Microlinc, LLC v. Intel Corporatioh et al
Filing
529
MEMORANDUM OPINION AND ORDER. Signed by Judge Michael H. Schneider on 6/7/13. (bas, )
United States District Court
EASTERN DISTRICT OF TEXAS
MARSHALL DIVISION
MICROLINC, LLC
§
§
§
§
§
v.
INTEL CORP., et al.
Case No. 2:07-cv-488
MEMORANDUM OPINION AND ORDER
This claim construction order construes the disputed claim terms of U.S. Patent No.
6,009,488 (’488 Patent) entitled “Computer Having Packet-Based Interconnect Channel.” On
May 16, 2013, the Court held a claim construction hearing to construe the disputed terms.
For the following reasons, the Court adopts the constructions set forth below. See also Appendix
A.
I.
PROCEDURAL BACKGROUND
The procedural background for this case is atypical. It involves two lawsuits, four ex parte
reexaminations of the asserted patent by the U.S. Patent and Trademark Office (USPTO), an
appeal to the Board of Patent Appeals and Interferences (BPAI), two court orders staying the
case, and the issuance of two Reexamination Certificates. Plaintiff filed the original lawsuit
against Intel and other defendants on November 10, 2005. See Microlinc, L.L.C. v. Acer Am.
Corp., et al., No. 2:05-cv-514 (the First Lawsuit). Shortly thereafter, Plaintiff dismissed the First
Lawsuit and submitted the ’488 Patent for reexamination by the USPTO. Plaintiff filed its
Request for Ex Parte Reexamination of all claims of the ’488 Patent on June 30, 2006 (the First
Reexamination) (No. 90/008,106). The USPTO agreed to reexamine the patent on August 22,
Page 1 of 60
2006. Defendant Intel filed a separate Request for Ex Parte Reexamination on August 28, 2006
(the Second Reexamination) (No. 90/008,196). The First Reexamination and the Second
Reexamination were then merged on October 12, 2007. On October 29, 2007, the USPTO issued
a non-final Office Action confirming the patentability of claims 10 and 13 of the ’488 Patent.
Plaintiff then filed the present suit on November 7, 2007, even though the ’418 Patent was
still in reexamination. Intel then filed its second Request for Ex Parte Reexamination on March
26, 2008 (the Third Reexamination) (No. 90/009,098), which was later merged with the First and
Second Reexaminations on July 19, 2008. On May 29, 2008, the defendants sought to stay the
case pending the reexamination of the ’488 Patent (Doc. No. 111). Given the exceptional
circumstances surrounding the pending reexamination, the Court granted the defendants’ motion
on September 16, 2008 (Doc. No. 154).
On October 28, 2008, the USPTO issued a Notice of Intent to Issue Reexamination
Certificate for the ’488 Patent confirming the patentability of original claim 13 and new claims
14-28. On December 23, 2008, Intel filed a third Request for Ex Parte Reexamination (the Fourth
Reexamination) (No. 90/009,376), which was granted on March 16, 2009. Despite the pending
fourth reexamination request, the USPTO issued the first Reexamination Certificate for the ’488
Patent on February 10, 2009. Accordingly, the Court lifted the stay in this case on August 14,
2009 (Doc. No. 201).
On June 30, 2010, the defendants sought to stay this case for a second time because the
USPTO issued a final rejection of all pending claims in the fourth reexamination (Doc. No. 314).
On September 20, 2010, the Court granted the defendants’ motion because Plaintiff was
continuing to amend the claims and prosecute new claims in the pending reexamination (Doc.
No. 374). On January 12, 2011, Plaintiff appealed the final rejection to the Board of Patent
Appeals and Interferences (BPAI). On June 6, 2012, the BPAI reversed the examiner on certain
Page 2 of 60
rejections pertaining to independent claims 14, 27–29, 31, 34, and 35, and dependent claims 15–
17, 20–26, 30, and 37–40, and found that claims 14–17 and 27–31 were patentable over the prior
art of record. On November 13, 2012, the USPTO issued the second Reexamination Certificate
for the ’488 Patent, which sets forth that: (l) the patentability of claims 14–17 and 20–28 was
confirmed; (2) claims 1–12 were previously canceled; (3) claims 13, 18, and 19 were canceled;
and (4) new claims 29–39 were added and determined to be patentable. The USPTO issued a
Certificate of Correction on February 12, 2013, which corrected errors appearing in the claims of
the ’488 Patent. On October 25, 2012, the Court entered the Fifth Amended Scheduling and
Discovery Order (Doc. No. 408).
II.
BACKGROUND OF THE TECHNOLOGY
The ’488 Patent issued on December 28, 1999 and generally relates to a physically nondistributed microprocessor-based computer that includes a “packet-based” communication
system that allows a computer’s components—such as a microprocessor, main memory, and disk
drive—to communicate with each other. Specifically, the specification states that the “present
invention provides a system with the configuration flexibility of a bus-based PC while reducing
the electrical problems” by applying “a point-to-point packetized interconnection structure to
facilitate communication between functional units (e.g., processor, memory, disk, I/O, etc.) within
a PC.” ’488 Patent at 2:12–18. Thus, the invention purports to overcome “several limitations and
unique problems associated” with prior art bus systems. ’488 Patent at 1:26–2:63.
The ’488 Patent explains that in the prior art computer, components talked to each other
over a shared “bus.” ’488 Patent at 1:10–2:11, 4:26–31, 4:63–5:18. Figure 1 from the ’488 Patent
is an example of a prior art bus system, where components such as the video card 45 and the
disk/storage device 46 all share access to one shared bus 43 in order to communicate with the
Page 3 of 60
microprocessor 40 or system memory 42:
In this shared bus system, each message is broadcast to all components, which means that
only one conversation can occur at a time over the bus. The specification identifies this as the
single transaction and sequential problem because only two functional units can communicate
with one another at any given time and establishing this communication required considerable
handshaking. ’488 Patent at 1:26–33.
In addition to this problem, the ’488 Patent identified a number of other problems with the
prior art bus system. ’488 Patent at 1:33–2:63. For example, the specification states that the prior
art bus system required that “all functional units connected to the bus must meet the electrical
specifications and requirements of the bus even if these specifications and requirements are quite
dated, technologically.” Id. Additionally, the specification states that “because the bus is a generic
interconnect in nature, it can not be truly optimized for communication between any specific
subset of functional units without adversely affecting communication performance between
another subset of functional units.” Id. The specification further states that “the speed of the bus is
substantially slower than might otherwise be obtainable” because of the electrical noise margin
requirements that restrict the speed of the bus and the fact that the highly variable electrical
loading of the bus requires it to be “slowed down to a rate that would sustain a worst case loading
situation even though this may occur in one PC in a thousand.” Id. Finally, the specification states
Page 4 of 60
that “[o]ther major drawbacks of a bus are the need for electrical handshake signals and its fixed
electrical data width (i.e., 8 bits, 16 bits, 32 bits, etc.).” Id.
The ’488 Patent purports to solve these stated problems by disclosing a personal
computer system in which data is transmitted in the form of “packets.” Rather than broadcasting
a message to all components, a packet is transmitted through a network analogous to how
packages are delivered by the postal service. Each packet (like a package) contains a “payload”
of data (contents of a package) and a “header” (address on a package). The header routes the
packet to the intended receiver, which then accesses the payload (opening the package and
retrieving the contents). See ’488 Patent at 5:45–6:25, Figs. 3–8. Because the components of a
personal computer may not be designed to communicate with each other using packets, the ’488
Patent provides for packet communication through “packet nodes.” This is shown in Figures 2,
24, and 25:
Page 5 of 60
These figures show that packet nodes 51 (outlined in green) include a functional unit 48
(shaded in purple) and a “link interface” 49 (shaded in yellow). The functional unit 48 portion of
a packet node includes a component of a personal computer (e.g., a processor, a memory, a
display, or an input/output device). See ’488 Patent at Fig. 2. The packet nodes are connected
with a packet-based data channel that extends between the interfaces of the packet nodes. ’488
Patent, Abstract at 1:5–8, 4:23–25. This “high speed message passing packetized interconnect
channel and protocol is intended to allow asynchronous communication between any two nodes,
so equipped, in any system configuration using packets of data or instructions.” ’488 Patent at
4:32–35. Thus, according to the specification, the claimed invention avoids being “a hardware
restricted, single transaction, high overhead interconnect (i.e., a bus).” ’488 Patent at 4:52–53.
The asserted claims focus on two sets of structures included in a functional unit’s
interface: (1) a “linc cache,” and (2) a set of two buffers, in which one buffer is used to queue
packets for transmission, and a separate buffer is used to store packets for retransmission, if
necessary. Those features are illustrated in Figure 9, which is an expansion of a functional unit’s
interface 49 (shaded in yellow in the previous figures):
Page 6 of 60
The Court turns first to the “cache” structure. A “cache” may be generically described as
high-speed memory used for temporary storage of frequently-used data (see, e.g., Doc. No. 5114 at 1). The patent’s “linc cache” 63 (shaded above in blue) is one type of cache located in the
functional unit’s link interface. The linc cache stores packets for use by functional unit 48
(shaded above in purple). ’488 Patent at 8:63–64. This can reduce the need for the functional
unit 48 to send requests for data since the data may already be stored in the linc cache.
The Court turns next to the “two buffer” structure. The “first buffer” in the interface is
illustrated above as “request queue” 64 (shaded in yellow). This buffer queues packets until they
are transmitted over data channel 50. ’488 Patent at 8:28–29, Fig. 9. A copy of the transmitted
packet is stored in a separate “second buffer,” shown above as the “echo waiting queue” 65
(shaded in green). ’488 Patent at 8:45–49, Fig. 9. If a transmission error occurs, the copy in the
second buffer is retransmitted. ’488 Patent at 9:61–10:1.
Thus, according to the specification of the ’418 Patent, the point-to-point interconnection
Page 7 of 60
scheme of the invention solves the problems with prior art bus systems because it: (1) “is of
relatively fixed electrical load and can, therefore, be optimized for speed,” (2) “provides a means
of eliminating the typical physical layer control signals of a bus and replacing them with linklayer control which is much more flexible,” and (3) “improves speed and performance at
reduced cost and with better noise characteristics (both internal electrical noise and radiated
EMI) as compared to the bus interconnect currently employed within a PC.” ’488 Patent at 2:19–
43.
The abstract of the ’488 Patent is reproduced below:
A physically non-distributed microprocessor-based computer
includes a microprocessor, and a random access memory device, a
mass storage device, and an input-output port device, all operable
from the microprocessor and including an interface for receiving
and transmitting data in packet form. A novel packet-based data
channel extends between the microprocessor and the interfaces of
the devices to provide communication between the microprocessor
and the devices. By varying the size of the packets in accordance
with actual data transmission requirements improved computer
performance is achieved.
As an exemplary claim of the ’488 Patent, claim 14 is reproduced below:
A physically non-distributed
computer system comprising:
microprocessor-based
a microprocessor;
a random access memory device;
a mass storage device;
an input-output port device;
said devices each being operable in conjunction with said
microprocessor; said microprocessor and said devices
each including an interface for receiving and
transmitting data in packet form;
a packet-based data channel extending between said
microprocessor and said interfaces of said devices for
providing simultaneous bi-directional communication
Page 8 of 60
between said microprocessor and said devices; and
a first one of said interfaces including a first buffer for
queuing packets for a transmission, and a second buffer
for storing a copy of a packet transmitted by said first
interface to a second one of said interfaces, said second
buffer storing said copy pending receipt of a reply
packet acknowledging receipt of said transmitted packet
by said second interface.
III.
LEGAL STANDARD
Claim construction is a matter of law. Markman v. Westview Instruments, Inc., 52 F.3d
967, 979 (Fed. Cir. 1995). The purpose of claim construction is to resolve the meanings and
technical scope of claim terms. U.S. Surgical Corp. v. Ethicon, Inc., 103 F.3d 1554, 1568 (Fed.
Cir. 1997). When the parties dispute the scope of a claim term, “it is the court’s duty to resolve
it.” O2 Micro Int’l Ltd. v. Beyond Innovation Tech. Co., 521 F.3d 1351, 1362 (Fed. Cir. 2008).
The claims of a patent define the scope of the invention. Teleflex, Inc. v. Ficosa N. Am.
Corp., 299 F.3d 1313, 1324 (Fed. Cir. 2002). They provide the “metes and bounds” of the
patentee’s right to exclude. Corning Glass Works v. Sumitomo Elec. U.S.A., Inc., 868 F.2d 1251,
1257 (Fed. Cir. 1989). Accordingly, claim construction begins with and “remain[s] centered on
the claim language itself.” Innova/Pure Water, Inc. v. Safari Water Filtration Sys., Inc., 381 F.3d
1111, 1116 (Fed. Cir. 2004).
Claim terms are normally given their “ordinary and customary meaning.” Phillips v.
AWH Corp., 415 F.3d 1303, 1312 (Fed. Cir. 2005) (en banc) (quoting Vitronics Corp. v.
Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed. Cir. 1996)). Generally, “the ordinary and
customary meaning of a claim term is the meaning that the term would have to a person of
ordinary skill in the art in question at the time of the invention.” Id at 1313.
Page 9 of 60
The best guide for defining a disputed term is a patent’s intrinsic evidence. Teleflex, 299
F.3d at 1325. Intrinsic evidence includes the patent’s specification and the prosecution history.
Id.
The claims are part of the specification. Markman, 52 F.3d at 979. “[T]he context in
which a term is used in the asserted claim can be highly instructive.” Phillips, 415 F.3d at 1314;
see also Abtox, Inc. v. Exitron Corp., 122 F.3d 1019, 1023 (Fed Cir. 1997) (“[T]he language of
the claim frames and ultimately resolves all issues of claim interpretation.”). “Differences among
claims can also be a useful guide in understanding the meaning of particular claim terms.”
Phillips, 415 F.3d at 1314.
In addition to the claims, the specification’s written description is an important
consideration during the claim construction process. See Vitronics Corp., 90 F.3d at 1582. The
written description provides further context for claim terms and may reflect a patentee’s intent to
limit the scope of the claims. See Watts v. XL Sys., Inc., 232 F.3d 877, 882 (Fed. Cir. 2000).
“[T]he specification ‘is always highly relevant to the claim construction analysis. Usually, it is
dispositive; it is the single best guide to the meaning of a disputed term.’” Phillips, 415 F.3d at
1315 (quoting Vitronics, 90 F.3d at 1582).
The specification may also resolve ambiguous claim terms “where the ordinary and
accustomed meaning of the words used in the claims lack sufficient clarity to permit the scope of
the claim to be ascertained from the words alone.” Teleflex, Inc., 299 F.3d at 1325. For example,
“[a] claim interpretation that excludes a preferred embodiment from the scope of the claim ‘is
rarely, if ever, correct.’” Globetrotter Software, Inc. v. Elam Computer Grp., Inc., 362 F.3d
1367, 1381 (Fed. Cir. 2004) (quoting Vitronics Corp., 90 F.3d at 1583).
Page 10 of 60
But care must be taken to avoid unnecessarily reading limitations from the specification
into the claims. Teleflex, 299 F.3d at 1326; see also Raytheon Co. v. Roper Corp., 724 F.2d 951,
957 (Fed. Cir. 1983) (“That claims are interpreted in light of the specification does not mean that
everything expressed in the specification must be read into all the claims.”). “[P]articular
embodiments appearing in the written description will not be used to limit claim language that
has broader effect.” Innova/Pure Water, 381 F.3d at 1117; see also Phillips, 415 F.3d at 1323
(“[A]lthough the specification often describes very specific embodiments of the invention, we
have repeatedly warned against confining the claims to those embodiments.”).
The prosecution history is also part of the intrinsic evidence. Phillips, 415 F.3d at 1317. It
“consists of the complete record of the proceedings before the PTO and includes the prior art
cited during the examination of the patent.” Id. “As in the case of the specification, a patent
applicant may define a term in prosecuting a patent.” Home Diagnostics, Inc. v. LifeScan, Inc.,
381 F.3d 1352, 1356 (Fed. Cir. 2004). Statements made during the prosecution of the patent may
limit the scope of the claims. Teleflex, 299 F.3d at 1326; see Omega Eng’g Inc. v. Raytek Corp.,
334 F.3d 1314, 1323 (Fed. Cir. 2003) (explaining that the doctrine of prosecution disclaimer
“preclud[es] patentees from recapturing through claim interpretation specific meanings
disclaimed during prosecution”).
Finally, the Court may rely on extrinsic evidence to aid with understanding the meaning
of claim terms. Markman, 52 F.3d at 981. Extrinsic evidence includes “all evidence external to
the patent and prosecution history, including expert and inventor testimony, dictionaries, and
learned treatises.” Id. at 980. Extrinsic evidence is generally less useful or reliable, Phillips, 415
F.3d at 1317, and it should not be relied on when it contradicts the intrinsic evidence, Markman,
52 F.3d at 981.
Page 11 of 60
IV.
DISCUSSION
The parties’ dispute focuses on the meaning and scope of 14 terms/phrases in the claims
of the ’488 Patent. Additionally, the parties have agreed on the construction of a number of other
terms.
a. “microprocessor” and “[a] physically non-distributed microprocessorbased computer system”
Disputed Claim Term
Plaintiff’s Proposed
Construction
Ordinary meaning, in the
alternative: “a central
processing unit residing
within a single chip”
“microprocessor”
“[a] physically nondistributed microprocessorbased computer system”
Defendants’ Proposed
Construction
“the central processing unit of
a computer on a single
integrated chip”
“a microprocessor-based
computer system in which all
the components are close to
one another and contained
within a single enclosure,
e.g., a personal computing
device or server”
“a microprocessor-based
computer system (other than
a multiprocessor-based
system) in which all the
components are close to one
another (e.g., a personal
computer within a single
enclosure)”
The parties dispute whether the term “microprocessor” and the phrase “[a] physically
non-distributed microprocessor-based computer system” should be limited to a single
microprocessor or whether the terms include multiprocessor-based systems. Defendants contend
that it should be limited to a single microprocessor based on prosecution history estoppel.
Standard Oil Co. v. Am. Cyanamid Co., 774 F.2d 448, 452 (Fed. Cir. 1985) (“[T]he prosecution
history (or file wrapper) limits the interpretation of claims so as to exclude any interpretation that
may have been disclaimed or disavowed during prosecution in order to obtain claim
allowance.”). Specifically, Defendants contend that the patentee argued to the USPTO that
certain prior art was different from the claimed invention because it disclosed “a packet-based
Page 12 of 60
interconnect for use in high-performance multiprocessor systems” (Doc. No. 511-32 at 7
(emphasis added)). Defendants argue that this statement represents a clear disavowal of a
multiprocessor-based system and that their proposed constructions properly limit the claims to
systems with one microprocessor. Plaintiff disagrees and contends that the claim terms should
not be limited to a single microprocessor or a microprocessor with only a single core.
The Court agrees with Plaintiff and finds that the patentee did not make a clear and
unmistakable disclaimer of a physically non-distributed unit containing more than one
microprocessor or a microprocessor with more than one core. However, as discussed in more
detail below, the Court concludes that Plaintiff did make a clear disclaimer of a physically
distributed system (i.e., where the multiple microprocessors are not contained within a single
enclosure or common housing). This conclusion is consistent with the representations made by
Plaintiff during the claim construction hearing that the patentee limited the invention to
physically non-distributed systems. Accordingly, the Court concludes that Plaintiff’s proposed
constructions are consistent with the claim language, the specification, and the prosecution
history.
i.
The Claim Language
The Court first turns to the language of the claims, as it provides “substantial guidance as
to the meaning of particular claim terms.” Phillips, 415 F.3d at 1313 (citing Vitronics Corp. v.
Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed. Cir. 1996)). The term “microprocessor” appears in
claims 14, 24, 27–29, 31–35, and 37. The phrase “[a] physically non-distributed microprocessorbased computer system” appears in claims 14, 27–29, and 31–35. Having reviewed the claims,
the Court concludes that the term “microprocessor” term, and the phrase “[a] physically nondistributed microprocessor-based computer system” are used consistently in each claim and are
Page 13 of 60
meant to have the same meaning in each claim. The Court further concludes that the plain
language of the disputed phrase requires a “physically non-distributed” system (e.g., a computer
within a single enclosure).
Furthermore, the Court concludes that the plain language of the claims requires at least
one microprocessor, but that the claims are not limited to systems having only a single
microprocessor or a microprocessor with only a single core. Indeed, Defendants’ proposed
constructions ignore a fundamental principle of claim construction by replacing the article “a”
with the article “the” and suggesting that there can only be a single central processing unit. It is
well established that “a” is generally understood to mean “one or more.” Baldwin Graphic Sys.,
Inc. v. Siebert, Inc., 512 F.3d 1338, 1342 (Fed. Cir. 2008) (“That ‘a’ or ‘an’ can mean ‘one or
more’ is best described as a rule, rather than merely as a presumption or even a convention.”).
As discussed in more detail below, nothing in the claim language, specification, or prosecution
history indicates that the patentee departed from the general rule for the article “a.” Accordingly,
the Court finds that Plaintiff’s constructions are consistent with the ordinary and customary
meaning of the claim language as understood by one of ordinary skill in the art. The Court will
now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, supports the adoption of Plaintiff’s constructions.
Although the specification does not explicitly disclose an embodiment including a
multiprocessor system or a multicore processors, the specification states that “[w]hile a particular
embodiment of the invention has been shown and described, it will be obvious to those skilled in
the art that changes and modifications may be made therein without departing from the invention
Page 14 of 60
in its broader aspects, and, therefore, the aim in the appended claims is to cover all such changes
and modifications as fall within the true spirit and scope of the invention.” ’488 Patent at 11:49–
55. Moreover, it is well established that claims are not limited to the embodiments disclosed in
the specification. Rexnord Corp. v. Laitram Corp., 274 F.3d 1336, 1344 (Fed. Cir. 2001) (“[A]n
applicant is not required to describe in the specification every conceivable and possible future
embodiment of his invention.”).
Defendants respond that general claim interpretation guidelines do not trump a clear
disclaimer. Omega Eng’g, Inc, v. Raytek Corp., 334 F.3d 1314, 1324 (Fed. Cir. 2003) (“[W]here
the patentee has unequivocally disavowed a certain meaning to obtain his patent, the doctrine of
prosecution disclaimer attaches and narrows the ordinary meaning of the claim congruent with
the scope of the surrender.”). However, the Court does not find that the patentee unequivocally
disavowed a physically non-distributed unit containing more than one microprocessor or a
microprocessor with more than one core.
Defendants’ disclaimer argument hinges on a single statement made during the first
reexamination regarding the Scalable Coherent Interface References (“SCI References”).
Specifically, in disclosing these references, the patentee stated:
[The SCI References] disclose a packet-based interconnect
for use in high-performance multiprocessor systems. While these
references disclose a packet-based interconnection, and therefore
are arguably relevant to claims 1-13 of the ’488 reference, these
references do not disclose an interconnect for use in physically
non-distributed microprocessor based computer systems.
(Doc. No. 511-32 at 7). Defendants contend that this single statement represents a clear
disavowal of multiprocessor systems. The Court disagrees and finds that as it relates to the
“microprocessor” element of a non-distributed system, one of ordinary skill in the art would find
Page 15 of 60
that the patentee’s statement is ambiguous and amenable to multiple reasonable interpretations,
and thus, is not a clear disavowal. Elbex Video, Ltd. v. Sensormatic Elecs. Corp., 508 F.3d
1366, 1373 (Fed. Cir. 2007) (“For a prosecution statement to prevail over the plain language
of the claim, the statement must be clear and unmistakable . . . .”); see also N. Telecom Ltd. v.
Samsung Elecs. Co., 215 F. 3d 1281, 1293–95 (Fed. Cir. 2000) (holding that prosecution history
failed to support the judicial narrowing of a claim term because the inventors’ statements
concerning the invention and the prior art were amenable to multiple reasonable interpretations).
The parties agree that the SCI References are directed towards a physically distributed
system. In fact, one of the SCI specifications states that its purpose is “[t]o define an interface
standard for very high performance multiprocessor systems that supports . . . systems with
up to 64K nodes” (Doc. No. 504-4 at 3). Accordingly, a person of ordinary skill in the art
would understand that the SCI References at least pertain to physically distributed
computing systems that may be used for identifying a large number of physically distributed
nodes. Thus, the Court concludes that the patentee clearly disclaimed physically distributed
systems in distinguishing the SCI References.1
However, in distinguishing the SCI References, it is unclear if the patentee disclaimed
anything other than a physically distributed system. Specifically, the patentee stated that “these
references do not disclose an interconnect for use in physically non-distributed microprocessor
based computer systems” (Doc. No. 511-32 at 7 (emphasis added)). The patentee did not
mention multiprocessor systems or multicore systems in this statement, but instead uses the term
1
The Court’s conclusion is consistent with the claim language, which recites a “physically nondistributed” computer system. Likewise, the Court’s conclusion is consistent with the
representations made by Plaintiff during the claim construction hearing that the patentee limited
the invention to physically non-distributed systems.
Page 16 of 60
“microprocessor.” In other words, if the patentee’s statement was that “these references do not
disclose an interconnect for use in physically non-distributed [multiprocessor] based computer
systems,” then a disclaimer finding would be warranted. Accordingly, the Court finds that the
patentee did not unequivocally disclaim a multiprocessor system or a multicore system included
in a physically non-distributed system.
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
adopts Plaintiff’s proposed constructions and construes a “microprocessor” to mean “a
central processing unit residing within a single chip.” The Court construes “[a] physically
non-distributed microprocessor-based computer system” to mean “a microprocessor-based
computer system in which all the components are close to one another and contained
within a single enclosure.”
b.
“said microprocessor and said devices each including an interface for
receiving and transmitting data in packet form”
Disputed Claim Term
Plaintiff’s Proposed
Construction
“said microprocessor and said “each of the microprocessor,
devices each including an
the random access memory
interface for receiving and
device, the mass storage
transmitting data in packet
device, and the input-output
form”
port device has at least one
interface connected to it, for
receiving and transmitting
data in packet form”
Defendants’ Proposed
Construction
“each of the microprocessor,
the random access memory
device, the mass storage
device, and the input-output
port device includes its own
interface for receiving and
transmitting data in packet
form”
The parties’ dispute is focused on two issues. The first issue is whether the claimed
“interface” must be physically incorporated within each of the claimed devices (i.e., the
microprocessor, the mass storage device, the random access memory). The second issue is
Page 17 of 60
whether the claimed devices may share an interface or whether each of the claimed devices must
have a separate interface.
Regarding the first issue related to the physical incorporation, Defendants contend that
their proposed construction uses the word “including” exactly as the inventor used the word in
the specification and that the interfaces must be physically incorporated within each device.
Defendants further contend that Plaintiff’s position would effectively overrule the USPTO’s
determination that the interface must be included in the functional unit. Plaintiff responds that
the specification of the ’488 Patent expressly depicts and discusses embodiments of an
interface that is separate from but connected to a functional unit, in addition to embodiments of
a link interface that is integrated within a functional unit. See e.g., ’418 Patent at 9:28–40, Figs.
13, 14. The Court agrees with Plaintiff and finds that the claimed “interface” may either be
incorporated within each of the claimed devices or connected to the claimed devices.
Regarding the second issue related to the shared interface, Plaintiff concedes that the
amended claims require an interface for the microprocessor and for each device but contends
that the “interface” language was not at issue during the reexamination. Plaintiff further
contends that Defendants have not pointed to any discussion in the prosecution history that
prohibits functional units from sharing an interface. Defendants respond that the plain language
of the claims, specification, and file history make clear that each device must have its own
interface. The Court agrees with Defendants because the patentee amended the claims during the
reexamination to emphasize that the claimed devices each require a separate interface. Moreover,
during the claim construction hearing, Plaintiff agreed with the Court that each of the claimed
devices must have its own interface. Accordingly, the Court does not adopt either side’s
construction. Instead, the Court provides a construction that is consistent with the claim
Page 18 of 60
language, the specification, and the prosecution history.
i.
The Claim Language
The phrase “said microprocessor and said devices each including an interface for
receiving and transmitting data in packet form” appears in claims 14, 27–29, and 31–35. Having
reviewed the claims, the Court concludes that the phrase is used consistently in each claim and is
meant to have the same meaning in each claim. Regarding the second issue related to the shared
interface, the Court finds that the plain language of the claims requires not only that the devices
are “operable in conjunction with said microprocessor,” but the claims also require that the
microprocessor and the devices “each includ[e] an interface for receiving and transmitting data
in packet form.” Thus, the claim language requires that the microprocessor and each claimed
device must have its own interface for receiving and transmitting data in packet form.
Regarding the first issue related to the physical incorporation, Plaintiff correctly contends
that as a general matter of patent claim construction, “includes” and “comprising” are openended terms which are well understood to mean “including but not limited to.” Cias, Inc. v.
Alliance Gaming Corp., 504 F.3d 1356, 1360–61 (Fed. Cir. 2007); Sandisk Corp. v. Memorex
Prods., Inc., 415 F.3d 1278, 1284 (Fed. Cir. 2005) (“As a patent law term of art, ‘includes’
means ‘comprising.’”) (citations omitted). Thus, the Court agrees that the use of “including” in
the claims is not a limiting term, and that the “elements” linked together by the term “including”
can have additional elements, such as, connections.
Additionally, Plaintiff contends that claims 20 and 21 are dependent claims which
further define the relationship of the interface to the functional unit, respectively, as either
connected to or integrated within the functional unit. The ’488 Patent specification supports
these narrower claims because it discloses multiple arrangements of the interface and functional
Page 19 of 60
unit that would fall within the embodiments claimed in dependent claims 20 and 21. See e.g.,
’488 Patent, 9:28–40, Figs. 13, 14. Accordingly, the Court concludes that the claimed “interface”
may be either connected to or integrated within the functional unit. Phillips v. AWH Corp., 415
F.3d, 1303, 1314–15 (en banc) (“[T]he presence of a dependent claim that adds a particular
limitation gives rise to the presumption that the limitation in question is not present in the
independent claim.”). The Court will now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, supports the Court’s construction. Regarding the first
dispute related to the physical incorporation, the specification depicts and discusses
embodiments of an interface (i.e., link interface) that is separate from but connected to a
functional unit, in addition to embodiments of an interface that is integrated within a functional
unit. See e.g., ’488 Patent at 9:28–40, Figs. 13, 14.
Defendants do not dispute that the specification discloses these embodiment but instead
argue that their proposed construction uses the word “including” exactly as the inventor used the
word in the specification. Specifically, when referring to Figure 14, the specification describes
the first interface as “include[d]” within the processor but is silent about the second interface.
’488 Patent at 9:37–40. Thus, according to Defendants, because the inventor did not use
“including” when referring to both of the illustrated interfaces, the claimed “interface” must be
“included” or physically incorporated within the functional unit. The Court disagrees because
“[a] claim construction that excludes a preferred embodiment . . . ‘is rarely, if ever, correct.’”
Sandisk Corp., 415 F.3d at 1285 (quoting Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576,
1583 (Fed. Cir. 1996)).
Page 20 of 60
Plaintiff correctly notes that the language in the specification is “include[d] within the
processor silicon itself.” ’488 Patent at 9:39–40 (emphasis added). Moreover, as discussed
above, the claim language does not include the further qualification that the interface be
included within the functional unit itself. Furthermore, the specification continues to state that
“[i]t will be shown in the description to follow how the memory node interface may be extended
to include any peripheral device.” ’488 Patent at 9:44–46 (emphasis added). The Court agrees
that a person of ordinary skill in the art would not interpret this use of “include” to be restricted
to “physically incorporating” a peripheral device into a memory node interface. Finally,
Defendants failed to point to any evidence in the prosecution history that would warrant
restricting the meaning of “including an interface for receiving and transmitting data in
packet form” to a structure wherein a link interface is incorporated within the functional
unit.
Regarding the second issue related to the shared interface, Defendants contend that the
specification and file history make clear that each device must have its own interface. The Court
agrees because in every disclosed embodiment, each claimed device has its own interface. See
’488 Patent at Figs. 2, 24, 25. Plaintiff responds that Figure 24 of the ’488 Patent explicitly
discloses a configuration where one packet node (a device plus an interface) provides an
interface for other devices. Plaintiff, however, does not provide any citation to the specification
that confirms Plaintiff’s theory.
Moreover, the patentee amended the claims during reexamination to emphasize that the
claimed devices “each” include an interface. Specifically, the patentee canceled claim 1 and
added new claims (14, 27, and 28) that require one interface for each claimed device, as
illustrated below:
Page 21 of 60
Claim 1 (canceled)
Claims 14, 27, 28, 31, 32, 33, 34, 35
said devices each being operable in
said devices each being operable in
conjunction with said microprocessor and conjunction with said microprocessor, said
including an interface for receiving and microprocessor and said devices each
including an interface for receiving and
transmitting data in packet form
transmitting data in packet form
(Doc. No. 511-23 (Suppl. Am. in Consolidated Ex Parte Reexams., Control No. 90/008,106, at
2–3, 6–7 (Jan. 21, 2008))). These amendments recite that the microprocessor must include an
interface and that “each” device must include its own interface. Pfund v. United States, 40 Fed.
Cl. 313, 367 (Fed. Cl. 1998) (“In interpreting the scope of claims, [courts] should endeavor to
give meaning to each claim term.”). Thus, the patentee narrowed the claim scope during the
reexamination, and the disputed phrase will not be construed to include a broader scope. Elekta
Instrument S.A. v. O.U.R. Scientific Int’l, Inc., 214 F.3d 1302, 1308 (Fed. Cir. 2000) (“Claims
that have been narrowed in order to obtain issuance over the prior art cannot later be interpreted
to cover that which was previously disclaimed during prosecution.”). Finally, during the claim
construction hearing, Plaintiff agreed with the Court that each claimed device must have its own
interface. Accordingly, the Court does not adopt either side’s construction but instead provides
a construction that is consistent with the claim language, the specification, and the prosecution
history
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
construes “said microprocessor and said devices each including an interface for receiving
and transmitting data in packet form” to mean “each of the microprocessor, the random
access memory device, the mass storage device, and the input-output port device has its
own interface for receiving and transmitting data in packet form.”
Page 22 of 60
c. “interface”
Disputed Claim Term
Plaintiff’s Proposed
Construction
“a circuit that links a packetbased data channel to either a
microprocessor or a device”
“interface”
Defendants’ Proposed
Construction
“a circuit that links a packetbased data channel to either a
microprocessor or a device
and that speaks the same
language electrically and
logically with each other such
circuit”
The parties agree that the claimed “interface” is “a circuit that links a packet-based data
channel to either a microprocessor or a device.” Thus, the dispute is whether the claimed
interfaces must “speak the same language electrically and logically with each other.” Defendants
contend that the interfaces must speak the same language based on the problem statement of the
invention and statements made in the specification. Plaintiff responds that the portion of the
specification referenced by Defendants is directed to a “packet node,” not the claimed
“interface,” and that nothing in the intrinsic record justifies Defendants’ attempt to further limit
the meaning of the claimed interface. The Court agrees with Defendants. Accordingly, the Court
finds that Defendants’ proposed construction is consistent with the claim language, the
specification, and the prosecution history.
i.
The Claim Language
The term “interface” appears in claims 14, 16–17, 20–21, 27–29, 31–35, and 37. Having
reviewed the claims, the Court concludes that the term is used consistently in each claim and is
meant to have the same meaning in each claim. The Court also finds that the claim language
recites an “interface for receiving and transmitting data in packet form.” Thus, the data that is
transmitted and received by the interface is in packet form. The Court will now turn to the
Page 23 of 60
other intrinsic evidence.
ii.
The Intrinsic Evidence
The other intrinsic evidence in this case, specifically the specification and prosecution
history, supports the adoption of Defendants’ construction. The specification states that “[t]he
present invention applies a point-to-point packetized interconnection structure to facilitate
communication between functional units (e.g., processor, memory, disk, I/O, etc.) within a PC.”
’488 Patent at 2:16–19 (emphasis added). The patentee accomplished this point-to-point
packetized interconnection structure by using what the specification refers to as a packet node,
which is the combination of a link interface and a functional unit. ’488 Patent at 4:36, Fig. 2.
Specifically, “[i]nterface 49, which is the embodiment of the present invention, provides a
seamless interface between functional unit 48 and the physical layer 50. Physical layer 50
connects to any other packet node 51, thereby, in conjunction with interface 49, allowing
communication between any group of functional units.” ’488 Patent at 4:37–43 (emphasis
added). Thus, the invention purports to overcome the drawbacks associated with “defining
handshaking and data size at the physical layer” by “eliminating the typical physical layer
control signals of a bus and replacing them with link-layer control which is much more flexible.”
’488 Patent at 1:61–2:26.
To support their contention that the interface must speak the same language electrically
and logically with each other, Defendants point to the following passage from the specification:
[T]he present invention makes each node . . . interface with each
other using a uniform packe[t] based message passing interface. The
invention creates packet nodes which speak the same language
electrically and logically. This simplifies communication and minimizes
traditional handshaking and overhead.
’488 Patent at 4:44–50 (emphasis added). Defendants argue that their construction expressly
Page 24 of 60
follows the language and teaching from the patent that each interface is required to “speak the
same language electrically and logically.” The Court agrees.
As an initial matter, the “use of the phrase ‘the present invention’ does not
‘automatically’ limit the meaning of claim terms in all circumstances, and . . . such language
must be read in the context of the entire specification and prosecution history.” Netcraft Corp. v.
eBay, Inc., 549 F.3d 1394, 1398 (Fed. Cir. 2008). Moreover, the Court must be careful “to avoid
the danger of reading limitations from the specification into the claim.” Phillips, 415 F.3d at
1323. In this case, however, the Court concludes that the patentee limited the scope of the claims
by describing features as “the present invention.”
The patentee’s repeated use of the phrases “the present invention” and “the invention” in
describing the claimed interface makes clear that the specification is describing the invention as
a whole, and not merely referring to a particular embodiment. See, e.g., Netcraft Corp., 549 F.3d
at 1398 (“[T]he common specification’s repeated use of the phrase ‘the present invention’
describes the invention as a whole . . . .”); Verizon Servs. Corp. v. Vonage Holdings Corp., 503
F.3d 1295, 1308 (Fed. Cir. 2007) (“When a patent thus describes the features of the ‘present
invention’ as a whole, this description limits the scope of the invention.”); Honeywell Int'l, Inc.
v. ITT Indus., Inc., 452 F.3d 1312, 1318 (Fed. Cir. 2006) (limiting claims to a fuel filter where
“the written description refers to the fuel filter as ‘this invention’ or ‘the present invention’”).
For example, the specification describes the advantages of using packet nodes by stating that
“the present invention allows for more intelligent and flexible information exchange between
nodes while requiring only that overhead and handshaking required for a specific transaction.”
’488 Patent at 4:56–59 (emphasis added). This statement, taken in context with the portion of the
specification relied on by Defendants, confirms that the claimed invention requires interfaces
Page 25 of 60
that “speak the same language electrically and logically” because the goal of the “present
invention” was to avoid imposing “a set of costly rules and formalism” that slow down
communication. ’488 Patent at 4:60–61.
Plaintiff responds that the portion of the specification relied on by Defendants does not
define the claimed “interface” but instead relates to a packet node. As it relates to the present
issue, this is a distinction without a difference because the specification defines a packet node as
combination of the claimed “interface” and the claimed “devices” (i.e., a functional unit). ’488
Patent at 4:36, Fig. 2. Moreover, it is “[i]nterface 49, which is the embodiment of the present
invention, [that] provides a seamless interface between functional unit 48 and the physical layer
50. Physical layer 50 connects to any other packet node 51, thereby, in conjunction with interface
49, allowing communication between any group of functional units.” ’488 Patent at 4:37–43
(emphasis added). Thus, it is the “interface” portion of each packet node that provides the
“seamless interface,” not the functional unit. Accordingly, one of ordinary skill in the art would
understand that it is the interface that allows the packet nodes to “speak the same language
electrically and logically.”
Plaintiff also contends that Defendants’ construction would ignore disclosed situations
where interfaces would not “speak the same language electrically and logically.” The Court
disagrees. Defendants’ construction does not require that the interfaces use a particular electrical
voltage or exclude claimed devices that have different physical data widths. ’488 Patent at 5:25–
31, 6:8–10. Instead, Defendants’ construction only requires that each interface speak the same
language electrically and logically. Accordingly, the Court finds that Defendants’ construction
is consistent with the intrinsic evidence.
Page 26 of 60
iii.
Extrinsic Evidence
In addition to the intrinsic evidence, there is also extrinsic evidence that supports
Defendants’ construction. The inventor, Mr. Gautam Kavipurapu, testified that that the
interfaces speak a common language by describing that “from a link interface to a link interface,
you have the same electrical and logical interface” ( Doc. No. 511-24 at 2 (Kavipurapu Tr.
453:9–21) (emphasis added))). This is consistent with a construction that requires each interface
“speak the same language electrically and logically.”
iv.
Court’s Construction
In light of the claim language, the specification, the prosecution history, and the
extrinsic evidence submitted by the parties, the Court adopts Defendants’ proposed
construction and construes “interface” to mean “a circuit that links a packet-based data
channel to either a microprocessor or a device and that speaks the same language
electrically and logically with each other such circuit.”
d. “a packet-based data channel extending between said microprocessor
and said interfaces of said devices”
Disputed Claim Term
“a packet-based data channel
extending between said
microprocessor and said
interfaces of said devices”
Plaintiff’s Proposed
Construction
“a circuit capable of
transporting a packet-based
message between a
microprocessor interface and
the interfaces of the devices”
Defendants’ Proposed
Construction
“a packet-based data
channel that extends from
the microprocessor interface
to the interfaces in the
random access memory
device, the mass storage
device, and the input-output
port device for carrying
uniform packet types”
The parties’ dispute is focused on two issues. The first issue is whether a single path must
Page 27 of 60
extend between a microprocessor interface and the interfaces of each of the devices or whether
the path need only extend between a microprocessor interface and one device interface. The
second issue is whether the packet-based data channel must carry uniform packet types.
Regarding the first issue related to the extension of the path, Plaintiff contends that
requiring the same single path between the microprocessor and each device is inconsistent with
the ’488 specification and would be limiting the data channel to one disclosed embodiment.
Defendants respond that Plaintiff’s construction is incorrect because it would allow the channel
to transmit packets from the microprocessor interface to only one of the device interfaces and
would not require transmission to all of the device interfaces. As discussed below, the Court
finds merit in certain aspects of both sides’ arguments. Accordingly, the Court will require that
the data channel be capable of transporting packets between a microprocessor interface and each
interface of the devices.
Regarding the second issue related to uniform packet types, Defendants contend that
the packet-based data channel must carry uniform packets. Plaintiff responds that while a
packet-based data channel certainly may carry uniform packet types, neither the claim
language nor the prosecution history requires incorporation of this limitation into the claim.
The Court agrees with Defendant and will require that the packet-based data channel transport
uniform packet types. Accordingly, the Court does not adopt the parties’ proposed construction
but instead provides a construction that is consistent with the claim language, the specification,
and the prosecution history.
i.
The Claim Language
The phrase “a packet-based data channel extending between said microprocessor and said
interfaces of said devices” appears in claims 14, 27–29, and 31–35. Having reviewed the claims,
Page 28 of 60
the Court concludes that the phrase is used consistently in each claim and is meant to have the
same meaning in each claim. The Court also finds that the claim language requires that the
claimed interface must be capable of receiving and transmitting data in packet form. Further,
the Court finds that the claim language requires that the packet-based data channel extends
between the microprocessor and the interfaces of the devices, not just between the
microprocessor interface and the interface of one device. Thus, the Court finds that Plaintiff’s
construction that would allow the channel to transmit packets from the microprocessor interface
to only one device interface is not consistent with the ordinary and customary meaning of the
claim language. The Court will now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, supports the Court’s construction. Regarding the first
issue related to the extension of the path, the specification describes and illustrates a number of
interconnect topologies for interconnecting more than two functional units. ’488 Patent at
2:27–36. For example, Figure 24 illustrates a ring interconnect in which each link interface is
connected to a neighboring node’s link interface until the ring is closed. ’488 Patent at 7:36–
40, Fig. 24. The specification also illustrates another common interconnect topology referred
to as a switch. ’488 Patent at 7:50–63, Fig 25. Plaintiff contends that Defendants’ construction
would limit the data channel to the ring topology, thereby excluding the disclosed switch
topology because the same single path would not extend between a microprocessor interface and
the interfaces of each of the devices in the switch topology. The Court agrees with Plaintiff.
Defendants have not pointed to any evidence that indicates why the claims should not cover this
embodiment. In fact, Defendants argue that the switch configuration in Figure 25 is not
Page 29 of 60
precluded by Defendants’ proposed construction because the crossbar 68 creates a packet-based
data channel that extends all the way between interfaces that speak the same language
electrically and logically. Thus, the Court’s construction is consistent with the arguments made
by both sides because it requires a data channel capable of transporting packet between a
microprocessor interface and each interface of the claimed devices.
Regarding the second issue related to uniform packet types, Defendants contend that the
packet-based data channel must carry uniform packet types. As with the disputed term
“interface,” Defendants point to the following passage from the specification to support their
argument:
[T]he present invention makes each node … interface with each other
using a uniform packe[t] based message passing interface. The invention
creates packet nodes which speak the same language electrically and
logically. This simplifies communication and minimizes traditional
handshaking and overhead.
’488 Patent at 4:44–50 (emphasis added). Defendants contend that this emphasizes that the
interfaces must use uniform packet types. Having considered all of the intrinsic evidence, the
Court agrees. First, for the reasons discussed above, the Court finds that this passage is describing
the invention as a whole. Moreover, the passage explicitly describes the interface as using
“uniform [packet].” The use of uniform packets is important because it enables the present
invention to overcome the drawbacks associated with “defining handshaking and data size at the
physical layer” by “eliminating the typical physical layer control signals of a bus and replacing
them with link-layer control which is much more flexible.” ’488 Patent at 1:61–2:26. It does this
by creating a uniform packet based message passing interface that creates packet nodes that
speak the same language electrically and logically, thereby simplifying communication and
providing a significant increase in data-rate speed. ’488 Patent at 4:44–56. Thus, the
Page 30 of 60
specification describes that the invention, as a whole, uses uniform packet types because this
eliminates the “handshaking” and “overhead” issues that the invention was designed to solve.
See ’488 Patent at 4:44–50, 4:63–5:19, 5:49–50. Thus, the Court’s construction is consistent
with the claim language, the specification, and the prosecution history.
iii.
Extrinsic Evidence
In addition to the intrinsic evidence, extrinsic evidence also supports the Court’s
construction. Plaintiff previously submitted a declaration of Dr. Richard Belgard in support of its
claim construction positions (Doc. No. 312-4 (Belgard Decl. (June 18, 2010))). In his
declaration, Dr. Belgard concedes that in describing the problems in the prior art bus system “the
inventor believed that by providing uniform packet types, his invention would eliminate the need
for electrical handshake, or control signals” (Doc. No. 312-4 (Belgard Decl. at ¶ 85 (June 18,
2010)) (emphasis added)). Thus, while Dr. Belgard disagreed that the paragraph relied upon by
Defendants “necessarily suggests that the packet-based data channel carries ‘uniform data
packets,’” he confirmed that a person of ordinary skill in the art would understand that the
invention overcame the problems with the prior art by providing “uniform packet types” (Doc.
No. 312-4 (Belgard Decl. at ¶¶ 85–85 (June 18, 2010))). This is consistent with a construction
that requires “a data channel capable of transporting uniform packet types.”
iv.
Court’s Construction
In light of the claim language, the specification, the prosecution history, and the
extrinsic evidence submitted by the parties, the Court construes “a packet-based data channel
extending between said microprocessor and said interfaces of said devices” to mean “a data
channel capable of transporting uniform packet types between a microprocessor interface
and each interface of the devices.”
Page 31 of 60
e. “packet-based data channel”
Disputed Claim Term
“packet-based data channel”
Plaintiff’s Proposed
Construction
“a circuit capable of
transporting a packet-based
message between one
interface and another
interface”
Defendants’ Proposed
Construction
“a set of wires and associated
components necessary for
carrying packets, without
dedicated wires for address or
data signals”
The parties dispute whether the claimed “packet-based data channel” should be construed
with the negative limitation of “without dedicated wires for address or data signals.” Defendants
contend that the negative limitation is warranted because Plaintiff’s construction fails to define
what distinguishes a packet-based data channel from other types of data channels. Defendants
also argue that Plaintiff’s construction allows for a “partial” data channel, which would not meet
the “extending between” interfaces claim limitation. Plaintiff responds that the claim language
itself unambiguously states that the data channel is a “packet-based” data channel. The Court
agrees with Plaintiff and does not find that a negative limitation is warranted. Accordingly, for
the following reasons and for the reasons stated above, the Court provides a construction that
is consistent with the claim language, the specification, and the prosecution history.
i.
The Claim Language
The phrase “packet-based data channel” appears in claims 14, 20, 23, 27–29, 31–35, and
39. Having reviewed the claims, the Court concludes that the phrase is used consistently in each
claim and is meant to have the same meaning in each claim. The Court notes that the claim
language itself recites that the data channel is a “packet-based” data channel. Accordingly, the
Court is not persuaded by Defendants’ argument that a negative limitation is required to
distinguish the claimed data channel from other types of data channels. The Court will now
Page 32 of 60
turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, confirms that a negative limitation is not warranted.
Defendants first argue that Figure 13a illustrates that the packet-based data channel is composed of
“wires” (shown as arrows) and associated components necessary for carrying packets. Moreover,
Defendants argue that it is a fundamental characteristic of packet networks that address and data
information are treated interchangeably and transmitted together over the same wires.
Specifically, Defendants argue that the ’488 Patent consistently shows address and data
information being combined onto the same packet channel. See, e.g., ’488 Patent at Figs. 9, 20.
Although these statements are true, it does not support inserting a negative limitation because
such limitations, absent an explicit disavowal, are generally disfavored. See Omega Eng’g, Inc,
v. Raytek Corp., 334 F.3d 1314, 1332–33 (Fed. Cir. 2003) (noting that a negative limitation
imposed by the district found no support in the claim language, the plain and ordinary meaning
of the phrase, or in any express disclaimer within the specification).
Defendants do not point to any evidence that the patentee defined this data channel by
what it is not. Instead, the specification and intrinsic evidence confirms that “a packet-based data
channel” is a data channel capable of transporting a packet. For example, the packet-based data
channel may include at least two signal wires in each direction as illustrated in Fig. 13a, but that
does not limit the claimed invention to only two signal wires. Accordingly, the Court provides a
construction that is consistent with the claim language, the specification, and the prosecution
history.
Page 33 of 60
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
construes “packet-based data channel” to mean “a data channel capable of transporting a
packet.”
f. “packet”
Disputed Claim Term
Plaintiff’s Proposed
Construction
“a message including payload
data and the control data
indicative of the intended
recipient”
“packet”
Defendants’ Proposed
Construction
“a message, including
payload data and control
data, that contains all the
necessary information for
the intended receiver”
The parties agree that a “packet” is a message including payload data and control data.
The parties also agree that packets are for an intended receiver/recipient. Thus, the dispute is
whether the packet must “contain all the necessary information for the intended receiver” as
Defendants contend. Defendants are correct that the patentee stated that the “each packet
contains all the necessary information that is required for the intended receiver,” but
Defendants ignore the context of this statement. Accordingly, the Court finds that Plaintiff’s
proposed construction is consistent with and supported by the claim language, the
specification, and the prosecution history.
i.
The Claim Language
The term “packet” appears in claims 14–17, 20, 23, 27–35, and 39. Having reviewed the
claims, the Court concludes that the term is used consistently in each claim and is meant to have
the same meaning in each claim. The Court finds that the claim language recites that the
microprocessor and devices each include an interface for communicating in packet form. Thus,
Page 34 of 60
the claims support the parties’ proposed construction that a packet is a message including
payload data and control data and that packets are intended for a recipient. However, the claims
are silent as to Defendants’ language requiring the packet to include “all the necessary
information that is required for the intended receiver.” Accordingly, the Court finds that
Plaintiff’s proposed construction is consistent with the ordinary and customary meaning of the
claim language as understood by one of ordinary skill in the art. The Court will now turn to the
other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, supports the adoption of Plaintiff’s construction. Figure
3 illustrates the structure of a packet, and the specification provides a number of general
characteristics of an exemplary packet. ’488 Patent at 5:45–65. The Court finds this discussion
is consistent with the parties agreed construction that a packet is a message including payload
data and control data and that packets are intended for a recipient. Importantly, the patentee
states that that “[t]he above definitions may easily be changed without affecting the nature of the
invention.” Id. at 5:64–65. Thus, the Court concludes that the patentee did not set out to define
“packet” any different from its plain and ordinary meaning.
Defendants, however, argue that the patentee acted as his own lexicographer and
explicitly defined “packet” with the following statement:
A packet node in the present invention communicates with another
packet node using “packets.” Each packet contains all the necessary
information that is required for the intended receiver without the added
overhead of setting up the receiver or formatting the data to a nodespecific set up.
’488 Patent at 5:45–49 (emphasis added). But Defendants ignore the context of the statement. As
Page 35 of 60
discussed above, the invention purports to overcome the drawbacks associated with “defining
handshaking and data size at the physical layer” by “eliminating the typical physical layer
control signals of a bus and replacing them with link-layer control which is much more flexible.”
’488 Patent at 1:61–2:26. Thus, the “necessary information” contrasts how the prior art bus
system operated at the physical layer. Moreover, Defendants do not point to any evidence in
the prosecution history where a definition of packets was restricted to “containing all the
necessary information for the intended receiver.” Accordingly, the Court finds that Defendants’
proposed language is superfluous and would only confuse the jury.
iii.
Extrinsic Evidence
The parties submitted a number of technical dictionary excerpts that included definitions
for the term “packet.” The definitions for “packet” provided in these dictionaries generally refer
to payload data and control data, but none state that a packet “contains all the necessary
information for the intended receiver” (see e.g., Doc. No. 504-13 at 3 (IBM Dictionary of
Computing (10th ed. 1993) (excerpt))). Accordingly, the Court finds that the patentee used the
term “packet” consistent with its plain and ordinary meaning.
iv.
Court’s Construction
In light of the claim language, the specification, the prosecution history, and the
extrinsic evidence submitted by the parties, the Court adopts Plaintiff’s proposed construction
and construes “packet” to mean “a message including payload data and the control data
indicative of the intended recipient.”
g. “a first buffer for queuing packets for a transmission”
Disputed Claim Term
“a first buffer for queuing
Plaintiff’s Proposed
Construction
Ordinary meaning, in the
Page 36 of 60
Defendants’ Proposed
Construction
“a first buffer for storing
packets for a transmission”
alternative: “a first buffer
capable of storing two or
more packets prior to
transmission”
two or more packets until
they are transmitted from
the buffer onto the packetbased data channel”
The parties’ dispute is focused on two issues. The first issue is whether the first buffer
must store more than one packet at a time. The second issue is whether the first buffer must only
queue packets for transmission onto the packet-based data channel without some other
intermediate activity (e.g., further formatting or manipulation).
Regarding the first issue related to storing more than one packet, the Court finds
Defendants’ argument are without merit. First, both sides have agreed that a “buffer” should be
construed to mean “a storage element used to store data temporarily when sending data from one
functional unit to another.” Defendants attempt to add a further limitation to this agreed
construction under the guise of the term “queuing.” Specifically, Defendants argue that the entire
concept of “queuing” means holding at least two items, in order, such that an item that arrives
later will not pass an item that arrived earlier. Although the Court agrees that queuing means
holding items in order of arrival, the Court disagrees that this requires the claimed buffer to
always store “two or more packets.”
As described in the specification, the claimed buffer may operate in a first in, first out
(FIFO) manner. ’488 Patent at 9:61–62. This means that the claimed buffer (i.e., the storage
element used to store data temporarily) may go from an empty state to a state of storing multiple
items. This FIFO operation, however, does not require that the buffer always store two or more
items because the buffer may be empty or contain only one item. Accordingly, the Court rejects
Defendants’ argument.
Page 37 of 60
Regarding the second issue, the Court disagrees that the claims bar some other
intermediate activity between the buffer and the packet-based data channel. As discussed below,
the claim language is unambiguous about the nature of the transmission described in the claim.
Accordingly, the Court concludes that a construction is not necessary for this disputed phrase,
especially given the parties’ agreed construction for the term “buffer.”
i.
The Claim Language
The phrase “a first buffer for queuing packets for a transmission” appears in claims 14,
27–29, 31, and 33–35. Having reviewed the claims, the Court concludes that the phrase is used
consistently in each claim and is meant to have the same meaning in each claim. The Court also
finds that the claim language states that the buffer is for queuing packets “for” a transmission.
Because the claims use the term “comprising,” the claim language does not exclude further
formatting or manipulation of the data before it is placed onto the packet-based data channel.
Instead, the claims require that the first buffer provide the order that the packets will be
transmitted on the packet-based data channel. Accordingly, given the claim language, the Court
finds that no construction is necessary for this disputed phrase. A jury will understand that the
first buffer queues the packets for a transmission. The Court will now turn to the other intrinsic
evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case supports the Court’s
conclusion that no construction is necessary. As discussed, Defendants’ construction adds an
unwarranted temporal limitation because it excludes any intermediate steps before the
transmission occurs. Defendants fail to point to any substantive evidence to justify adding this
limitation. Instead, Defendants only point to Plaintiff’s characterization of the first buffer as a
Page 38 of 60
“transmit buffer” in the prosecution history to support their construction. However, there is
nothing in the intrinsic record that shows packets are transmitted by buffers onto the packetbased data channel. Indeed, the examiner confirmed this conclusion when he stated that “a buffer
is a region of memory storage. It is not a transmitter that can retransmit its content” (Doc. No.
504-15 at 4 (Office Action dated May 26, 2010 at ¶7)). In other words, the claimed first
buffer establishes the order of transmission but not the manner of transmission. This is consistent
with the construction the parties agreed to for the term “buffer” as “a storage element used to
store data temporarily.”
During the claim construction hearing, Defendants argued that the phrase should be
construed to prevent Plaintiff from arguing that the first buffer and second buffer may be
arranged in a serial configuration. Such an alignment is contrary to Figure 9, which illustrates the
first and second buffer in a parallel configuration. The Court agrees with Defendants’ argument
but finds that the plain and ordinary meaning of this disputed phrase, in conjunction with the
Court’s construction of the disputed phrase “a second buffer for storing a copy of a packet
transmitted,” properly addresses this issue. Thus, the Court need not construe the “first buffer”
phrase. Accordingly, the Court concludes that the disputed “first buffer” phrase is unambiguous,
is easily understandable by a jury, and requires no construction. Therefore, the phrase “a first
buffer for queuing packets for a transmission” will be given its plain meaning to one of
ordinary skill in the art.
h. “a second buffer for storing a copy of a packet transmitted”
Disputed Claim Term
“a second buffer for storing a
copy of a packet transmitted”
Plaintiff’s Proposed
Defendants’ Proposed
Construction
Construction
Ordinary meaning, in the
“a second buffer, separate
alternative: “a second buffer
from the first buffer, for
capable of temporarily storing storing only copies of
Page 39 of 60
a copy of a packet
transmitted”
packets already transmitted
from the first buffer onto the
packet-based data channel”
The parties dispute whether the second buffer only stores packets previously transmitted
by the first buffer or whether the second buffer can store packets before they are transmitted by the
first buffer. Plaintiff contends that the copy of a packet stored in the second buffer may be
created before the “original” packet has been transmitted (i.e., before the original packet has
left the first buffer). Defendants respond that the claim language and intrinsic evidence confirms
that the second buffer only stores packets previously transmitted by the first buffer. The Court
generally agrees with Defendants and adopts a modified version of their construction that is
consistent with the claim language, the specification, and the prosecution history.
i.
The Claim Language
The phrase “a second buffer for storing a copy of a packet transmitted” appears in claims
14, 27–29, 31, and 33–35. Having reviewed the claims, the Court concludes that the phrase is
used consistently in each claim and is meant to have the same meaning in each claim. The Court
finds that the claim language recites that the second buffer is separate from the first buffer. The
claim language also recites “a second buffer for storing a copy of a packet transmitted by said
first interface.” By using the past tense, the claims recite that the packets have already been
“transmitted” by the first interface. Plaintiff contends that the claim language “for storing”
means that the second buffer is capable of storing packets that have already been transmitted but
it does not exclude the capability to store such copies before transmission. The Court disagrees.
The patentee could have drafted the claim without the “transmitted” language. Accordingly, the
Court finds that Defendants’ proposed construction is consistent with the ordinary and
Page 40 of 60
customary meaning of the claim language as understood by one of ordinary skill in the art. The
Court will now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case, specifically the
specification and prosecution history, supports the adoption of Defendants’ construction.
Plaintiff contends that Defendants’ construction is limited to a “serial” arrangement (requiring
that the first buffer be in-line with the second buffer), which would exclude a preferred
embodiment from the claim scope. Specifically, Figure 9 depicts an embodiment that has a
“parallel” arrangement for the packet to be transmitted showing a communication path from a
first buffer (“Request Queue”) to a second buffer (“Echo Waiting Queue”) and a communication
path from the first buffer to another interface.
Defendants respond that their construction is entirely consistent with Figure 9. The Court
agrees. As stated in Plaintiff’s opening brief, the parties are in agreement that the “Echo
Waiting Queue” is claimed as the “second buffer” (Doc. No. 504 at 21). The specification
states that a request “is queued in the echo waiting queue” (the second buffer) “[o]nce any
request has been sent out.” ’488 Patent at 9:67–10:1 (emphasis added). Thus, Defendants’
construction is consistent with Figure 9 and the specification, which describes transmitting
packets from the first buffer (“Request Queue”) and providing copies of “outstanding” packets,
in parallel, to the second buffer (“Echo Waiting Queue”). ’488 Patent and 8:45–49; 9:67–10:1.
Moreover, Defendants point to statements made by the patentee during the file history to
support their construction. Defendants argue that during reexamination, the patentee
distinguished prior art that allowed for the storage of packets before being transmitted.
Specifically, the prior art used a “sliding window implementation” in a single “transmit buffer”
Page 41 of 60
holding both “packets that have been transmitted, and packets that have yet to be transmitted”
(Doc. No. 511-30 at 18–19 (Am. in Consol. Reexam.) (emphasis added)). To distinguish this
prior art, the patentee argued to the USPTO that this was different from the ’488 Patent because
“the transmit interface disclosed by the ’488 Patent makes use of two buffers[:] a Request Queue
for holding packets for transmission, as well as an Echo Waiting Queue that holds only those
packets that have already been transmitted by the Request Queue.” (Doc. No. 511-30 at 18–19
(Am. in Consol. Reexam.) (emphasis added)). The Court agrees that Defendants’ construction is
consistent with the statements made to the USPTO to distinguish the invention from the prior art.
Finally, during the claim construction hearing, the Court proposed its construction to the
parties. Defendants stated they could agree with the Court’s construction if the Court were to
adopt their proposed construction for the disputed “first buffer” phrase. For the reasons stated
above, the Court finds that its construction for this disputed phrase, in conjunction with the plain
and ordinary meaning of the disputed “first buffer” phrase, properly addresses the issues raised
by the parties and does not require the Court to construe the “first buffer” phrase.
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
adopts a modified version of Defendants’ proposed construction and construes “a second buffer
for storing a copy of a packet transmitted” to mean “a second buffer, separate from the
first buffer, for storing copies of packets that have already been transmitted from the first
buffer.”
Page 42 of 60
i. “cache”
Disputed Claim Term
Plaintiff’s Proposed
Construction
Ordinary meaning, in
the alternative: “a circuit for
temporarily storing data”
“cache”
Defendants’ Proposed
Construction
“a section of fast memory
that stores frequently used
data until it is full and then
replaces old data with more
recently requested data”
The parties’ dispute is focused on three issues. The first issue is whether the cache must
be “fast” memory. Plaintiff contends that “fast” memory introduces unnecessary ambiguity into
the claim. The Court agrees and finds that the claim language itself provides for Defendants’
“fast” memory construction. The second issue is whether the cache is limited to storing only
“frequently used data.” Plaintiff contends that not all caches are limited to storing only
“frequently used data.” The Court agrees with this statement, but Defendants’ construction is
not limited to only storing frequently used data. It also allows for storing other data (e.g., prefetched data). The third issue is whether the cache must “store frequently used data until it is
full.” Plaintiff contends that there are different types of caches, and requiring the cache to “store
frequently used data until it is full” would exclude one cache type described in the
specification. The Court agrees. Accordingly, the Court provides a construction that is
consistent with the claim language, the specification, and the prosecution history.
i.
The Claim Language
The term “cache” appears in claims 32 and 35. Having reviewed the claims, the Court
concludes that the term is used consistently throughout the claims and is meant to have the same
meaning. Claim 35 specifies that the cache is included in the microprocessor interface, stores a
quantity of data requested from one of the devices, and is organized with a variable line size
Page 43 of 60
adjusted dynamically based on changes in hit rate. Thus, the claim language captures the “fast”
memory aspect of Defendants’ construction because the cache is included in the microprocessor
interface itself. Similarly, the claim language supports a construction that includes storing
frequently used data because the claim requires the cache to store data requested from one of the
devices. Finally, the claim language is silent as to whether the cache must store frequently used
data until it is full. Accordingly, the Court finds that its construction is consistent with the
plain and ordinary meaning of the claim language as understood by one of ordinary skill in the
art. The Court will now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
In addition to the claims, the other intrinsic evidence in this case supports the Court’s
construction. First, Defendants contend that Plaintiff’s construction is so broad that it would
encompass almost any form of memory, including the “main memory,” “buffers,” and “mass
storage device” that the ’488 Patent identifies as separate elements. The Court agrees and finds
that Plaintiff’s construction would not help the jury understand the differences between the
claimed cache and the other claimed devices. Indeed, during the claim construction hearing,
Plaintiff conceded that its proposed construction is broad and would read on other claimed
memory devices (e.g., random access memory device, mass storage device). Although the
Court agrees with Defendants’ argument, it does not adopt Defendants’ construction because it
goes too far.
First, Defendants argue that a cache must “store frequently used data until it is full.”
The parties do not dispute that this is a characteristic of a particular type of cache that is
commonly referred to as “fully associative” (see, e.g., Doc. No. 504-2 at 25, ¶64 (Dec. of
William Mangione-Smith)). Fully associative caches allow data to be stored at any location
Page 44 of 60
within the cache (Doc. No. 504-2 at 25, ¶64 (Dec. of William Mangione-Smith)). In contrast,
“direct-mapped” or “set-associative” caches restrict where a particular piece of data can reside
within a cache and typically require cache management before the cache is full (Doc. No. 5042 at 25, ¶64 (Dec. of William Mangione-Smith)). Thus, Defendants’ construction is not
consistent with a direct-mapped cache, which is described as an embodiment of the linc cache.
Specifically, the specification states that the “[l]inc cache 63 is basically a directed [sic] mapped
cache for caching response data for the functional unit.” ’488 Patent at 8:63–64. Accordingly,
the Court will not include Defendants’ “until it is full” language in its construction because it is
not consistent with a “direct-mapped” cache.
Defendants also contend that it is well document that a cache stores data until it is full
and replaces old data with more recently requested data. However, none of the definitions for
cache included in the extrinsic evidence provided by the parties, or even the definition quoted by
Defendants in their brief, specify that the cache must “store data until it is full and replaces old
data with more recently requested data.” Although the Court recognizes that a cache must have
some mechanism for replacing data, the Court does not find that Defendants’ proposed language
is necessary for its construction. Moreover, Defendants point to no intrinsic evidence that would
compel the Court to include this language in its construction. The Court will now turn to the
extrinsic evidence submitted by the parties.
iii.
Extrinsic Evidence
In addition to the intrinsic evidence, there is also extrinsic evidence that supports the
Court’s construction. First, the Court finds that the definitions for cache provided in the
extrinsic evidence commonly references frequently used data (see, e.g., Doc. No. 511-6 at 3
(Dictionary of Computer Words 34 (rev. ed. 1995)); Doc. No. 511-8 at 3 (Prentice Hall’s
Page 45 of 60
Illustrated Dictionary of Computing 73 (2d ed. 1995))). Plaintiff’s basis for opposing this
language is that not all caches are limited to storing only “frequently used data.” To be clear, the
Court’s construction does not require that the cache store only frequently used data, but also
allows the cache to store other data (e.g., pre-fetched data). See, e.g., ’488 Patent, 11:15–20.
Accordingly, in order to help the jury understand the difference between a “cache” and the other
claimed elements (e.g., “main memory,” “buffers,” “mass storage device”), the Court’s
construction includes “frequently used data.”
This conclusion is also consistent with the testimony of Plaintiff’s expert, Mr. William
Henry Mangione-Smith. Mr. Mangione-Smith agreed that it would be fair to characterize the
cache’s intent as storing frequently used data. Doc. No. 511-19 at 13 (Mangione-Smith Tr.
172:20–23). Accordingly, the Court’s construction that requires the cache to store “frequently
used data” is consistent with the extrinsic evidence submitted by the parties.
iv.
Court’s Construction
In light of the claim language, the specification, the prosecution history, and the
extrinsic evidence submitted by the parties, the Court construes “cache” to mean “a section of
memory that temporarily stores frequently used data.”
j. “linc cache”
Disputed Claim Term
Plaintiff’s Proposed
Construction
Ordinary meaning in the
context of the claim
language, in the alternative:
“a circuit that contains a first
memory block for storing
recently received data and a
second memory block for
storing additional data related
to said received data”
“linc cache”
Page 46 of 60
Defendants’ Proposed
Construction
“a single link-layer cache”
The parties dispute whether “linc cache” should be construed as “a single link-layer
cache.” Plaintiff contends that Defendants’ construction introduces unnecessary ambiguity to
the claim and could improperly limit the claim to exclude systems using more than one “linc
cache.” Defendants respond that the patentee made clear and unequivocal statements to the
USPTO that a “linc cache” is a single link-layer cache. The Court agrees with Defendants and
finds that their proposed construction is consistent with the claim language, the specification,
and the prosecution history.
i.
The Claim Language
The phrase “linc cache” appears in claim 32. Having reviewed the claim, the Court
concludes that the phrase is used consistently throughout the claim. Plaintiff contends that the
remainder of the claim language sufficiently and unambiguously defines the meaning of the
phrase “linc cache.” Specifically, Plaintiff argues that the claim language describes the “linc
cache” as follows: (1) included in the microprocessor interface; (2) used for storing data
requested by the microprocessor from at least one of said devices (i.e., random access memory
device, mass storage device and/or input-output port device); and (3) made up of at least two
memory blocks—one block for storing at least data most recently requested (by the
microprocessor) and one block for storing data not most recently requested. The Court agrees
with Plaintiff’s summary of the claim elements but finds that it does not take into account the
statements made to the USPTO that describe the linc cache as a single link-layer cache.
Moreover, Plaintiff’s proposed construction does not provide any guidance to the jury because it
only repeats the elements of the claim. The Court will now turn to the other intrinsic evidence.
Page 47 of 60
ii.
The Intrinsic Evidence
The other intrinsic evidence in this case, specifically the prosecution history, supports the
adoption of Defendants’ construction. Defendants contend that during the reexamination, the
patentee explicitly defined the “linc cache” to mean a single link-layer cache. Specifically,
Defendants point to the following statement:
The ’488 linc cache defines a hardware-controlled data cache that is
incorporated at the link layer in each node of a packet-based computer
bus. This hardware cache implementation is a non-obvious design choice
and represents a valid patent claim.
(Doc. No. 511-5 (Gardner Decl. ¶ 50) (emphasis added)). Using this statement, the patentee then
distinguished three prior art references on the ground that they did not disclose a link-layer
cache. Specifically, Defendants point to the following arguments made by the patentee to the
USPTO:
1) “Yoshizawa’s “link cache” . . . does not teach the structural and
functional limitations of the link-layer cache in ’ 488” (Doc. No.
511-5 (Gardner Decl. at ¶ 63) (emphasis added)).
2) “A local memory cache is at a higher layer in the OSI
communications model and is fundamentally different from cache at
the link layer of the communications channel” (Doc. No. 511-5
(Gardner Decl. at ¶ 57) (emphasis added)).
3) “Since Yoshizawa . . . does not use packets, does not operate as a bus
interface, and does not operate at the link layer of a communications
system, a person of ordinary skill in the art would not look to
Yoshizawa . . . to arrive at the linc cache invention of ’488” (Doc.
No. 511-5 (Gardner Decl. at ¶ 64) (emphasis added)).
4) “A software disk cache like Holzhammer’s has to operate at a
communications layer higher than the link layer implementation of the
’488 patent” (Doc. No. 511-5 (Gardner Decl. at ¶ 74) (emphasis
added)).
Similarly, Defendants point to the patentee’s repeated statements to the USPTO that its
“linc cache” is a single cache, not a combination of two separate caches (see Doc. No. 511-10 at
75 (Patent Owner’s Appeal Brief (Feb. 8, 2011)) (“[T]he specification of the ’488 Patent, in
Page 48 of 60
every instance, depicts and/or discusses the link cache of the claimed invention as single cache,
and not as two separate caches.”); Doc. No. 511-12 at 25 (Resp. in ’376 Reexam. (Nov. 11,
2009)); Doc. No. 511-13 at 51–52 (Resp. in ’376 Reexam. (May 14, 2010)); Doc. No. 511-34
(Belgard Decl. ¶¶ 14, 15, 38); Doc. No. 511-14 at 42–43 (Resp. in ’376 Reexam. (July 26,
2010)); Doc. No. 511-3 (Reinman Decl. ¶¶ 216–18)).
Plaintiff contends that the USPTO expressly rejected these characterization as not
supported by either the specification or the claim language. This argument is without merit
because the patentee representation is binding regardless of whether the USPTO accepted the
patentee’s argument. Am. Piledriving Equip., Inc. v. Geoquip, Inc., 637 F.3d 1324, 1336 (Fed.
Cir. 2011) (finding disclaimer based on statements with which “examiner explicitly disagreed”);
Lifestream Diagnostics, Inc. v. Polymer Tech. Sys., Inc., 109 F. App’x 411, 414–16 (Fed. Cir.
2004) (“We agree with the district court that arguments deliberately and repeatedly advanced by
the patent applicant in regard to the scope of a claim term during prosecution may be used for
purposes of claim construction even though the Patent Office rejected the arguments.”).
Accordingly, the Court agrees with Defendants and finds that the patentee defined the “linc
cache” to mean “a single link-layer cache” given the repeated statements made to the USPTO
during the reexaminations.
Finally, Plaintiff’s argument that this construction could improperly limit the claim to
exclude systems using more than one “linc cache” is unfounded. Defendants’ construction only
requires that the linc cache must be a single cache and not the combination of two or more
caches. In other words, while the system may have multiple linc caches, there must be at least
one linc cache that contains the two required blocks in a single cache. Accordingly, the Court
adopts Defendants’ construction.
Page 49 of 60
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
construes a “linc cache” to mean “a single link-layer cache.”
k. “said linc cache comprising first and second memory blocks”
Disputed Claim Term
“said linc cache comprising
first and second memory
blocks”
Plaintiff’s Proposed
Defendants’ Proposed
Construction
Construction
Ordinary meaning in the
“the linc cache contains two
context of the claim language, blocks with fixed locations,
in the alternative: “a circuit
each of which has identical
that contains a first memory
functionality but stores
block for storing recently
different data”
received data and a second
memory block for storing
additional data related to said
received data”
The parties’ dispute is focused on two issues. The first issue is whether the “first and
second memory blocks” of the linc cache must have fixed locations. The second issue is whether
the “first and second memory blocks” of the linc cache must have identical functionality.
Regarding the first issue related to the fixed locations, Defendants argue that the claim language
dictates that the locations of the first and second memory blocks of the linc cache are fixed. The
Court is not persuaded by Defendants’ argument because their construction is based on pure
speculation that the “patentee must have meant the two-block limitation to mean something
other than what every cache already has” (Doc. No. 511 at 28). Accordingly, the Court rejects
Defendants’ argument because the addition of the “fixed locations” limitation is not supported
by the claim language, the specification, or the prosecution history.
Regarding the second issue related to identical functionality, Defendants contend that the
intrinsic evidence requires that the first and second memory blocks have identical caching
Page 50 of 60
functionality. The Court agrees with Defendants on this issue and adopts a modified version of
Defendants’ construction.
i.
The Claim Language
The phrase “said linc cache comprising first and second memory blocks” appears in
claim 32. Having reviewed the claim, the Court concludes that the phrase is used consistently
throughout the claim. As with the term “linc cache,” Plaintiff contends that the remainder of the
claim language sufficiently and unambiguously defines the meaning of this phrase and that no
separate definition is required for this phrase. The Court agrees that the claim further defines the
disputed phrase but finds that this alone would not capture the statements made to the USPTO
that further define the disputed phrase. Moreover, Plaintiff’s proposed construction does not
provide any guidance to the jury because it merely repeats the elements of the claim. The Court
will now turn to the other intrinsic evidence.
ii.
The Intrinsic Evidence
Defendants contend that the intrinsic evidence requires that the first and second
memory blocks have identical caching functionality (although they may store different data).2
Specifically, Defendants point to the portion of the specification that states that “the Linc
Cache 63 is partitioned into two identical blocks.” ’488 Patent at 8:63–9:3. Defendants further
point to arguments made by the patentee in the prosecution history. During the reexamination,
the patentee argued that a prior art reference was “different from the dual-block design of
’488,” because the ’488 Patent’s linc cache “divides a cache into two parts with identical
functionality” (Doc. No. 511-5 (Gardner Decl. ¶ 80)). Plaintiff responds that Defendants seek
2
During the claim construction hearing, Defendants agreed that their proposed language of “but
stores different data” was not necessary for the construction of this disputed phrase.
Page 51 of 60
to impose additional unstated limitations into the claim and Defendants’ additional limitations
are not only unnecessary but also introduce ambiguity into the claim. The Court disagrees.
Accordingly, the Court concludes that the patentee defined the “first and second memory blocks”
as “two blocks with identical functionality” as indicated in the intrinsic record.
iii.
Court’s Construction
In light of the claim language, the specification, and the prosecution history, the Court
construes “said linc cache comprising first and second memory blocks” to mean “the linc
cache contains two blocks with identical functionality.”
l.
“said first memory block storing a first portion of said stored data
comprising the data most recently requested”3
Disputed Claim Term
“said first memory block
storing a first portion of said
stored data comprising the
data most recently requested”
Plaintiff’s Proposed
Defendants’ Proposed
Construction
Construction
“the first memory block
Ordinary meaning, in the
always stores the data most
alternative:
“a first memory block capable recently requested”
of storing the data most
recently requested”
The parties dispute whether the linc cache’s first memory block must “always” store the
data most recently requested. Plaintiff contends that the claim language is unambiguous in its
meaning. Defendants respond that the intrinsic evidence uniformly describes the first memory
block as storing the data most recently requested. The Court agrees with Defendants’ argument
but finds that this is captured by the plain and ordinary meaning of the disputed phrase. The
phrase requires that the first memory block store a first portion of the stored data “comprising the
3
The phrase “said first memory block storing a first portion of said stored data comprising the
data most recently requested” appears in claim 32.
Page 52 of 60
data most recently requested.”
Defendants also contend that their “always stores” construction is compelled by the
patentee’s statement to the USPTO. During reexamination, the patentee argued that the link
cache’s dual-block design allowed simultaneous access to both blocks, “with one block always
having the latest data and the other block being one access behind” (Doc. No. 511-5 (Gardner
Decl. ¶ 80) (emphasis added)). The Court has reviewed the claims in light of this intrinsic
evidence and concludes that the claim language faithfully captures the statements made by the
patentee to the USPTO. Moreover, the Court concludes that the disputed phrase is unambiguous,
is easily understandable by a jury, and requires no construction. Therefore, the phrase “said
first memory block storing a first portion of said stored data comprising the data most recently
requested” will be given its plain meaning to one of ordinary skill in the art.
m. “the balance of said stored data”
Disputed Claim Term
“the balance of said stored
data”
Plaintiff’s Proposed
Construction
Ordinary meaning, in the
alternative: “additional data
related to the received data”
Defendants’ Proposed
Construction
“all of the stored data except
the data most recently
requested”
The parties dispute whether the term “balance” should be replaced with “all of the stored
data except the data most recently requested.”4 Plaintiff contends that the claim language is clear
enough that the ordinary meaning should be applied. Defendants respond that Plaintiff’s
alternative construction ignores the word “balance” and could allow the second memory block to
store some “stored data” but not necessarily the “balance of said stored data.” Defendants’
argument is unfounded because Plaintiff’s primary position is for the plain meaning of the
4
The phrase “the balance of said stored data” appears in claim 32.
Page 53 of 60
disputed phrase, which includes the word “balance.” Defendants fail to point to any evidence
that would justify significantly redrafting one term in the disputed phrase. Moreover, the Court
concludes that the disputed phrase is unambiguous, is easily understandable by a jury, and
requires no construction. Therefore, the phrase “the balance of said stored data” will be given its
plain meaning to one of ordinary skill in the art.
n. Agreed Constructions
In addition to the above disputed terms, the parties have come to several agreed
constructions. Based upon the joint submission of a claim construction chart, the parties agree to
the construction of the following terms. These constructions are therefore adopted by the Court:
Claim Term/Phrase
“buffer”
Location
14, 17, 27,
28, 29, 31,
33, 34, 35
Agreed Construction
“a storage element used to store data
temporarily when sending data from one
functional unit to another”
“parallel data/address bus
circuit”
20
“a circuit having multiple address, data and
control lines”
“switching means”
24
Function: selectively connecting said
microprocessor to selected ones of said
devices
Structure: a ring and a switch and/or
crossbar interconnect (’488 Patent, Figs.
24, 25; Col. 7, Lines 36–63 at
“a quantity of data
requested from one of
said devices”
32, 35
“a quantity of data requested by the
microprocessor from one of the devices”
“physical address”
33, 34
“an address corresponding to a physical
memory location or a particular functional
unit”
“virtual address”
34
“a non-physical memory address that can be
converted to a corresponding physical
address to access data in physical memory”
Page 54 of 60
“translation lookaside
buffer"
34
“a memory that stores the physical address
associated with each virtual address”
“hit rate”
35
“a measure of how often data requested by
the microprocessor is in the cache”
“said cache being organized
with a variable line size”
35
“the amount of data corresponding to a
cache tag is variable”
“said cache being organized
with a variable line size
adjusted dynamically based
on changes in hit rate”
35
“the amount of data corresponding to a
cache tag is variable, and is adjusted
dynamically based on changes in hit rate”
a. “for providing simultaneous bi-directional communication
between said microprocessor and said devices”5
Prior to the claim construction hearing, the parties disputed whether packet-based data
channel must provide simultaneous bi-directional communication between the microprocessor
and all of the devices, or whether it needs to provide communication between the microprocessor
and only one of the claimed devices. The Court reviewed the intrinsic evidence and concluded
that Defendants’ proposed construction was ambiguous on its face and that Plaintiff’s proposed
construction was broader than the scope of the claims.
As with the disputed phrase “said microprocessor and said devices each including an
interface for receiving and transmitting data in packet form,” the Court concluded that the bidirectional communication must be capable of extending from a microprocessor to each interface
because every disclosed embodiment illustrates a data path channel extending to each device.
See ’488 Patent at Figs. 2, 24, 25. However, the Court also concluded that the claims do not
require that the simultaneous bi-directional communication between the microprocessor and each
5
The phrase “for providing simultaneous bi-directional communication between said
microprocessor and said devices” appears in claims 14, 27-29, and 31-35.
Page 55 of 60
of the multiple devices must occur at the same time. For example, Figure 24 illustrates a ring
interconnect where “each link interface's physical link output is connected to a neighboring
nodes link interface physical input until the ring is closed.” ’488 Patent at 7:37–40, Fig. 24. This
embodiment confirms that while the communication between the microprocessor and any one
device must be simultaneous bi-directional communication, the simultaneous bi-directional
communication between the microprocessor and more than one device may occur at the same
time or different times. In other words, there would be no need to circulate or forward the packet
from one interface in the ring to another interface in the ring if the claims required the bidirectional communication to occur at the same time to each device.
Thus, during the claim construction hearing, the Court proposed a construction to the
parties that is consistent with the claim language, the specification, and the prosecution
history. The parties agreed with the construction presented by the Court. Accordingly, the
Court adopts the parties’ agreed construction and construes “for providing simultaneous bidirectional communication between said microprocessor and said devices” to mean “for
providing simultaneous bi-directional communication between the microprocessor and each of
the devices.”
V.
CONCLUSION
For the forgoing reasons, the Court adopts the constructions as set forth above, and as
listed in the attached chart. The parties are ordered that they may not refer, directly or indirectly,
to each other’s claim construction positions in the presence of the jury. Likewise, the parties are
ordered to refrain from mentioning any portion of this opinion, other than the actual definitions
adopted by the Court, in the presence of the jury. Any reference to claim construction
Page 56 of 60
proceedings is limited to informing the jury of the definitions adopted by the Court.
It is SO ORDERED.
SIGNED this 7th day of June, 2013.
____________________________________
MICHAEL H. SCHNEIDER
UNITED STATES DISTRICT JUDGE
Page 57 of 60
APPENDIX A
Claim Term/Phrase
“microprocessor”
Location
14, 24, 2729, 31-35,
37
14, 27-29,
31-35
“[a] physically nondistributed microprocessorbased computer system”
Court’s Construction
“a central processing unit residing within a
single chip”
“a microprocessor-based computer system in
which all the components are close to one
another and contained within a single
enclosure”
“said microprocessor and
said devices each including
an interface for receiving
and transmitting data in
packet form”
14, 27-29,
31-35
“each of the microprocessor, the random
access memory device, the mass storage
device, and the input-output port device has
its own interface for receiving and
transmitting data in packet form”
“interface”
14, 16-17,
20-21, 2729, 31-35,
37
“a circuit that links a packet-based data
channel to either a microprocessor or a
device and that speaks the same language
electrically and logically with each other
such circuit”
“a packet-based data channel
extending between said
microprocessor and said
interfaces of said devices”
14, 27-29,
31-35
“a data channel capable of transporting
uniform packet types between a
microprocessor interface and each interface
of the devices”
“packet-based data channel”
14, 20, 23,
27-29, 3135, 39
“a data channel capable of transporting a
packet”
“packet”
14-17, 20,
23, 27-35,
39
“a message including payload data and the
control data indicative of the intended
recipient”
“a first buffer for queuing
packets for a transmission”
14, 27-29,
31, 33-35
No construction necessary
“a second buffer for storing
a copy of a packet
transmitted”
14, 27-29,
31, 33-35
“a second buffer, separate from the first
buffer, for storing copies of packets that have
already been transmitted from the first
buffer”
Page 58 of 60
“cache”
32, 35
“A section of memory that temporarily stores
frequently used data”
“linc cache”
32
“a single link-layer cache”
“said linc cache comprising
first and second memory
blocks”
32
“the linc cache contains two blocks with
identical functionality”
“said first memory block
storing a first portion of said
stored data comprising the
data most recently
requested”
32
No construction necessary
“the balance of said stored
data”
32
No construction necessary
“for providing simultaneous
bi-directional
communication between
said microprocessor and said
devices”
14, 27-29,
31-35
[Agreed] “for providing simultaneous bidirectional communication between the
microprocessor and each of the devices”
“buffer”
14, 17, 27,
28, 29, 31,
33, 34, 35
[Agreed] “a storage element used to store
data temporarily when sending data from one
functional unit to another”
“parallel data/address bus
circuit”
20
[Agreed] “a circuit having multiple address,
data and control lines”
“switching means”
24
[Agreed] Function: selectively connecting
said microprocessor to selected ones of said
devices
Structure: a ring and a switch and/or
crossbar interconnect (’488 Patent at Figs.
24, 25; Col. 7, Lines 36-63)
32, 35
“a quantity of data
requested from one of
said devices”
[Agreed] “a quantity of data requested by
the microprocessor from one of the devices”
Page 59 of 60
“physical address”
33, 34
[Agreed] “an address corresponding to a
physical memory location or a particular
functional unit”
“virtual address”
34
[Agreed] “a non-physical memory address
that can be converted to a corresponding
physical address to access data in physical
memory”
“translation lookaside
buffer"
34
[Agreed] “a memory that stores the physical
address associated with each virtual address”
“hit rate”
35
[Agreed] “a measure of how often data
requested by the microprocessor is in the
cache”
“said cache being organized
with a variable line size”
35
[Agreed] “the amount of data corresponding
to a cache tag is variable”
“said cache being organized
with a variable line size
adjusted dynamically based
on changes in hit rate”
35
[Agreed] “the amount of data corresponding
to a cache tag is variable, and is adjusted
dynamically based on changes in hit rate”
Page 60 of 60
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?