Ex Parte Sharma

Board of Patent Appeals and Interferences

Mar 26, 2003

09042202 (B.P.A.I. Mar. 26, 2003)

-1–
The opinion in support of the decision being entered
today was not written for publication in a law journal
and is not binding precedent of the Board.
Paper No. 18
UNITED STATES PATENT AND TRADEMARK OFFICE

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES

Ex parte DEBENDRA DAS SHARMA

Appeal No. 2002-1498
Application No. 09/042,202

HEARD: MARCH 13, 2003

Before KRASS, JERRY SMITH and BARRETT, Administrative Patent
Judges.
KRASS, Administrative Patent Judge.
DECISION ON APPEAL
This is a decision on appeal from the final rejection of
claims 1-32.
The invention is directed to a system for managing
communications between nodes utilizing credit flow control. A
“credit flow control” is used in order to avoid losing data when
a queue associated with a destination node is completely full.
Appeal No. 2002-1498
Application No. 09/042,202
-2–
In a reactive flow control manner, the destination node tracks
its own queues and sends a queue full signal to a sending node
when the queue is full. In a proactive flow control manner, the
sending node keeps track of the amount of queue space available
at the destination queue in the form of “credits” stored in a
credit register. Each credit corresponds to a number of entries
available at the destination queue. The sending node will then
only send data to the destination node when the sending node has
sufficient credit. When the destination node clears space in the
queue, the destination queue releases credits to the sender,
allowing the sender to send more data, as needed. The
destination queue stores the number of credits to be released in
the debit register.
The problem lies in the initialization of credits in
respective registers. Since queue sizes may change over time,
with design revisions, the sender nodes cannot assume queue sizes
in a destination node. Therefore, upon start-up, the sender will
not know how many credits to allocate in the registers for each
queue in the destination node. While the prior art initially set
the debit and credit registers to zero and then software was used
to read the maximum number of credits for each destination node
and the number was written into the sender node, the problem
Appeal No. 2002-1498
Application No. 09/042,202
-3–
with this approach is alleged to be in the paradox arising from
the need for transaction credits to be available in order to
perform this credit initialization but the credit initialization
cannot take place until credits are available. The solution of
the prior art was to use a “power on mode” which allowed all
transactions to be sent without using credits but this is alleged
to require extra logic and certain minimum queue sizing so that
the user needed to have extensive knowledge of the system.
The instant invention is said to overcome these
disadvantages by employing the ordinary flow transfer processes
of the system. That is, if the sender has credits, then it may
send transactions to the destination node. If the destination
node has debits, then it releases them back to the sender node.
During initialization, the credit register is loaded with a zero
so that the sender node begins operation with no credits and no
transactions can be sent. The debit register is not loaded with
a zero but, rather, with the maximum credits representing the
size of its queue. Accordingly, the destination register begins
operations with the debit register full and, through normal
operations, releases the credits back to the sender node. Then
the sender node will have credits and can send transactions and
the destination node will have an empty debit register. This is
Appeal No. 2002-1498
Application No. 09/042,202
1It appears that the examiner intended to identify line 5 of
claim 14.
-4–
alleged to result in faster boot times since initialization is
completed as soon as the chips are out of their reset state
because no time is spent on software operations in a start-up
mode.
Representative independent claim 1 is reproduced as follows:
1. A method for initializing credit in a credit register
used for flow control that is resident on a first node of a
multinode computer system, the method comprising the steps of:
loading an initial credit value into a debit register
resident on a second node of the multinode computer system; and
transferring the initial credit value into the credit
register by using operational mechanisms of the system.
The examiner relies on the following reference:
Barkey et al. [Barkey] 6,044,406 Mar. 28, 2000
(filed Apr. 8, 1997)
Claims 1-32 stand rejected under 35 U.S.C. 102(e) as
anticipated by Barkey.
Claim 14 stands rejected under 35 U.S.C. 112, second
paragraph, as reciting “the step of transferring” on line 61 for
which there is insufficient antecedent basis.
Appeal No. 2002-1498
Application No. 09/042,202
-5–
Reference is made to the briefs and answer for the
respective positions of appellant and the examiner.
OPINION
With regard to the rejection under 35 U.S.C. 112, second
paragraph, under this statutory section, the inquiry is whether a
claim does, in fact, set out and circumscribe a particular area
with a reasonable degree of precision and particularity. It is
here where the definiteness of the language employed must be
analyzed–not in a vacuum, but always in light of the teachings of
the prior art and of the particular application disclosure as it
would be interpreted by one possessing the ordinary skill in the
pertinent art. In re Moore, 439 F2d 1232, 1235, 169 USPQ 236,
238 (CCPA 1971).
While the examiner protests that there is no antecedent
basis for the phrase, “the step of transferring” on line 5 of
claim 14, and we do agree that the claim structure is a bit
awkward in this regard, we will not sustain this rejection
because, after a careful reading of the claim and the
specification, it is clear that “the step of transferring” refers
to the subsequent paragraph, beginning “transferring the initial
Appeal No. 2002-1498
Application No. 09/042,202
-6–
credit value....” Accordingly, while it may be a bit awkward to
have the antecedent basis for a recited limitation appear
subsequent to, rather than previous to, the recited limitation,
the instant claim language does, “in fact, set out and
circumscribe a particular area with a reasonable degree of
precision and particularity.” Accordingly, since the artisan
would have clearly understood the nature of the invention from
the instant claim language and would clearly have understood as
to which limitation “the step of transferring” referred, we will
not sustain the rejection of claim 14 under 35 U.S.C. 112, second
paragraph.
Under 35 U.S.C. 102, a reference must disclose, explicitly
or implicitly, every limitation of the claimed invention. Glaxo
Inc. v. Novopharm Ltd., 52 F.3d 1043, 1047, 34 USPQ2d 1565, 1567
(Fed. Cir.), cert. denied, 516 U.S. 988 (1995).
Barkey clearly discloses a credit-based flow control system
wherein an initial credit value is loaded into a debit register
resident on a second node. In Figure 1, element 34, equivalent
to the claimed “debit register,” within receiver 14, i.e., second
node, is loaded with an initial credit value of n (column 4,
lines 39-41). The issue, as we view it, is whether Barkey
discloses “transferring the initial credit value into the credit
Appeal No. 2002-1498
Application No. 09/042,202
-7–
register by using operational mechanisms of the system,” as
claimed. In our view, the term, “using operational mechanisms of
the system” is broad enough to cover any way Barkey’s system
transfers the initial credit value into the credit register, if,
in fact, Barkey performs such a transfer at all.
The register 30 within sender 12 in Barkey would be the
“credit register,” as referred to in the claims. From the
disclosure at column 4, lines 26-65, of Barkey, it appears that
at the time of initialization, an initial credit value of n is
loaded into the credit register of sender 12 and the same value n
is loaded into the debit register 34 of the second node, or
receiver, 14.
Unlike appellant’s invention, Barkey does not appear to
disclose that the credit value n is initially loaded into the
debit register and then the contents of the debit register are
transferred to the credit register in sender 12. Rather, Barkey
appears to load the same credit value n into both the credit and
debit registers. Having said that, however, and with a full
understanding as to how appellant’s disclosed invention differs
from that of Barker, we now focus on the instant invention, as
claimed.
With regard to claim 1, there is no requirement that the
Appeal No. 2002-1498
Application No. 09/042,202
-8–
contents of the debit register be transferred into the credit
register. There is nothing even requiring that anything in the
debit register be transferred to the credit register. The claim
merely requires that an initial credit value be loaded into a
debit register on a second node (Barker clearly does this, at
column 4, lines 39-41, wherein an initial credit value of n is
loaded into debit register 34 on second node, or receiver, 14)
and “transferring the initial credit value into the credit
register....” This latter portion of the claim does not recite
transferring the contents of the debit register. In fact, it
does not require any connection with the debit register. It
merely requires the transferring of the initial credit value
(which, in Barker’s case, is “n”) into the credit register.
Since the initial credit value of n is “transferred” into the
credit register 30 of sender 12, no matter where from, the terms
of the claim are met.
Accordingly, we will sustain the rejection of claim 1 under
35 U.S.C. 102(e) and also the rejection of claim 4 since the
merits of this claim are not separately argued.
Similarly, since independent claim 17 is merely the
apparatus counterpart to method claim 1 and includes similar
language, we will also sustain the rejection of claim 17, and of
Appeal No. 2002-1498
Application No. 09/042,202
-9–
claim 20, dependent therefrom, and not separately argued, under
35 U.S.C. 102(e).
With regard to independent claims 14 and 30, we take a
different view. These claims include an intermediate step of, or
second means for, loading the credit register with a zero prior
to the step of transferring the initial credit value into the
credit register. Just prior to initialization in Barker, it is
not clear what, exactly, the value is in the credit register but
it is clear that, at initialization, the credit value of n is
loaded into both the debit register and the credit register.
Thus, we cannot say, with assurance, that Barker, at any time,
actually loads the credit register with a zero, let alone at some
time prior to the step of transferring, as required by claims 14
and 30. For similar reasons, we will not sustain the rejection
of claims 3 and 19 under 35 U.S.C. 102(e) since these claims also
contain the limitation of loading the credit register with a zero
prior to the step of transferring.
With regard to claims 5 and 21, these claims recite further
specifics regarding the transferring step, viz., that the initial
credit value is placed into a packet which is bound for the first
node, that the debit register is decremented by the initial
credit value and the packet is sent to the first node. We find
Appeal No. 2002-1498
Application No. 09/042,202
-10–
no such specifics, or even the suggestion of placing the initial
credit value in a packet, in Barker. Accordingly, we will not
sustain the rejection of claims 5 and 21, or of claims 6-13 and
22-29, dependent thereon, under 35 U.S.C. 102(e).
Since Barker does not appear to disclose each and every
limitation of claims 5-16 and 21-32, we will not sustain the
rejection of these claims under 35 U.S.C. 102(e).
With regard to dependent claims 2 and 18, these claims
recite that the initial credit value “is equal to a size of a
queue that is resident on the second node and receives
information from the first node.” While appellant does not
specifically argue these limitations, it appears that Barker
discloses this at column 4, lines 57-62, wherein it is explained
how and why a certain initial value n is chosen. Thus, we will
sustain the rejection of claims 2 and 18 under 5 U.S.C. 102(e).
We have not sustained the rejection of claim 14 under
35 U.S.C. 112, second paragraph, nor have we sustained the
rejection of claims 3, 5-16, 19 and 21-32 under 35 U.S.C.
102(e). We have, however, sustained the rejection of claims 1,
2, 4, 17, 18 and 20 under 35 U.S.C. 102(e).
Accordingly, the examiner’s decision is affirmed-in-part.
Appeal No. 2002-1498
Application No. 09/042,202
-11–
No time period for taking any subsequent action in
connection with this appeal may be extended under 37 CFR
§ 1.136(a).
AFFIRMED-IN-PART
ERROL A. KRASS )
Administrative Patent Judge )
)
)
)
)
JERRY SMITH )BOARD OF PATENT
Administrative Patent Judge ) APPEALS AND
)INTERFERENCES
)
)
)
LEE E. BARRETT )
Administrative Patent Judge )

Appeal No. 2002-1498
Application No. 09/042,202
-12–
EK/RWK
HEWLETT PACKARD COMPANY
P O BOX 272400, 3404 E. HARMONY ROAD
INTELLECTUAL PROPERTY ADMINISTRATION
FORT COLLINS, CO 80527-2400