
maintenance information. Thus, for example, the power on
time information from a particular device can be used to design
preventative maintenance schedules for that device.

(Nickels, ¶ 27). We agree with the Examiner that, in view of Nickels’
disclosure quoted above, it would have been obvious to use power cycle
information in deciding whether to power down a server in Bradley’s
network after a workload migration (see Ans. 6, 29–30). We are not
persuaded by Appellants’ argument that Nickels powers down a device
before replacing it, contrary to claim 1 (see App. Br. 8–12), because the
Examiner relies on Bradley, not Nickels, for teaching the order of
performing a workload migration before powering down a server (see Ans.
25–30). The Examiner only relies on Nickels for teaching the technique of
monitoring power cycling information and using this information for
maintenance purposes, for example, powering down and replacing a device
(see Ans. 5–6, 27). Appellants do not specifically explain why it would
have been beyond the skill in the art to use Nickels’ power cycle information
as criteria in deciding whether to power down one of Bradley’s servers. See
KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 417 (2007) (“[I]f a technique
Appeal 2012-012113
Application 11/831,541

5
has been used to improve one device, and a person of ordinary skill in the art
would recognize that it would improve similar devices in the same way,
using the technique is obvious unless its actual application is beyond his or
her skill.”).
We are, therefore, not persuaded the Examiner erred in rejecting claim
1, and claims 2–5 and 7–9 not specifically argued separately.

Claim 6
Appellants contend “nothing in the cited section of Moore . . . teaches
or suggests that power cycling information is compared against warranty
information to determine whether to power down a first computer, as recited
in claim 6” (App. Br. 19). Rather, Appellants assert “the cited section of
Moore is related to a method for determining (e.g., by a manufacturer)
parameters of a warranty” (id.). We disagree with Appellants.
Moore discloses:
[A] hardware device is implemented to track and store the run
time and number of power on cycles (or other duration and
event count items for a hardware subassembly. . . . [A]
diagnostic tool can be used to access the hardware device and
retrieve the run time/number of power cycles information. This
information can be useful in measuring reliability parameters
for the subassembly in field use. It may also be possible to use
this information for the determination of pre-failure warranty
and pre-failure warning features for the subassembly.

(Moore, ¶ 54). We disagree with Appellants (App. Br. 19) that this portion
of Moore should be read as disclosing determining warranty parameters by a
manufacturer. Moore discloses a “diagnostic tool” to retrieve power cycle
information and using the information in measuring reliability parameters
Appeal 2012-012113
Application 11/831,541

6
“in field use.” Accordingly, we read Moore as suggesting the use of power
cycle information to determine whether parameters of a warranty are met
during operation of a device. We thus agree with the Examiner (see Ans. 22,
49–50) that it would have been obvious to compare power cycle information
with warranty information in the combination of Bradley and Nickels.
We are, therefore, not persuaded the Examiner erred in rejecting claim
6.

Claims 10–16
Appellants contend that modifying Bradley in view of Nickels would
result in an inoperable device because “the virtual machine migration
techniques disclosed in Bradley are incompatible with techniques for the
repair/replacement of a sensor 120n disclosed in Nickels” (App. Br. 13).
Specifically, Appellants argue:
If the system in Nickels were modified such that the
workload of the first CAN [Controller Area Network]-based
sensor . . . were to be migrated to a second CAN-based sensor
. . . before the first CAN-based sensor is removed from the
system and physically replaced by the second CAN-based
sensor, the second CAN-base[d] sensor would not be installed
in the proper physical location. . . . Accordingly, the system
resulting from the combination and modification of Bradley and
Nickels would be inoperable, as any data received from the
second CAN-based sensor would not report actual operating
conditions.

(App. Br. 14). We disagree with Appellants.
Appellants’ arguments are unpersuasive because they mischaracterize
the Examiner’s proposed combination. The Examiner does not propose
combining Bradley’s virtual machine migration techniques with Nickels’
Appeal 2012-012113
Application 11/831,541

7
sensor network. Rather, the Examiner relies on Bradley for teaching “a
migration manager that causes a workload in a first computer . . . of the
datacenter to be migrated to a second computer . . . of the datacenter” (Ans.
11–12) and relies on Nickels for teaching tracking power cycle information
and using power cycle information for maintenance purposes in a network
(Ans. 13). Thus, the proposed combination applies Nickels’ technique of
tracking and using power cycle information to Bradley’s datacenter as a
basis for migrating a workload from one computer to another (see Ans. 13,
34). This does not require incorporating the structure of Nickels’ sensor
network into Bradley’s system. Accordingly, we are not persuaded that the
combination of Bradley and Nickels would result in an inoperable device.
Further, we are not persuaded by Appellants’ additional argument that
“the Examiner has not explained how Nickel[s’] teaching of determining
whether to replace a device (e.g., the faulty sensor 120n) could lead one of
ordinary skill in the art to implement a migration manager that causes a
workload in a first computer to migrate to a second computer based on
tracking information . . . from a tracking of at least one power cycle” (App.
Br. 15). The Examiner explains that one would have wanted to use power
cycle information to identify possible failures of devices to keep the system
continuously running (Ans. 9). Appellants do not specifically explain why
this reason does not suffice as a motivation to combine Nickels with Bradley
(see App. Br. 15).
We are, therefore, not persuaded the Examiner erred in rejecting claim
10, and claims 11–16 not specifically argued separately.

Appeal 2012-012113
Application 11/831,541

8
Claims 17 and 18
Appellants contend, similarly to the argument presented for claim 6,
that “the cited section of Moore is related to a method for determining (e.g.,
by a manufacturer) parameters of a warranty” (App. Br. 21). Appellants also
contend “nothing in the cited section of Moore . . . teaches or suggests to one
of ordinary skill in the art how to implement a migration manager that
considers warranty information for a first computer in its evaluation of
whether a workload in the first computer should be migrated to a second
computer, as recited in claims 17 and 18” (id.). We disagree with
Appellants.
For the reasons discussed above regarding claim 6, we do not find that
Moore’s disclosure is limited to determining warranty parameters by a
manufacturer. Further, as mentioned above, we find Moore suggests using
power cycle information to determine whether parameters of a warranty are
met during operation of a device (see Moore, ¶ 54). We thus agree with the
Examiner (see Ans. 22–24, 51) that it would have been obvious, in view of
Moore, to consider not only power cycle information, but also warranty
information, in determining to perform a workload migration in the system
of Bradley and Nickels.
We are, therefore, not persuaded the Examiner erred in rejecting
claims 17 and 18.

Claims 19 and 20
Regarding claim 19, Appellants contend: “The Examiner has not
explained how Bradley or Nickels (taken alone or in combination) would
teach or suggest how to implement powering down a target computer as a
Appeal 2012-012113
Application 11/831,541

9
result of a comparison of power cycling information upon migration of a
workload” (App. Br. 18). We disagree with Appellants.
The Examiner finds that Bradley discloses comparing information
such as the utilization of server resources among a number of servers in a
datacenter to determine an optimal distribution of virtual machines across
the servers, performing a workload migration from one server to another,
and determining whether to power down the server (see Ans. 17–18). The
Examiner relies on Nickels for teaching tracking and comparing power cycle
information among a number of devices and using this information to
perform maintenance on the devices, including powering down a device as a
result of the comparison (see Ans. 18, 45). Specifically, Nickels discloses:
“Problems related to malfunctioning device connections can be isolated to a
specific device by allowing the computer 103 to obtain and compare the
number of power cycles between each device 120n resulting in maintenance
information” (Nickels, ¶ 27). The Examiner concludes it would have been
obvious in view of Nickels to compare power cycle information among
Bradley’s servers and use it as the basis for powering down a server upon a
workload migration (see Ans. 47–48). Appellants do not specifically
explain why the Examiner’s proposed combination fails to show how
Nickels’ power cycle information would be implemented in Bradley’s
system (see App. Br. 17–18).
In the Reply Brief, Appellants assert, “The Examiner has not cited any
section of Bradley or Nickels that would teach or suggest to one of ordinary
skill in the art powering down a target computer as a result of a comparison
of power cycling information upon migration of a workload, as recited in
claim 19” (Reply Br. 8). However, as discussed above, the Examiner relies
Appeal 2012-012113
Application 11/831,541

10
on the collective teachings of Bradley and Nickels for disclosing this feature.
Appellants do not specifically explain why one of ordinary skill in the art
would not have been motivated to apply Nickels’ technique of comparing
power cycle information among a number of devices in order to perform
maintenance, including powering down a device, to the servers in Bradley’s
system when powering down a server after a workload migration. To the
contrary, we find that, given Nickels’ teaching that power cycle information
can help diagnose an impending failure of a device, one of ordinary skill in
the art would have compared power cycle information—as taught by
Nickels—among Bradley’s servers, and powered down one of Bradley’s
servers based on this comparison after first performing a workload migration
from the server. The Examiner has stated a reason for so combining the
references, namely, to keep Bradley’s system continuously running (Ans.
19–20, 47).
We are, therefore, not persuaded the Examiner erred in rejecting claim
19, and claim 20 not specifically argued separately.

Claims 21
Appellants contend that Nickels does not disclose storing any
information for a replacement sensor because “presumably the replacement
sensor would be a new sensor that would have no information related to the
number of times that the replacement sensor [h]as been turned on or off”
(App. Br. 16–17). “Thus, Bradley taken in view of Nickels does not teach or
suggest to one of ordinary skill in the art how to implement a migration
manager that considers the service life information for both the first
computer and the second computer in its evaluation of whether the workload
Appeal 2012-012113
Application 11/831,541

11
in the first computer should be migrated to the second computer, as recited
in claim 21” (id. at 17). We disagree with Appellants.
The Examiner relies on Bradley for disclosing a workload migration
from one server to another based on server resource utilization, and relies on
Nickels for teaching that service life information is maintained among
devices in order to monitor the health of the devices (see Ans. 20–21). Thus,
the Examiner does not rely on Nickels’ replacement sensor to meet the
limitation of a “second computer” to which a workload is migrated. Rather,
the Examiner proposes combining Nickels’ service life information with
Bradley’s system such that service life information would be maintained for
Bradley’s servers (see Ans. 21, 42). Accordingly, Appellants’ argument that
a replacement sensor in Nickels’ system would not have any service life
information associated with it (see App. Br. 16–17) is not persuasive. We
agree with the Examiner and find that it would have been obvious to apply
Nickels’ technique of storing and using service life information to Bradley’s
system such that the service life information would be taken into account in
performing a workload migration from one server to another.
We are, therefore, not persuaded the Examiner erred in rejecting claim
21.

CONCLUSION
The Examiner did not err in rejecting claims 1–21 under 35 U.S.C.
§ 103(a).

Appeal 2012-012113
Application 11/831,541

12
DECISION
For the above reasons, the Examiner’s decision rejecting claims 1–21
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)(1)(iv).

AFFIRMED