Ex Parte Rochon et al

Board of Patent Appeals and Interferences

Sep 20, 2010

09963487 (B.P.A.I. Sep. 20, 2010)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________________

Ex parte STEVE ROCHON, RICHARD S. NORMAN,
and ROBIN BOIVIN
____________________

Appeal 2009-006259
Application 09/963,4871
Technology Center 2400
____________________

Before KENNETH W. HAIRSTON, MARC S. HOFF,
and CARLA M. KRIVAK, Administrative Patent Judges.

HOFF, Administrative Patent Judge.

DECISION ON APPEAL2

1 The real party in interest is Hyperchip Inc.
2 The two-month time period for filing an appeal or commencing a civil
action, as recited in 37 C.F.R. § 1.304, or for filing a request for rehearing,
as recited in 37 C.F.R. § 41.52, begins to run from the “MAIL DATE”
(paper delivery mode) or the “NOTIFICATION DATE” (electronic delivery
mode) shown on the PTOL-90A cover letter attached to this decision.

Appeal 2009-006259
Application 09/963,487

STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 1-4, 25-37, 40-51, 53-55, and 57-58. We have jurisdiction under 35
U.S.C. § 6(b).
We reverse.
Appellants’ invention relates to a method for regulating packet flow at
the ingress stage of a packet switching device in an effort to avoid
congestion. Bandwidth utilization information is obtained for each of a
plurality of queues at an egress stage of the device to generate a discard
probability for each queue. The discard probability is transmitted to the
ingress stage, either periodically or at other controlled time periods.
Depending on the queue for which the packet is destined and the discard
probability for that queue, the ingress stage decides to transmit the packet to
the switch fabric. This method “leads to improved bandwidth utilization,
since packets which are discarded at the ingress stage will not unnecessarily
take up other resources in the device” (Spec. 3:15-4:2).
Claim 1 is exemplary:
1. A method of regulating packet flow through a device having a
processing fabric with at least one input port and at least one output port, a
control entity connected to the at least one input port for regulating packet
flow thereto, and a plurality of egress queues connected to the at least one
output port for temporarily storing packets received therefrom, said method
comprising:
obtaining bandwidth utilization information regarding packets
received at the egress queues, wherein obtaining said bandwidth utilization
information includes determining the amount of bandwidth consumed by
packets received at each of said egress queues;
determining, from the bandwidth utilization information and the
amount of bandwidth consumed by packets received at each of said egress
queues, a discard probability associated with each egress queue; and
2
Appeal 2009-006259
Application 09/963,487

providing the discard probability associated with each egress queue to
the control entity, for use by the control entity in selectively transmitting
packets to the at least one input port of the processing fabric.

The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Blumer US 2002/0105908 A1 Aug. 8, 2002
Jefferies US 6,728,253 B1 Apr. 27, 2004
Haskin US 6,813,242 B1 Nov. 2, 2004
Brewer US 7,002,980 B1 Feb. 21, 2006
Lyon CA 2,292,828 Jun. 22, 2001

Claims 1-3, 25-27, 30, 36-37, 40-51, and 57 stand rejected under 35
U.S.C. § 103(a) as being unpatentable over Lyon in view of Brewer.
Claims 4, 28, 29, 31-35, 53, 54, and 55 stand rejected under 35 U.S.C.
§ 103(a) as being unpatentable over Lyon in view of Brewer and Blumer.3
Claim 50 stands rejected under 35 U.S.C. § 103(a) as being
unpatentable over Lyon in view of Brewer and Haskin.
Claim 58 stands rejected under 35 U.S.C. § 103(a) as being
unpatentable over Lyon in view of Brewer and Jefferies.
Rather than repeat the arguments of Appellants or the Examiner, we
make reference to the Appeal Brief (filed August 29, 2007), the Reply Brief
(filed February 28, 2008), and the Examiner’s Answer (mailed February 6,
2008) for their respective details.

3 The Examiner has rejected claim 53 but has not set forth the reason for
rejection of this claim in the Final Office Action nor the Answer; however
the reason is set forth in the Non-Final Action (p. 13; mailed July 01, 2005).
3
Appeal 2009-006259
Application 09/963,487

ISSUE
Appellants contend that neither Lyon nor Brewer discloses
"determining the amount of bandwidth consumed by packets received at
each of said egress queues" and "determining, from the bandwidth utilization
information and the amount of bandwidth consumed by packets received at
each of said egress queues, a discard probability associated with each egress
queue" (App. Br. 8-9).
Appellants’ contentions present us with the following issue:
Do the combined teachings of Lyon and Brewer disclose a control
entity that regulates packet flow through a device by obtaining bandwidth
utilization information regarding packets received at the egress queues to
determine the amount of bandwidth consumed by packets received at each of
the egress queues and generating a discard probability associated with each
egress queue from the bandwidth utilization information and the amount of
bandwidth consumed by the packets?

FINDINGS OF FACT
The following Findings of Fact (FF) are shown by a preponderance of
the evidence.
Brewer
1. Brewer discloses a multi-Quality of Service (QOS) level
queuing structure where packet payload pointers are stored in multiple
queues and packet payloads are stored in a common memory pool.
Instantaneous drop probabilities are derived from algorithms that measure
instantaneous queue size and compare it with calculated minimum and
maximum queue sizes. A traditional Weighted Random Early Discard
4
Appeal 2009-006259
Application 09/963,487

(WRED) mechanism generates time averaged drop probabilities of the
packets. Algorithms are adapted to the multi-level QOS queuing structure,
floating point format, and hardware implementation. An arbitration
algorithm controls the packet flow traffic from a router egress queuing
structure to a single egress port tributary (Abstract).
2. Brewer discloses in Equation 3.1 of the WRED algorithm, used
in the router for management of congestion that an average byte count
avgByteCnt is obtained for a particular queue into which a packet is being
placed, is indexed by both the destination tributary and the QOS level of the
incoming packet. The time weighted average for the actual byte count for a
particular queue is used to determine the avgByteCnt. A probability (prob1)
is calculated by multiplying this avgByteCnt by a constant and then
subtracting a different constant as shown in Equation 3.2. The drop
probability (dropProb) or discard probability is calculated by multiplying the
probability (prob1) by the size of the packet: the smaller the packet the less
likely it is to be dropped (col. 8, ll. 30-60).

PRINCIPLES OF LAW
On the issue of obviousness pursuant to 35 U.S.C § 103, the Supreme
Court has stated that “the obviousness analysis cannot be confined by a
formalistic conception of the words teaching, suggestion, and motivation.”
KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 419 (2007). Further, the Court
stated “[t]he combination of familiar elements according to known methods
is likely to be obvious when it does no more than yield predictable results.”
Id. at 416.
5
Appeal 2009-006259
Application 09/963,487

ANALYSIS
Claims 1-3, 25-27, 30, 36-37, 40-51, and 57
Independent claims 1 and 37 recite "determining the amount of
bandwidth consumed by packets received at each of said egress queues …
determining, from the bandwidth utilization information and the amount of
bandwidth consumed by packets received at each of said egress queues, a
discard probability associated with each egress queue." Independent claims
40 and 46 recite claim limitations similar in scope.
We consider Appellant’s arguments to be persuasive to show
Examiner error (see App. Br. 11-13; Reply Br. 11-12). We do not agree
with the Examiner’s finding that Brewer discloses the claimed limitation by
specifically disclosing that the traffic flow method of Brewer determines the
QOS level (which the Examiner finds to be equivalent to the bandwidth
utilization level) and an average byte count (which the Examiner finds to be
equivalent to a bandwidth consumed by packets at the egress queue) (Ans.
19). In particular, we do not agree with the Examiner’s finding that the QOS
level is equivalent to the bandwidth utilization level (Ans. 19).
Brewer discloses a multi-Quality of Service (QOS) level queuing
structure where instantaneous drop probabilities are derived from algorithms
and the instantaneous queue size is compared with calculated minimum and
maximum queue sizes (FF 1). Specifically, Brewer discloses that the
average byte count is obtained for the associated queue into which a packet
is being placed (FF 2). The time weighted average for the actual byte count
is used to determine the average byte count (FF 2). A probability (prob1) is
calculated by multiplying this average byte count by a constant and then
subtracting a fixed constant (FF 2). The drop probability (dropProb) or
6
Appeal 2009-006259
Application 09/963,487

discard probability is calculated by multiplying this probability (prob1) by
the size of the packet (FF 2). Therefore, the drop probability (dropProb) is
proportional to the probability (prob1) which is proportional to the average
byte count (FF 2). The average byte count, however, is proportional to the
time weighted average for the actual byte count for the particular queue (FF
2). Since the actual byte count is the number of bytes waiting to be
transmitted, rather than the number of bytes actually sent, the byte count is
not indicative of consumption (bytes per unit time), but rather of allocation
(number of bytes). Thus, the probability (prob1) and the drop probability
(dropProb) are not indicative of the consumption of bandwidth. Rather,
Brewer merely discloses a time-weighted average byte count of a queue
(Reply Br. 11) Therefore, we do not agree with the Examiner’s finding that
the QOS is equivalent to the bandwidth utilization level (Ans. 5 and 19).
Rather, we find that the QOS corresponds to authorization and allocation for
a specified amount of bandwidth, but it does not equate to the amount of
bandwidth actually consumed.
We find that the combination of Lyon and Brewer does not disclose
"determining the amount of bandwidth consumed by packets received at
each of said egress queues." As a result, we will not sustain the Examiner’s
rejection of independent claims 1, 37, 40, and 46 under 35 U.S.C. § 103, nor
that of claims 2, 3, 25-27, 30, 36-37, 40-51, and 57 not separately argued
with particularity.
Claims 4, 28, 29, 31-35, 50, 53, 54, 55, and 58
As noted supra, we reversed the rejection of claims 1 and 46 from
which claims 4, 28, 29, 31-35, 53, 54, and 55 depend. We have reviewed
Blumer, Haskin, and Jeffries (the additional references applied by the
7
Appeal 2009-006259
Application 09/963,487

Examiner to reject these claims), and find that none of the references teaches
the limitation deemed to be absent from the combination of Lyon and
Brewer.
We therefore reverse the Examiner’s rejections of claims 4, 28, 29,
31-35, 53, 54, and 55 under 35 U.S.C. § 103, for the same reasons expressed
with respect to the § 103 rejection of parent claims 1 and 46, supra.

CONCLUSION
The references do not disclose a control entity that regulates packet
flow through a device by obtaining bandwidth utilization information
regarding packets received at the egress queues to determine the amount of
bandwidth consumed by packets received at each of the egress queues and
generating a discard probability associated with each egress queue from the
bandwidth utilization information and the amount of bandwidth consumed
by the packets.

ORDER
The Examiner’s rejection of claims 1-4, 25-37, 40-51, 53-55 and 57-
58 is reversed.

8
Appeal 2009-006259
Application 09/963,487

REVERSED

ELD

SMART & BIGGAR
1000 DE LA GAUCHETIERE ST. W.
SUITE 3300
MONTREAL QC H3B 4W5 CA CANADA

9