Ex Parte Kim et al

Patent Trial and Appeal Board

Jun 24, 2014

12195198 (P.T.A.B. Jun. 24, 2014)

UNITED STATES PATENT AND TRADEMARK OFFICE
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
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
12/195,198 08/20/2008 John Kim 42P46096 8039
45209 7590 06/25/2014
Mission/BSTZ
BLAKELY SOKOLOFF TAYLOR & ZAFMAN
1279 Oakmead Parkway
Sunnyvale, CA 94085-4040
EXAMINER
NAJJAR, SALEH
ART UNIT PAPER NUMBER
2492
MAIL DATE DELIVERY MODE
06/25/2014 PAPER
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.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________________

Ex parte JOHN KIM, DENNIS C. ABTS,
STEVEN L. SCOTT, and WILLIAM J. DALLY
____________________

Appeal 2012-003568
Application 12/195,198
Technology Center 2400
____________________

Before: EDWARD A. BROWN, JAMES P. CALVE, and
THOMAS F. SMEGAL, Administrative Patent Judges.

CALVE, Administrative Patent Judge.

DECISION ON APPEAL

STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134 from the final rejection of
claims 1–20. App. Br. 4. We have jurisdiction under 35 U.S.C. § 6(b).
We AFFIRM.
Appeal 2012-003568
Application 12/195,198

2
CLAIMED SUBJECT MATTER
Claims 1, 10, 19, and 20 are independent. Claim 1 is reproduced
below.
1. A multiprocessor computer system comprising a dragonfly
processor interconnect network, the dragonfly processor interconnect
network comprising:
a plurality of processor nodes;
a plurality of routers, each router directly coupled to a plurality
of terminal processor nodes, the routers coupled to one another and
arranged into a group,
and a plurality of groups of routers, such that each group is
connected to each other group via at least one direct connection.

REJECTIONS1
Claims 1–9 and 11 are rejected under 35 U.S.C. § 112, second
paragraph, as being indefinite for failing to particularly point out and
distinctly claim the subject matter which Applicants regard as the invention.2
Claims 1–20 are rejected under 35 U.S.C. § 103(a) as unpatentable
over McCanne (US 2003/0088696 A1; pub. May 8, 2003).

1 The Examiner also rejected claims 2–9 and 11–18 under 35 U.S.C.
§ 103(a) as unpatentable over McCanne and Scott (Scott, “The Black Widow
High-Radix Clos Network, Proc. of the 33rd Int’l Symposium on Computer
Architecture (ISCA’06), (2006)). Ans. 5–6. Appellants do not indicate that
this rejection is being appealed. App. Br. 8, 9–12. Accordingly, we
summarily affirm this rejection.
2 Although the Examiner rejected claims 1-11 under 35 U.S.C. § 112, second
paragraph, the Examiner did not indicate any basis for rejecting independent
claim 10. Ans. 4.
Appeal 2012-003568
Application 12/195,198

3
ANALYSIS
Claims 1–9 and 11 under 35 U.S.C. § 112, second paragraph, as indefinite
Claims 1-9
The Examiner found claim 1 to be indefinite because it recites a
plurality of “processor nodes” and also recites “terminal processor nodes”
and it is unclear what distinction exists between these two limitations. Ans.
4. Appellants do not present any argument for this rejection. App. Br. 11.
We summarily sustain the rejection of claim 1 and its dependent claims 2-9.
Claims 2 and 11
Claims 2 and 11 depend from claims 1 and 10 respectively and recite
“wherein each group acts as virtual router with radix of approximately 2
times the square root of the number of nodes in the network.” The Examiner
found that the term “approximately” was indefinite because “it [was] unclear
how approximate the radix should be from the 2 times the square root of the
number of nodes in the network.” Ans. 4. We agree with Appellants that
“approximate” would be understood by one skilled in the art to mean
“about” in view of Appellants’ disclosure of this arrangement. App. Br. 11;
see Spec. 6, ll. 1–12. We do not sustain the rejection of claims 2 and 11 on
this ground.
Thus, we sustain the rejection under 35 U.S.C. § 112, second
paragraph, as to claims 1-9, but not as to claim 11.
Claims 1–20 as unpatentable over McCanne
Claims 1–18
Appellants argue claims 1–18 as a group. App. Br. 9–11. We select
claim 1 as representative. 37 C.F.R. § 41.37(c)(1)(vii) (2011). Claims 2–18
stand or fall with claim 1.
Appeal 2012-003568
Application 12/195,198

4
The Examiner found that McCanne discloses a multiprocessor
computer system comprising an interconnect network as recited in claim 1,
including a plurality of processor nodes (para. 62) and a plurality of routers
where each router is coupled directly to a plurality of terminal processor
nodes (fig. 2), the routers coupled to one another and arranged into a group,
and a plurality of groups of routers (para. 50) where each group is connected
to each other group via at least one direct connection (para. 83). Ans. 4–5
(citing McCanne,¶¶ 62, 50, 83). The Examiner found that the routers are
coupled directly to a plurality of terminal processor nodes (clients) and the
routers are “grouped together” in Transit Domains that are connected by at
least one direct connection. Ans. 6–7(citing McCanne, ¶43; fig. 2).
Appellants argue that “McCanne fails to consider routers coupled to
one another and arranged into a group, each router directly coupled to a
plurality of terminal processor nodes.” App. Br. 9, 10.3 Appellants assert
that the dragonfly network includes routers coupled to multiple processor
nodes and one another to form multiple groups. Id. at 10. Appellants further
argue that paragraph 50 of McCanne discloses an “overlay protocol router
used for multicast routing, but fails to discuss any of the router limitations .
. . such as (1) the routers being directly coupled to a plurality of terminal
processor nodes, (2) the routers being coupled to one another arranged into
a group, or (3) a plurality of groups of such routers, where each group is
connected to each other group via at least one direct connection.” Id. at 9.

3 Appellants provide these arguments generally without subheadings for a
particular independent claim. App. Br. 9–11. Therefore, we treat these
arguments as pertaining to independent claims 1 and 10, which contain
these limitations, and to their dependent claims, but as not applicable to
claims 19 and 20, which do not contain such limitations.
Appeal 2012-003568
Application 12/195,198

5
These arguments are not persuasive because they do not address the
Examiner’s findings that McCanne discloses these features in Figure 2 and
paragraphs 43 and 83. Ans. 5, 6–7. The Examiner relies on Figure 2 of
McCanne to show a plurality of processor nodes (clients) that are connected
to overlay routers in a transit domain and the overlay routers are coupled to
one another in transit domains and arranged in groups that are connected by
at least one direct connection. Ans. 4–5, 6–7. Figure 2 also illustrates
overlay routers (OR) arranged in leaf and transit domains to control traffic
through a leaf VIF (LVIF). McCanne, ¶ 62. McCanne discloses that “an
overlay network model is a collection of isolated (but possibly overlapping)
regions of native multicast connectivity” with overlay routers deployed
across the multicast clouds to peer with each other via unicast or multicast
and form a network of application or unicast forwarding agents. Id., ¶ 43.
Once a packet is delivered to an overlay router via multicast or direct unicast
communication, “it is propagated to all other LVIFs and unicast receivers in
the over lay network that include members interested in receiving traffic sent
to that overlay group.” Id., ¶ 83.
Appellants also argue that the claims recite a processor interconnect
network called dragonfly which links processor nodes to one another in a
processor interconnect network and the Final Office Action cites only a
computer terminal which is not a multiprocessor computer. App. Br. 9–10.
This argument is not persuasive of error in the Examiner’s finding that
McCanne discloses the functionality of a dragonfly processor network by
teaching a processor architecture with connections, routers, and processors,
as claimed. Ans. 5, 7. Claim 1 recites a dragonfly processor interconnect
network that comprises a plurality of processor nodes, a plurality of routers,
Appeal 2012-003568
Application 12/195,198

6
and plurality of groups of routers. The Examiner found all of these claimed
features in McCanne. Appellants have not identified a definition or other
disclosure in the Specification that distinguishes the claimed dragonfly
processor interconnect network over McCanne. Appellants disclose that the
Dragonfly topology is a hierarchical network with three levels as shown in
Figure 1 of terminal processor nodes “tcx”, routers 104, 105, 106, and groups
101, 102, 103. Spec. 8, ll. 22–23.
Figure 2 of McCanne is reproduced below along with Appellants’
Figure 1 below it:

Figure 2 of McCanne illustrates an overlay network

Appellants’ Figure 1 illustrates a dragonfly network topology.
Appeal 2012-003568
Application 12/195,198

7
We agree with the Examiner that McCanne discloses the claimed
dragonfly network topology comprising the claimed three levels of
(1) processor nodes (clients), (2) a plurality of routers coupled to one another
and arranged into groups (overlay routers (OR) arranged in transit domains),
and (3) plurality of groups of routers connected to other groups (ORs of
transit domains connected to ORs or other transit domains) as shown in
Figure 2 of McCanne. Appellants’ arguments do not persuade us of error in
those findings of the Examiner.
We sustain the rejection of claims 1–18.
Claims 19 and 20
Claim 19 recites “A multiprocessor computer system, comprising a
Dragonfly processor interconnect network.” Claim 20 recites “A method of
communicating data between processing nodes in a multiprocessor computer
system, comprising routing the data over a Dragonfly processor interconnect
network.” Appellants do not present separate arguments for independent
claims 19 and 20 under separate heading. See App. Br. 9–11. Appellants’
arguments regarding the dragonfly processor interconnect network, which is
recited in claims 19 and 20, are not persuasive for the reasons discussed
supra for claim 1. We sustain the rejection of claims 19 and 20.

DECISION
We AFFIRM the rejection of claim 1-9 and REVERSE the rejection
of claim 11 under 35 U.S.C. § 112, second paragraph.
We AFFIRM the prior art rejections of claims 1–20 under 35 U.S.C.
§ 103(a).
Appeal 2012-003568
Application 12/195,198

8
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
llw