Ex Parte Rana et al

Board of Patent Appeals and Interferences

Dec 23, 2008

09792267 (B.P.A.I. Dec. 23, 2008)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte ASWINKUMAR VISHANJI RANA
and COREY ALAN GARROW
____________

Appeal 2009-0158
Application 09/792,267
Technology Center 2600
____________

Decided: December 23, 2008
____________

Before MAHSHID D. SAADAT, ROBERT E. NAPPI,
and KARL D. EASTHOM, Administrative Patent Judges.

EASTHOM, Administrative Patent Judge.

DECISION ON APPEAL

Appeal 2009-0158
Application 09/792,267

2
STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134 from the Examiner’s Final
Rejection of claims 1-6 and 9-14. (App. Br. 1).1 We have jurisdiction under
35 U.S.C. § 6(b).
We affirm.
Appellants claim a method for reordering out of sequence or
fragmented data packets or datagrams (PDUs) using a link list. Data packets
can become out of order if some of the packets in the sequence are routed
differently. Data packets can also become fragmented when routers or
switches limit the maximum PDU size. The method extracts header
information in each PDU to determine if the data packet is fragmented or out
of order. A link list memory maintains certain header data, including inter
alia, a field holding the head pointer, or address of the first block in a PDU,
the length of the last bock of the PDU, and a next field for storing the pointer
to the next associated block or PDU. (Spec. 2; 5: 2-6; 10:11-15; Abstract).
Claim 1, illustrative of the invention, follows:
1. A method for reordering out of sequence data packets which are
associated with a traffic flow on a data network, the data packets including
header information with fields, some data packets including a sequence
number that determines packet order, the method comprising:
receiving the data packets from an input;
storing the data packets in a packet memory wherein the packet

1 The Examiner’s Answer (mailed June 20, 2008) (“Ans.”) and Appellants’
Brief (filed, December 31, 2007) (“App. Br.”) and Reply Brief (filed June 5,
2008) (“Reply Br.”) detail the respective positions of the parties.
Appeal 2009-0158
Application 09/792,267

3
memory is controlled by a link list controller that keeps track of data packets
in the packet memory;
extracting desired fields from the data packets including respective
sequence numbers; identifying out of sequence data packets;
reordering out of sequence data packets using a link list storing
locations of data packets in a traffic flow; and
sending the data packets to an output.
The Examiner relies on the following prior art references to show
unpatentability:
Liao US 6,185,208 B1 Feb. 6, 2001
Chapman US 6,246,684 B1 Jun. 12, 2001
Bellaton US 6,473,425 B1 Oct. 29, 2002
Almulhem US 6,587,431 B1 Jul. 1, 2003

The Examiner rejected:
Claims 1-3, 5, and 6 under 35 U.S.C. § 103(a) based upon the
teachings of Chapman and Bellaton;
Claim 4 under 35 U.S.C. § 103(a) based upon the teachings of
Chapman, Bellaton, and Almulhem;
Claims 9-11, 13 and 14 under 35 U.S.C. § 103(a) based upon the
teachings of Liao and Bellaton; and
Claim 12 under 35 U.S.C. § 103(a) based upon the teachings of Liao,
Bellaton, and Almulhem.

Appeal 2009-0158
Application 09/792,267

4
ISSUES
Appellants’ and the Examiner’s arguments and rejections raise the
following claim groupings: a) 1-3, 5 and 6; b) 4; c) 9, 13 and 14; d) 10; e)
11; and f) 12. The issues before us are:
1) Did Appellants demonstrate that the Examiner erred in finding that
Chapman and Bellaton collectively teach or suggest reordering out of
sequence data packets using a link list storing locations of data packets in a
traffic flow as set forth in claim 1?2
2) Did Appellants demonstrate that the Examiner erred in finding that
Chapman, Bellaton and Almulhem collectively teach or suggest “wherein
the session id associates the data packets with a particular traffic flow” as set
forth in claim 4?
3) Did Appellants demonstrate that the Examiner erred in finding that
Liao and Bellaton collectively teach or suggest “extracting desired fields
from the fragments respectively including the start offset and the expected
offset” and “reassembling the fragments into a complete data packet using a
link list storing locations of the fragments making up the data packet” as set

2 Appellants initially stated that claims 1-3, 5 and 6 stand or fall together
(App. Br. 4), but then raised new arguments with respect to claims 2, 3, 5
and 6 (Reply Br. 9-18). Appellants waived these new arguments as not
timely raised. Claims 1-3, 5 and 6 remain grouped according to Appellants’
initial grouping. Optivus Tech., Inc. v. Ion Beam Appl’ns S.A. 469 F.3d 978,
989 (Fed Cir. 2006)(“[A]n issue not raised by an appellant in its opening
brief . . . is waived.”)(citations and quotation marks omitted); Ex parte
Scholl No. 2007-3653, (BPAI Mar. 13, 2008)((Informative), at 17-18,
available at http://www.uspto.gov/web/offices/dcom/bpai/its/fd073653.pdf.

Appeal 2009-0158
Application 09/792,267

5
forth in claim 9; “using a window data structure to store the start offset and
expected offset of respective out of order fragments,” as set forth in claim
10; and “comparing the start offset of the respective fragments to an order
offset and to the start offset and expected offset of any window entries,” as
set forth in claim 11?3
4) Did Appellants demonstrate that the Examiner erred in finding that
Liao, Bellaton, and Almulhem collectively teach or suggest that
“reassembling includes [a] time-out mechanism for identifying
unreassembled fragments after a prescribed amount of time,” as set forth in
claim 12?

FINDINGS OF FACT (FF)
1. Chapman discloses completely re-ordering data traffic units or
simply reducing the level of miss-ordering to change the units to their
original order (col. 2, ll. 64-66; col. 7, ll. 13-18). A data traffic unit includes
a part of a packet, a complete packet, a set of packets or an arbitrary set of
data bytes (col. 1, ll. 28-31).
2. Re-orderable queue buffers allow changing the order of data
packets stored in the respective slots of queue buffers and input buffers
temporarily holding the data packets (col. 6, ll. 38-54). Sequence numbers
in the data packet headers are used to re-order the packets in the original
order. (Col. 7, ll. 42-50).

3 Appellants do not separately argue dependent claims 13 and 14, grouped
with claim 9 (App. Br. 15).
Appeal 2009-0158
Application 09/792,267

6
3. One of Chapman’s embodiments, employing a single buffer, re-
orders traffic units according to sequence numbers, but without taking into
account the source of each traffic unit. The first embodiment cannot
guarantee complete or correct packet re-ordering because data from different
sources can be reordered together (col. 7, ll. 51-67). Another embodiment,
employing one buffer per source ID number, re-orders data packets from
each source using sequence numbers and the source ID. The second
embodiment guarantees “a more definitive re-ordering” than the first.
(Chapman, col. 13, ll. 37-54).
4. Bellaton discloses a linked list queue system to control the flow of
data packets over a telecommunications network. The linked list queue
system, storing pointers to packet data, provides flexibility in size alterations
to the queue, and deletions and additions of data, and provides ready
comparison of the data in the packets. Packet sequence numbers and lengths
are compared to determine if a packet should be deleted or added to the
queue. (Col. 5, l. 57 to col. 6, l. 47; Abstract).
5. Bellaton’s system works in conjunction with prior art TCP/IP
protocols (col. 4, ll. 30-35), including a “sliding window” approach (col. 4,
ll. 4-29), that control packet order using, inter alia, a “SEQUENCE NO.” (to
indicate packet position in a message stream), source and destination port
numbers, and data offsets “OFF” (required since data options vary), all
stored in packet headers (col. 3, l. 65 to col. 4, l. 67; col. 10, ll. 46-50; Fig.
5B). Similarly, for fragmented data, a “FLAG” field in such prior art
headers indicates if the data is fragmented. An “IDENT” field uniquely
identifies the message packet containing the fragments. A “FRAGMENT
OFFSET” field specifies the original position of a fragment in a message,
Appeal 2009-0158
Application 09/792,267

7
while “MORE FRAGMENTS” indicates if any more fragments exist.
Other fields, such as data “TOTAL LENGTH” also exist in the header.
(Bellaton, col. 2, ll. 26-61).
6. Bellaton’s duplicate data removal system improves upon such prior
art systems. The system extracts and lists header information into linked
data queues or registers, including inter alia sequence numbers, message
length, data offsets, and pointers linking successive data packets. Packets
are linked in order by “NEXT” and “PREVIOUS” pointers to packet entries.
(Col. 8, l. 43 – col. 9, l. 20; Fig 10). Required data, extracted from incoming
packets, is compared in memory queues to existing queue data. Data
compared includes sequence data, length data, and source data, to determine
if packets are related to the same data flow, where to place the extracted
information in the queue, how to order the pointers linking the packets, and
whether to drop a data packet. (Bellaton, col. 9, l. 21 to col. 10, l. 24).
7. Liao discloses as background prior art, fragmenting message
packets that are too long for a particular wireless network into smaller
packets, and then transmitting the packets in order. However, because such
fragments do not always arrive in order at a destination, the fragments must
be reassembled for transmission and reception using sequence numbers,
which designate fragment order within a message, and reference numbers,
which designate all the fragments for each particular message. The
numbers, inter alia, are transmitted in the fragment headers. (Liao, col. 1, l.
22 to col. 2, l. 3).
8. Liao’s system improves on the prior art systems by using relatively
smaller reference numbers, thereby saving bandwidth and memory. Thus, as
in prior art systems, messages in Liao’s system are fragmented, and
Appeal 2009-0158
Application 09/792,267

8
reassembled in order, using Liao’s improved reference number and prior art
sequence number techniques. That is, servers or other message
transmission/reception units transmit the fragmented messages in correct
order using header information for later correct reassembly by a mobile unit.
(Col. 2, ll. 25-58; col. 3, ll. 22-34; col. 5, ll. 1-15; see also generally col. 5, l.
16 to col. 6, l. 45; see also FF 7).
9. For example, Liao’s system uses a flag in the last fragment header,
and a series number Sn incremented in each successive fragment header in a
data packet, each employed to reassemble the fragments in the correct order
(col. 5, l. 43 to 46, col. 6, ll. 59-63).
10. Almulhem’s system re-orders packets into the proper order using
a sorting algorithm/function (Abstract; col. 7, ll. 15-21). The sorting
function includes a time window period and a STS I signal signifying order.
After the time period expires, the sorting function assumes that a missing
data packet expected to be at a “min” memory address in EPQ is lost. On
the other hand, a previously unassembled data packet found at the “min”
address in EPQ 318 after the timeout is output to EPQ 320 and reassembled.
The sorting function then declares the data stream complete and in the
proper order and frees up the memory for other functions. (Almulhem, col.
9, ll. 22-28; 59-67; col. 11, l. 46 to col. 12, l. 4).
11. Appellant states that sequence numbers and offsets represent
“essentially the same information.” (Spec. 11: 9-10). An “expected offset . .
. is the start offset of the next fragment in the series.” (Spec. 11: 23-27).
The offsets are employed with other information to reassemble the
fragments. (Spec. 11: 15-27).
Appeal 2009-0158
Application 09/792,267

9
12. Appellant’s disclosure indicates that windows constitute areas in
memory (see e.g., Spec. 10:16-32; Fig. 3E, 3G).

PRINCIPLES OF LAW
“[T]here must be some articulated reasoning with some rational
underpinning to support the legal conclusion of obviousness.” In re Kahn,
441 F.3d 977, 988 (Fed. Cir. 2006). “On appeal to the Board, an applicant
can overcome a rejection by showing insufficient evidence of prima facie
obviousness . . . .” Kahn, 441 F.3d at 985-986 (quoting In re Rouffet, 149
F.3d 1350, 1355 (Fed. Cir. 1998).
“[W]hen a patent ‘simply arranges old elements with each performing
the same function it had been known to perform’ and yields no more than
one would expect from such an arrangement, the combination is obvious.”
KSR Int’l Co. v. Teleflex, Inc., 127 S.Ct. 1727, 1740 (U.S. 2007) (quoting
Sakraida v. Ag. Pro. Inc., 96 S.Ct. 1532 (1976)). In “. . . difficult . . . cases .
. . the claimed subject matter may involve more than the . . . mere
application of a known technique to a piece of prior art ready for the
improvement.” Id. “[O]ne cannot show non-obviousness by attacking
references individually where, as here, the rejections are based on
combinations of references.” In re Keller, 642 F.2d 413, 426 (CCPA 1981).
ANALYSIS
Issue 1
Appellants’ argument that neither Bellaton nor Chapman teach
reordering using a linked list (App. Br. 5-9) amounts to a separate attack on
Appeal 2009-0158
Application 09/792,267

10
the references and does not defeat the obviousness rejection under Keller,
supra. As the Examiner generally found, Chapman teaches reordering of
packets in a traffic flow by comparing sequence numbers in buffers, while
Bellaton teaches using a linked list in the same field of endeavor; i.e., the
control of packet data flow in telecommunications systems (FF 1-6, Ans. 4-
6, 10-11). Both systems extract sequence numbers from packet headers and
compare such numbers, in memory buffers or queues, to control the orderly
transmission of packets (FF 1-6). Finally, Bellaton’s system orders
incoming packets successively using pointers to previous and next packets in
a sequence (FF 6). Bellaton’s system also drops duplicate packets (FF 6),
constituting, or, at a minimum, suggesting, proper re-ordering.
Appellants’ quotation and description of Bellaton, alleging a lack of
re-ordering, simply demonstrates how Bellaton’s system controls packet
flow by extracting information from linked packet headers and storing same
in linked lists, or queues. (App. Br. 6-7, quoting Bellaton, col. 9, ll. 12-20;
col. 10, ll.7-32). Using Bellaton’s flexible pointer-linked packet header
sequence extraction technique (FF 4-6) to replace Chapman’s general packet
header sequence extraction technique, would have involved no “more than . .
. the mere application of a known technique to a piece of prior art ready for
the improvement.” KSR, 127 S. Ct. at 1740. (Compare Ans. 11. – Examiner
finding that the combination yields predictable results).
Issue 2
Appellants do not dispute that Almulhem teaches “wherein the session
id associates the data packets with a particular traffic flow” as set forth in
claim 4. Rather, Appellants assert that Chapman teaches away from using
such a session id to associate packets with any particular flow, because
Appeal 2009-0158
Application 09/792,267

11
Chapman is not concerned with multiple sources of traffic flow. (App. Br.
9-11). Appellants’ argument ignores one of Chapman’s two different
embodiments which re-order packets from multiple sources. (FF 3).
Further, Chapman’s embodiment, alleged by Appellants to teach away
from identifying sources, does not do so. Rather, as the passage quoted by
Appellants demonstrate, Chapman teaches that not identifying the source of
a multiple source system has “drawbacks” – thus teaching towards
identifying the source of a multiple source system (see App. Br. 10-11,
quoting Chapman at col. 7, ll. 51-67; FF 3). Moreover, Chapman
specifically teaches “a more definitive re-ordering” by using source ID
numbers to account for such multiple sources (FF 3).
Issue 3
Appellants’ argument that Liao’s system merely fragments a data
packet but does not reassemble it, as claim 9 requires (Reply Br. 25), lacks
factual support. Liao’s system improves on prior art fragmentation
processes and necessarily involves correct reassembly of such fragmented
data packets (see FF 7-9).
Appellants’ argument that Liao does not disclose a start offset and an
expected offset that helps to determine fragment order, fails to explain how
the Examiner’s position is in error. (Reply Br. 22-23). As the Examiner
explained, Liao’s system employs the “current fragment number (start
offset) and the total number of fragments (expected offset).” (Ans. 14). The
Appeal 2009-0158
Application 09/792,267

12
current number reasonably refers to Liao’s sequence/series number (see FF
7-9).4
The sequence number, by itself, or with the total fragment number,
helps Liao’s system determine fragment order, and reasonably corresponds
to the start offset and expected offset of the total number of fragments in a
packet (FF 7-9, see Ans. 7, 14). Thus, Liao discloses storing essentially the
same offset information as Appellants (see FF 11). That is, sequence
numbers, like the claimed start and expected offset information, simply
identify the relative order of the particular fragment within a message. (FF
7-9, 11). Liao’s sequence number necessarily embodies at least three items
of data – the previous, current, and next fragment numbers in a message
sequence. In any case, Bellaton explicitly discloses using prior art fragment
offset data and length data for packet ordering and otherwise. (FF 5).
Moreover, as indicated supra, Appellants acknowledge that sequence
numbers and offsets provide the same information. (FF 11, see also Reply
Br. 27 offsets “give information related to the sequence of the fragments”).
Appellants also state that expected offset merely identifies the next fragment
(FF 11) – much like Liao’s series number identifies the next fragment in the
series. Hence, any distinction between the claimed data and Liao’s data
either amounts to one based upon nonfunctional descriptive material, or
involves the simple substitution of prior art, well-known data elements and
their expected functions. In short, Appellants’ claimed system functions the

4 Liao interchanges “series” and “sequence” when referring to the same
number.
Appeal 2009-0158
Application 09/792,267

13
same as the proposed combination, as the Examiner generally found (Ans.
13-15), even if the claimed system employs different data nomenclatures.
Appellants similarly repeat the limitations of claim 10 and assert that
the Examiner erred with respect to Liao’s teachings (App. Br 16-17, quoting
Liao, col. 9, ll. 36-42). Appellants’ arguments fail, under Kahn, to rebut the
Examiner’s finding that Liao’s system specifically uses stored sequence
numbers and/or total fragment number (i.e., start offset and expected offset)
of fragments in buffers (i.e., windows), and that the combination generally
meets the claim (Ans. 7-9, 16).
For example, Appellants do not explain what constitutes a window
structure, or as indicated supra with respect to claim 9, start offset or
expected offset. Windows, according to Appellants’ disclosure, simply
constitute areas in memory (FF 12). Bellaton’s linked data list queues,
storing data contiguously and including sequence numbers and packet
lengths, constitute the data and windows recited in claim 10 (FF 6).
Similarly, Liao’s system memory or buffers, storing inter alia, the claimed
offsets, i.e., sequences and/or total fragment numbers, also constitute such
windows (see FF 7-9).
Appellants’ assertion that claim 10 requires storing the offsets “only
for respective out of order fragments” (Reply Br. 27) is not commensurate
with the claim scope. The term “only” does not appear in claim 10, and
hence Liao’s system, storing such data for all the fragments, meets the claim.
With respect to claim 11, Appellants similarly fail to demonstrate
error in the Examiner’s position. Appellants argue that the various offsets
are not compared, (see Reply Br. 30-32), but as discussed supra, Liao’s and
Bellaton’s sequence numbers, and/or other data, containing essentially the
Appeal 2009-0158
Application 09/792,267

14
same information as offsets, are compared. Sequence numbers must be
compared to reassemble the fragments (see FF 7-9). Further, Bellaton
discloses comparing sequence and length numbers in queues or windows,
and all manner of prior art fragment offset and other data. (FF 4-6).
Finally, Appellants’ arguments that neither Liao or Bellaton disclose
the limitation of “reassembling the fragments into a complete data packet
using a link list storing locations of the fragments making up the data
packet” as set forth in claim 9, using a window as set forth in claim 10, or by
comparing as set forth in claim 11, (see Reply Br. 23-24, 28, 31-32), amount
to separate attacks on the references, which as noted above, under Keller,
does not demonstrate error in the Examiner’s obviousness determination.
As noted above, Liao discloses reassembling the fragments, by comparing
inter alia, sequence numbers in buffers, and Bellaton discloses controlling
and comparing data packets in windows using a linked list to store fragment
data that provides flexibility. (FF 4-9). Bellaton’s pointer system also
orders the packets successively with pointers. (See FF 6).
Using Bellaton’s well-known linked pointer technique to extract
Liao’s sequence/series numbers from headers and store fragment data (FF 4-
9) would have involved no “more than . . . the mere application of a known
technique to a piece of prior art ready for the improvement.” KSR, 127 S.
Ct. at 1740. In other words, for claims 9-11, comparing any of the known
prior art data types used to control traffic flow and/or order of fragments or
packets to related window data amounts to combining predictable prior art
techniques. See KSR supra.

Appeal 2009-0158
Application 09/792,267

15
Issue 4
Appellants argue that Almulhem, which outputs a data packet after a
time out, does not meet claim 12, which recites “wherein reassembling
includes [sic] time-out mechanism for identifying unreassembled fragments
after a prescribed amount of time.” However, Almulhem’s data packets are
not reassembled at EPQ 318 until after the time-out mechanism identifies
one as a “min” packet, thereby reasonably meeting the claim as the
Examiner generally found. (Ans. 17-18). That is, claim 12, requires
“reassembling” to include “identifying unreassembled fragments.” Hence,
claim 12 reasonably does not preclude reassembling such previously
unreassembled packets, identified by the time-out as not yet reassembled,
and thereafter reassembled.
Appellants explain that their disclosed system discards any packets
still in memory, but not yet reassembled after a time-out mechanism, while
Almulhem’s system outputs the packet after the time-out mechanism (App.
Br. 20-21). That explanation does not distinguish Almulhem’s system. If
discarding fragments still in memory constitutes identifying fragments, then
outputting similar such packet segments still in memory, or designating them
for reassembly, also constitutes identifying. Moreover, Almulhem’s
algorithm similarly assumes that unreassembled packets that should be, but
are not, in the “min” location are lost, and also clears the memory thereafter.
(FF 10). Designating such unassembled packets as lost, in light of
Appellants’ disclosed system, which discards unassembled packets, also
reasonably constitutes “identifying.”
Appellants’ assertion that Liao does not disclose “wherein
reassembling includes time-out mechanism for identifying unreassembled
Appeal 2009-0158
Application 09/792,267

16
fragments after a prescribed amount of time,” again, amounts to a separate
attack on the references, which, under Keller, does not defeat obviousness,
as the Examiner found. (Ans. 17-18). As the Examiner also found, one of
ordinary skill in the art would have been motivated to implement
Almulhem’s sorting method in Liao’s reassembly method “in order to
transmit and receive messages that are larger than the maximum packet size
for a network in the proper order.” (Ans. 17, see FF 7-10).

CONCLUSION
Appellants fail to demonstrate error in the Examiner’s findings.
Accordingly, we sustain the Examiner’s rejections of the claims on appeal.

DECISION
We affirm the Examiner’s decision rejecting claims 1-6 and 9-14.
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)(2006).

AFFIRMED

KIS

HAYNES AND BOONE, LLP
IP Section
2323 Victory Avenue
Suite 700
Dallas, TX 75219