1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS WILLIAM F. LEE (admitted pro hac vice) (william.lee@wilmerhale.com) DOMINIC E. MASSA (admitted pro hac vice) (dominic.massa@wilmerhale.com) JOSEPH J. MUELLER (admitted pro hac vice) (joseph.mueller@wilmerhale.com) LOUIS W. TOMPROS (admitted pro hac vice) (louis.tompros@wilmerhale.com) WILMER CUTLER PICKERING HALE AND DORR LLP 60 State Street Boston, MA 02109 Telephone: (617) 526-6000 Facsimile: (617) 526-5000 Attorneys for Plaintiff BROADCOM CORPORATION UNITED STATES DISTRICT COURT CENTRAL DISTRICT OF CALIFORNIA BROADCOM CORPORATION, Plaintiff, v. EMULEX CORPORATION, Defendant. CASE No. SACV09-1058 JVS (ANx) consolidated SACV 10-03963-JVS (ANx) PLAINTIFF BROADCOM CORPORATION’S REPLY BRIEF IN SUPPORT OF ITS MOTION FOR JUDGMENT AS A MATTER OF LAW OF INFRINGEMENT OF CLAIM 1 OF THE ’194 PATENT (BE3 AND QT2025), AND CLAIM 5 OF THE ’124 PATENT (SOC442) Before: Hon. James V. Selna Date: December 12, 2011 Time: 3:00 p.m. Courtroom: 10C And Related Counterclaims Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 1 of 21 Page ID #:25066 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 i BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS TABLE OF CONTENTS TABLE OF AUTHORITIES .........................................................................ii I. INTRODUCTION..........................................................................................1 II. ARGUMENT .................................................................................................3 A. The BE3 and QT2025 Infringe Claim 1 as a Matter of Law.......................................................................................................3 1. Emulex-Not Broadcom-Is Seeking a New and Different Post-Trial Claim Construction. ..................................4 2. Claim 1 Does Not Require a C3MOS Demultiplexer. ...........................................................................7 3. Emulex Admits that the BE3 and QT2025 Include C3MOS and Deserialize...........................................................12 B. The SOC442 Infringes Claim 5 as a Matter of Law..........................13 III. CONCLUSION ............................................................................................17 Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 2 of 21 Page ID #:25067 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 ii BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS TABLE OF AUTHORITIES Cases Conoco, Inc. v. Energy & Environmental International, L.C., 460 F.3d 1349 (Fed. Cir. 2006) ......................................................................5, 6 Finjan, Inc. v. Secure Computing Corp., 626 F.3d 1197 (Fed. Cir. 2010) ......................................................................4, 6 Geo. M. Martin Co. v. Alliance Machine Sys. International LLC, 618 F.3d 1294 (Fed. Cir. 2010) ..............................................................4, 12, 14 Kara Tech. Inc. v. Stamps.com Inc., 582 F.3d 1341 (Fed. Cir. 2009) ........................................................................10 Lazare Kaplan Int’l, Inc. v. Photoscribe Technologies, Inc., 628 F.3d 1359 (Fed. Cir. 2010) ........................................................................11 Sitrick v. Dreamworks, LLC, 516 F.3d 993 (Fed. Cir. 2008) ..........................................................................14 Teleflex, Inc. v. Ficosa North America Corp., 299 F.3d 1313 (Fed. Cir. 2002) ........................................................................13 Upjohn Co. v. Mova Pharmaceutical Corp., 225 F.3d 1306 (Fed. Cir. 2000) ....................................................................3, 14 Verizon Servs. Corp. v. Cox Fibernet Virginia, Inc., 602 F.3d 1325 (Fed. Cir. 2010) ..........................................................................5 Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 3 of 21 Page ID #:25068 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 1 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS I. INTRODUCTION Like its defense at trial, Emulex’s opposition to JMOL rests on the premise that claim 1 of U.S. Patent No. 6,424,194 (the “’194 patent”) requires that the claimed C3MOS circuitry be “used for deserialization.” (Emulex’s Opposition to Broadcom’s Motion for Judgment as a Matter of Law Regarding Infringement of Claim 1 of the ’194 Patent (BE3 and QT2025) and Claim 5 of the ’124 patent (SOC442), Dkt. 973, Nov. 23, 2011 (“Emulex Op.”), at 3, 16.) Specifically, Emulex argues that the C3MOS circuits must themselves be used to divide down the frequency of a signal to satisfy claim 1. (See id. at 11 n.7 (equating “divid[ing] down” frequency with “deserializ[ing]).) In other words, Emulex’s defense rests on the premise that the C3MOS circuits must themselves be the circuits in the deserializer that divide down the frequency of the signals-i.e., “demultiplex” the signals. But the invention of the ’194 is not so limited. As the specification makes clear, C3MOS can be used for a variety of purposes in high-speed deserializers- including “divid[ing] down the frequency of the signals,” but also including to “interfac[ing] with and process[ing] the high speed signals” before that “dividing down” occurs. ’194 patent col.7 ll.6-12. Part of the genius of the invention is that power-hungry C3MOS can be used only to the extent necessary, in the particular parts of the deserializer where it is needed to handle a high-speed signal. (See, e.g., Emulex Op. at 9 (conceding that C3MOS is used when “vital”); Trial Tr. Sept. 22, 2011, at 33:15 (testimony of inventor Dr. Armond Hairapetian that C3MOS is used only “where it [i]s essential.”).) Emulex argues that the invention is somehow limited to the specific configuration in which C3MOS is used to divide down the frequency of the signal, but this mischaracterizes the invention and is not what claim 1 requires. Rather, claim 1 requires only that the first circuitry contain C3MOS, and that the first Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 4 of 21 Page ID #:25069 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 2 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS circuitry be “configured to deserialize.” (’194 patent col.9 ll.11-13.) There is no dispute that the circuitry identified by Broadcom’s expert Dr. Fair as the claimed “first circuitry” in the BE3 and the QT2025 products contains C3MOS, and there is no dispute that the first circuitry is configured to deserialize. (See, e.g., Trial Tr. Sept. 30, 2011, at 46:4-10.) Because no reasonable jury could find otherwise, Broadcom is entitled to JMOL. Emulex attempts to avoid JMOL by suggesting that Broadcom is somehow offering a “brand new claim construction.” But Emulex-not Broadcom-is asking the Court to construe claim 1 post trial. In its opposition, Emulex specifically requests that claim 1 be “construed to require that at least a ‘portion’ of the deserialization used C3MOS circuitry”-in other words, that the C3MOS circuits are used to divide down the frequency. (Emulex Op. at 16; see also id. at 11 n.7.) It is improper for Emulex to offer a new limiting construction post-trial, and Emulex’s construction is contrary to the claim language as well as the specific portion of the specification that Broadcom cited in its opening brief, but Emulex ignored in its opposition. To support its new construction, Emulex relies heavily on Dr. Fair’s testimony that because the claim says “C3MOS” and “is configured to deserialize,” the infringing circuitry must include “C3MOS” and be “configured to deserialize.” (Emulex Op. at 2, 10 (citing Trial Tr. Sept. 23, 2011, at 38:12-39:2).) But that is not what is actually in dispute. Neither Broadcom nor Dr. Fair has ever disputed that, to infringe, a product must include C3MOS, and must be configured to deserialize. Dr. Fair did, however, unequivocally reject Emulex’s view that the infringing C3MOS circuits had to be the circuits in the deserializer that were performing demultiplexing (i.e., “dividing down the frequency”). (See, e.g., Trial Tr. Sept. 22, 2011, at 139:8-15; id. at 157:5-6.) Because Emulex’s construction finds no support in either the intrinsic record or the testimony at trial, the Court Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 5 of 21 Page ID #:25070 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 3 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS should reject it and grant Broadcom’s motion for JMOL. Emulex’s opposition to JMOL of infringement of claim 5 of U.S. Patent No. 7,486,124 (the “’124 patent”) is equally flawed. Emulex relies solely on a purported factual dispute resulting from Emulex’s expert’s conclusory, impeached, and counterfactual testimony that the signal pair Q1P and Q1N is never “output” from the circuit in which it is generated. But an expert’s conclusory opinion, unsupported by the factual record, does not defeat JMOL. See, e.g., Upjohn Co. v. Mova Pharm. Corp., 225 F.3d 1306, 1311 (Fed. Cir. 2000) (“[T]here must be factual support for an expert’s conclusory opinion”). Based on the record as a whole, Broadcom’s motion for JMOL that the SOC442 infringes claim 5 should be granted. II. ARGUMENT A. The BE3 and QT2025 Infringe Claim 1 as a Matter of Law. There is no dispute that the BE3 and QT2025 use C3MOS, and that they do so “for a reason.” (Trial Tr. Sept. 30, 2011, at 46:4-10.) As Emulex itself concedes, the use of C3MOS can be “vital” in high-speed applications, including high-speed SerDes. (Emulex Op. at 9.) But, because C3MOS requires substantial power, it is only used in SerDes circuitry in the particular locations “where it [i]s essential.” (Trial Tr. Sept. 22, 2011, at 33:15.) Emulex contends that the use of C3MOS circuitry described in claim 1 has only a limited purpose-to divide down the frequency of a signal. (Emulex Op. at 3, 11 n.7.) But Emulex’s proposed limiting construction comes too late, and, in any event, is unsupported by the claim language, specification, prosecution history, and testimony at trial. Once Emulex’s construction is rejected, “the evidence, construed in the light most favorable to [Emulex], permits only one reasonable conclusion”-that the BE3 and QT2025 infringe claim 1. Geo. M. Martin Co. v. Alliance Machine Sys. Int’l LLC, 618 F.3d 1294, 1300 (Fed. Cir. 2010) (quoting Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 6 of 21 Page ID #:25071 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 4 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS White v. Ford Motor Co., 312 F.3d 998, 1010 (9th Cir. 2002)). 1. Emulex-Not Broadcom-Is Seeking a New and Different Post-Trial Claim Construction. Before trial, Emulex never sought the limiting construction of claim 1 that it now advocates. The Court construed six terms in the claims of the ’194 patent, and held that all other terms “shall have their plain and ordinary meaning.” (Glossary, Dkt. No. 899, Sept. 21, 2011, at 3-4); see also, e.g., Finjan, Inc. v. Secure Computing Corp., 626 F.3d 1197, 1207 (Fed. Cir. 2010) (approving use of plain and ordinary meaning when parties do not dispute meaning of claim term). Neither the “wherein the first circuitry” limitation, nor any of the component parts “first circuitry,” “input circuit,” or “configured to deserialize” was among the construed terms. (See id.) Contrary to Emulex’s assertions, Broadcom’s motion for judgment as a matter of law does not “demand[] a new claim construction.” (Emulex Op. at 4.) Broadcom’s motion does not advocate for any new claim construction-in fact, it does not mention claim construction at all. Rather, Broadcom’s motion is premised solely on the claim language, in view of the concessions that Emulex made at trial that the identified circuitry in the BE3 and QT2025 is C3MOS. (See Plaintiff Broadcom Corporation’s Memorandum in Support of Its Motion for Judgment a Matter of Law of Infringement of Claim 1 of the ’194 Patent (BE3 and QT2025) and Claim 5 of the ’124 Patent (SOC442), Dkt. No. 961, at 18 (“Broadcom Mem.”).) Emulex-not Broadcom-is now seeking to apply a new and narrower claim construction of claim 1. Emulex admits expressly in its brief that its primary argument applies only “when claim 1 of the ’194 Patent is construed to require that at least a ‘portion’ of the deserialization used C3MOS circuitry.” (Emulex Op. at 16 (emphasis added).) But as Emulex itself argues, seeking a new claim Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 7 of 21 Page ID #:25072 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 5 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS construction post trial is improper. See, e.g., Conoco, Inc. v. Energy & Environmental Int’l, L.C., 460 F.3d 1349, 1359 (Fed. Cir. 2006) (“[L]itigants waive their right to present new claim construction disputes if they are raised for the first time after trial.”); (see also Emulex Op. at 1, 4-5). Moreover, Emulex’s new position that claim 1 should be construed to require that the claimed C3MOS circuits be “used for deserialization” (by which it means used to divide down frequency) is inconsistent with its earlier position that the “input circuit” limitation did not require construction. At the outset of the claim construction process, Emulex submitted its list of proposed claim constructions pursuant to Northern District of California Patent Local Rule 4-2. (See Ex. A, Emulex Corporation’s Preliminary Claim Constructions and Extrinsic Evidence, June 18, 2010.)1 In that disclosure, Emulex proposed that the term “[i]nput circuit” be construed to mean “A C3MOS circuit configured to deserialize the first signal.” (Id. at 19.) However, three months later, Emulex reversed course and abandoned its claim that “input circuit” required construction. On September 10, 2010, Emulex filed amended proposed claim constructions that made no mention of any construction of “input circuit.” (See Ex. B, Emulex’s First Amended Preliminary Claim Constructions and Extrinsic Evidence, Sept. 10, 2010.) Emulex never subsequently amended its proposed constructions and never sought any construction of “input circuit.” Emulex has therefore waived its right to seek construction of the “input circuit” limitation. See Verizon Servs. Corp. v. Cox Fibernet Va., Inc., 602 F.3d 1325, 1334 (Fed. Cir. 2010) (holding that party is not entitled to post-trial claim construction where it “never identified at any time 1 All exhibits are exhibits to the Declaration of Jason H. Liss Accompanying Broadcom Corporation’s Reply Brief in Support of Its Motion for Judgment as a Matter of Law of Infringement of Claim 1 of the ’194 Patent (BE3 and QT2025), and Claim 5 of the ’124 Patent (SOC442), filed concurrently. Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 8 of 21 Page ID #:25073 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 6 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS during the proceedings before the district court any specific claim term that was misconstrued or that needed further construction”); see also, e.g., Finjan, 626 F.3d at 1207; Conoco, 460 F.3d at 1359. Emulex is also wrong to suggest that Broadcom’s understanding that claim 1 does not require the C3MOS circuits themselves to be used to divide down the frequency of the signal is somehow a “new position.” (Emulex Op. at 1.) Broadcom’s expert at trial repeatedly made clear that his position was that the demultiplexing did not have to occur in C3MOS to satisfy claim 1, and that he understood that Dr. Nikolic was taking a contrary position. (See, e.g., Trial Tr. Sept. 22, 2011, at 139:8-15 (testimony of Dr. Fair explaining disagreement with Dr. Nikolic’s view; id. at 157:5-6 (same).) Moreover, Emulex was aware well before trial-certainly at the latest, after the deposition of Dr. Nikolic-that Broadcom’s position was that a product containing C3MOS and performing deserialization infringed, even if the C3MOS circuits themselves did not perform the actual demultiplexing by dividing down the signal. (See, e.g., Ex. C, Deposition of Borivoje Nikolic, June 8, 2011, at 215:3-231:22.) In fact-at least at the time of his deposition-Dr. Nikolic himself not only understood Broadcom’s position, but agreed that, depending on the implementation, a product (like the BE3 and QT2025) that contained C3MOS and a deserializer could infringe, even if it had a non-C3MOS circuit performing deserialization: [Q.] Does a product that contains a C[3]MOS circuit that does not perform deserialization and also a circuit that does perform deserialization but that is not C[3]MOS, can such a product infringe any of the asserted claims? A. That’s a very hypothetical question, and I can give you a hypothetical answer: Possibly. It really depends on the implementation. Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 9 of 21 Page ID #:25074 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 7 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS (Id. at 216:16-23 (emphasis added).) Emulex has been aware of Broadcom’s position for months; it cannot claim surprise or argue that Broadcom’s position on claim 1 is somehow “new.” Rather, it is Emulex that is improperly seeking a new, never-before-requested, and limiting claim construction after trial. 2. Claim 1 Does Not Require a C3MOS Demultiplexer. Nothing in the language of claim 1, the specification, the prosecution history, or the evidence at trial supports Emulex’s view that claim 1 requires that the claimed C3MOS circuits must themselves demultiplex-i.e., perform the actual dividing down of the frequency of the incoming signal. (Emulex Op. at 3, 11 n.7.) Thus, even if Emulex were permitted to raise its new, limiting claim construction for the first time post-trial, the Court should reject it. Preliminarily, Emulex focuses much of its brief arguing that the “input circuit” portion of the “first circuitry” of claim 1 must be “configured to deserialize.” (Emulex Op. at 9-15.) In particular, Emulex relies on statements in the prosecution history and testimony by Broadcom’s expert Dr. Fair that suggest that the claimed input circuit is “configured to deserialize.” (Id. at 8, 10.) But the actual dispute between the parties is not about whether the deserialization occurs in the “input circuit” or elsewhere in the “first circuitry.” It is whether the claim requires that the C3MOS circuits themselves are the circuits in the deserializer that demultiplex-i.e., divide down the frequency of the signal. As Emulex admits in its brief, its noninfringement position is that “the alleged C3MOS was not used for deserialization.” (Emulex Op. at 3.) Emulex therefore argues in its opposition for a narrow construction of claim 1, in which the claim would require that the C3MOS circuits perform the actual demultiplexing- i.e., dividing down of the frequency of the incoming signal. (Emulex Op. at 16.) Neither the intrinsic nor the extrinsic evidence supports such a narrow construction. Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 10 of 21 Page ID #:25075 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 8 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS First, the language of claim 1 includes no requirement whatsoever of C3MOS used to demultiplex. The claim requires the first circuitry and the input circuit to be “implemented using” C3MOS-but, as the Court has already made clear, “implemented using” requires only that the circuitry includes but is not limited to C3MOS. (See Order Regarding Markman / Claim Construction Hearing, Dec. 17, 2010, Dkt. No. 261, at 17.) Thus, the claimed “first circuitry” can contain circuits that are not C3MOS circuits. The claim requires that the circuitry be “configured to deserialize”-but the claim says nothing about which portions of the circuitry-the C3MOS portions or the non-C3MOS portions-must perform the actual dividing down of the frequency of the incoming signal. By contrast, other claims of the SerDes patent family-including certain claims of the ’124 patent- specifically require that the C3MOS portion of the circuit itself perform the signal division. See, e.g., ’124 patent col.8 ll.40-45 (“a deserializer circuit block, implemented using [C3MOS] . . . that is operable to convert the input signal into a parallel signal”). Second, the specification makes clear that C3MOS can be used in an “input circuit” to process high-speed signals, without the C3MOS actually performing the signal division operation itself. Emulex makes the broad claim that, in the specification, “[i]n all cases, the structure that performs the deserialization is not some circuit “apart from” the input circuit but the input circuit itself.” (Emulex Op. 7-8.) But this argument is simply incorrect. To the contrary, the specification includes an express example-discussed in detail at trial and in Broadcom’s opening brief-of a C3MOS input circuit that is separate from the structure that performs the signal division. Specifically, the specification says: For example, to enable an integrated circuit to be used in ultra high speed applications, the input and output circuitry that interfaces with and processes the high speed signals is implemented using C3MOS. Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 11 of 21 Page ID #:25076 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 9 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS The circuit also employs C3MOS to divide down the frequency of the signals being processed to a low enough frequency where conventional CMOS logic can be used. ’194 patent col.7 ll.6-12 (emphases added). As Broadcom explained in its opening brief, this section confirms that the “input . . . circuitry” can be separate from the circuitry that “divide[s] down the frequency.” (See Broadcom Mem. at 4-5, 17.) Either or both can be implemented using C3MOS. See ’194 patent col.7 ll.6-12. Broadcom’s expert-Dr. Fair-also testified in detail about this portion of the specification at trial, and concluded that the example “describe[s] first circuitry that’s implemented such that it processes a signal and then it deserializes a signal. So you’ve got a number of things going on in C3MOS . . . .” (Trial Tr. Sept. 23, 2011, at 46:9-13 (emphasis added).) Notably, even though this section of the specification was a focus of Broadcom’s opening brief, and even though Broadcom’s expert Dr. Fair testified about it in detail at trial, Emulex’s opposition makes no mention of it, no effort to address Broadcom’s argument based on it, and no attempt to square it with its claim that the C3MOS input circuit must divide down the frequency “in all cases.” (Emulex Op. at 7-8.) Emulex also identifies examples in the specification in which input circuitry that is implemented using C3MOS performs deserialization. (See, e.g., Emulex Op. at 7 (citing ’194 patent at col.7 ll.27-28 (“C3MOS input circuitry 1100 is a deserializer”) and col.7 ll.16-18 (“C[3]MOS input circuit 1000 receives a high frequency input signal IN and outputs a divided down version of the signal IN/n.”)).) Preliminarily, both of Emulex’s citations are to exemplary configurations in particular embodiments in the patent-not to any portion of the specification that purports to describe the invention as a whole. Such examples do not limit the claims. See, e.g., Kara Tech. Inc. v. Stamps.com Inc., 582 F.3d 1341, 1348 (Fed. Cir. 2009) (“The patentee is entitled to the full scope of his claims, and Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 12 of 21 Page ID #:25077 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 10 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS we will not limit him to his preferred embodiment or import a limitation from the specification into the claims.”). But more importantly, those examples say nothing about whether the C3MOS circuits within that input circuitry are performing the demultiplexing itself, or whether those C3MOS circuits process the signal in the input circuit before other (non-C3MOS) circuits in the input circuitry actually perform the demultiplexing. Third, Emulex’s two partial and out-of-context quotes from the prosecution history of the ’194 patent do not support its construction of claim 1. (Emulex Op. at 3, 8.) Emulex relies on the statement that “[. . .] the claimed input circuit would be block 1100 (including latches 1104 and 1106) that ‘is configured to deserialize the input signal.’” (Emulex Op. at 8 (quoting Qureshi Decl. Ex. 2, JX-9 (’194 prosecution history), at ELX-PAT-00011520).) But Emulex omits the first half of the sentence, which reads: “In the context of the exemplary embodiment shown in Fig 11 for illustrative purposes . . . .” (’194 prosecution history at ELX-PAT- 00011520 (emphases added).) As the portion that Emulex omitted makes clear, the statement in the prosecution history was not intended to limit any claim, but merely describing one particular embodiment-the embodiment in Figure 11. Moreover, the statement in the prosecution history uses (in quotation marks) the language “configured to deserialize the input signal”-a phrase that does not appear in claim 1. (Id. (emphasis added).) In the language of claim 1, the circuitry is “configured to deserialize the first signal,” not “the input signal.” (’194 patent col.9 ll.12-13.) Given the express language in the history limiting the discussion to “illustrative purposes” and an “exemplary embodiment,” and given that the claim language at issue does not track the language used in the prosecution history, the statements on which Emulex relies fall well short of the “clear and unmistakable disavowal” necessary to limit claim scope. E.g., Lazare Kaplan Int’l, Inc. v. Photoscribe Techs., Inc., 628 F.3d 1359, 1370 (Fed. Cir. 2010). Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 13 of 21 Page ID #:25078 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 11 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS Fourth, the testimony of Dr. Hairapetian-the inventor of the ’194 patent- does not limit the claim to require C3MOS be used to divide down the signal. Emulex relies on Dr. Hairapetian’s testimony that C3MOS is “vital” in high speed applications (Emulex Op. at 9)-but this has never been in dispute. There is, in fact, no dispute that the two high-speed chips at issue in this motion use C3MOS, and there is no dispute that the C3MOS “has to be in there for a reason.” (Trial Tr. Sept. 30, 2011, at 46:4-10.) Emulex also relies on Dr. Hairapetian’s testimony that he “used C[3]MOS in areas where it was essential”-another fact that is not in dispute. But Emulex ignores Dr. Hairapetian’s later testimony explaining just which portions those were-including the “front end or the first stages” (i.e., the circuitry that processes the high speed signal entering the chip, before it is deserialized). (Trial Tr. Sept. 22, 2011, at 33:12-22.) Nothing in Dr. Hairapetian’s testimony supports Emulex’s construction requiring C3MOS to be used to divide down the signal. Finally, contrary to Emulex’s suggestion, Dr. Fair’s testimony is entirely consistent with Broadcom’s position. Emulex relies heavily on Dr. Fair’s testimony that, because the claim says “C3MOS” and “is configured to deserialize,” the infringing circuitry must include “C3MOS” and be “configured to deserialize.” (Emulex Op. at 10 (citing Trial Tr. Sept. 23, 2011, at 38:12-39:2).) But Broadcom has never disputed that-in Emulex’s counsel’s words-“it’s ‘and’ and not an ‘or’ situation.” (Id.) Dr. Fair never said, however, that the claim required that the C3MOS circuits themselves perform demultiplexing. Rather, he testified-repeatedly-that he disagreed with Dr. Nikolic’s view that the C3MOS circuits had to demultiplex to infringe claim 1. (See, e.g., Trial Tr. Sept. 22, 2011, at 139:8-15; id. at 157:5-6.) Thus, even if the Court were to address Emulex’s belated request to construe claim 1, it should not construe it as Emulex contends. (Emulex Op. at 16.) Rather, Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 14 of 21 Page ID #:25079 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 12 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS the Court should construe the “wherein the first circuitry” limitation to require simply a first circuitry that includes an input circuit that is implemented using C3MOS, and that the first circuitry be configured to deserialize-whether or not the C3MOS circuits themselves are actually performing demultiplexing or performing some other function. Thus, once Emulex’s late and unsupported limiting construction of claim 1 is rejected, the “only one reasonable conclusion” from the evidence is that the BE3 and QT2025 infringe, and Broadcom is entitled to judgment as a matter of law. See Geo. M. Martin Co., 618 F.3d at 1300. 3. Emulex Admits that the BE3 and QT2025 Include C3MOS and Deserialize. As Emulex admits in its brief, the only noninfringement defense that it put on as to the BE3 and QT2025 was that “the alleged C3MOS was not used for deserialization.” (Emulex Op. at 3.) Emulex does not dispute-and its expert Dr. Nikolic admitted-that the circuits that Dr. Fair identified as C3MOS in the BE3 and QT2025 are, in fact, C3MOS. (Trial Tr. Sept. 30, 2011, at 46:4-10.) Nowhere in its brief does Emulex dispute that the BE3 and QT2025 circuitry that Dr. Fair identified as “configured to deserialize” does, in fact, perform deserialization. (See Broadcom Mem. at 16, 19 (summarizing Emulex concessions).) Emulex makes two arguments against JMOL in the event that the Court rejects its claim construction-neither of which have merit. First, Emulex suggests that the jury could have found that Dr. Fair had “little or no credibility” and rejected his testimony outright on that basis. (Emulex Op. at 16-17.) But- apart from its dispute about whether the C3MOS circuitry performs demultiplexing-Emulex never disputed any of Dr. Fair’s opinions concerning the BE3 or the QT2025. Moreover, Emulex’s own expert-and Emulex’s counsel- agreed with Dr. Fair that the BE3 and QT2025 contain C3MOS and include a deserializer. (Trial Tr. Sept. 30, 2011, at 45:14-17, 46:4-5, 8-10; Trial Tr. Oct. 6, Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 15 of 21 Page ID #:25080 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 13 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS 2011, at 117:20-21, 122:1-3, 146:16-20, 151:16-18.) This is not a case in which Dr. Fair’s opinion was merely “uncontested” and subject to rejection by a reasonable jury; this was a case in which Emulex and its expert affirmatively conceded that the limitations were present in the accused devices. See, e.g., Teleflex, Inc. v. Ficosa N. Am. Corp., 299 F.3d 1313, 1322, 1335 (Fed. Cir. 2002) (affirming judgment based in part on concessions). Second, Emulex argues that a jury could have found that the “data recovery block” of the QT2025 and the “DFE” or “Decision Feedback Equalizer” of the BE3 were not part of the claimed “first circuitry,” because Broadcom allegedly “presented no testimony” on the issue. (Emulex Op. at 17.) But Broadcom did present testimony on precisely this issue. In fact, in two of the very demonstratives of Dr. Fair’s that Emulex reproduced in its own opposition brief, Broadcom identified the “data recovery block” and the “DFE” as part of the “first circuitry.” Specifically, on page 14 of Emulex’s brief, it reproduces PDX02-120, which was Dr. Fair’s demonstrative identifying the QT2025 block labeled “Data Recovery” as part of the “first circuitry comprising an input circuit.” (Emulex Op. at 14.) Likewise, on page 12 of its brief, Emulex reproduces PDX02-50, which was Dr. Fair’s demonstrative identifying the BE3 block “dfe_slice” as part of the “first circuitry.” (Emulex Op. at 12 (emphasis added).) Thus, Emulex is wrong to argue that JMOL should be denied because “Broadcom presented no testimony” on these issues. B. The SOC442 Infringes Claim 5 as a Matter of Law. Emulex’s only argument against judgment as a matter of law on claim 5 of the ’124 patent is that infringement turns on “a factual issue that . . . was disputed between the experts.” (Emulex Op. at 18.) But whether an issue is “disputed between the experts” is not the standard for judgment as a matter of law. Broadcom is entitled to JMOL if “the evidence, construed in the light most Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 16 of 21 Page ID #:25081 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 14 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS favorable to [Emulex], permits only one reasonable conclusion”-that the SOC442 infringes claim 5. Geo. M. Martin Co., 618 F.3d at 1300. An expert’s conclusory opinion, unsupported by the facts, does not preclude JMOL. See, e.g., Upjohn, 225 F.3d at 1311 (noting that “there must be factual support for an expert’s conclusory opinion”); cf. Sitrick v. Dreamworks, LLC, 516 F.3d 993, 1001 (Fed. Cir. 2008) (“Conclusory expert assertions cannot raise triable issues of material fact on summary judgment.”) Here, Emulex seeks to avoid JMOL by relying on the on the unsupported and factually incorrect conclusion of its expert that the signal pair that Dr. Fair identified as an “output” is not actually “output” from the deserializer. Claim 5 requires that the claimed circuitry have an “output that is operable to transmit the parallel signal.” (’124 patent col.8 ll.49-50.) Dr. Fair identified this “output” as the signal pair “Q1P” and “Q1N” in the circuit block “DFEC4_DEFF_R3.” (See Broadcom Mem. at 11; PDX02-144; JX-0388 at page TEAC 0278.) As Broadcom explained in its opening brief-and as Emulex concedes in its opposition (Emulex Op. at 18-19)-this circuit actually generates two copies of the signal pair Q1P and Q1N: (See JX-0388 at page TEAC 0278 (red and blue annotations added).) The first Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 17 of 21 Page ID #:25082 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 15 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS copy of Q1P and Q1N is shown with a bold arrow (indicating that it is output from the circuit diagram to another location). The second copy of Q1P and Q1N continues to the right side of the circuit, where it is input into a multiplexer along with another signal pair (Q2P and Q2N), generating a reserialized signal pair labeled “QP” and “QN.” (JX-0388 at page TEAC 0278.) It is the first copy of Q1P and Q1N that Dr. Fair identified as the “output.” (See Broadcom Mem. at 11; PDX02-144.) Emulex’s only purported evidence that the first copy of Q1P and Q1N is not “output” is the following testimony by Dr. Nikolic: Q. When we look at the actual schematic, and this is the first circuit, are these outputs coming out of the first circuit? A. Actually they are not. They are being used internally inside this first block to correct for DC offset -- to correct for the DC offset. What we have here is the input signal is coming in here, is split into two parts. We can say that it is being deserialized. But then it is being serialized again through a multiplexer, and one output is coming out. This is the only output that is coming out of that block of the data. (Trial Tr. Sept. 29, 2011, at 131:18-132:4.) This testimony is insufficient to defeat Broadcom’s JMOL motion for two reasons. First, contrary to Emulex’s suggestion, it is not at all clear that Dr. Nikolic was referring to the first copy of Q1P and Q1N-rather than the second copy- when he testified that signals are being “used internally inside this first block to correct for DC offset.” (Emulex Op. at 18, 19.) Dr. Nikolic’s testimony never even acknowledges that there are two copies of Q1P and Q1N within the deserializer circuit block. He appears instead to refer to a single signal pair that is “used internally . . . to correct for DC offset” and “then it is being serialized again Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 18 of 21 Page ID #:25083 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 16 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS through a multiplexer.” (Trial Tr. Sept. 29, 2011, at 131:21-132:2.) Nowhere does Dr. Nikolic acknowledge that (as Emulex now admits) there are two distinct copies of Q1P and Q1N within the deserializer circuit block. Thus, the testimony that Emulex identifies does not actually appear to dispute that the first copy of Q1P and Q1N is output without being reserialized. Second, even to the extent that Dr. Nikolic’s testimony could be read as concluding that the first copy of Q1P and Q1N is “used internally,” there is no factual support whatsoever for that conclusion in the record. To the contrary, the SOC442 schematics-which are the only evidence on which Dr. Nikolic purports to rely for his conclusions about Q1P and Q1N-unambiguously show that the first copy of the signal pair is output from the deserializer circuit block. The deserializer circuit block that Dr. Fair identified is the block labeled DFEC4_SYS_R6, which is shown in block form on in the SOC442 schematics. (See Broadcom Mem. 12; PDX02-146; JX-0388 at page TEAC 290.) The two copies of Q1P and Q1N are generated in a sub-block of that DFEC4_SYS_R6 circuit. (See JX-0388 at page TEAC 291.) As the circuit schematics unambiguously show, the signal pair Q1P and Q1N are output from the DFEC4_SYS_R6 block and into the next circuit block in the signal path: Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 19 of 21 Page ID #:25084 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 17 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS (JX-0388 at page TEAC 290 (highlighting added).) In fact, the circuit diagrams are so unambiguous about the path of the first copy of the Q1P and Q1N signal pair that Dr. Nikolic himself admitted on cross examination that same signal pair- Q1P and Q1N-is input into the next circuit block. (See Trial Tr. Sept. 30, 2011 at 78:5.) Dr. Nikolic’s conclusory testimony that the signal pair is not output therefore lacks factual support and does not preclude JMOL. III. CONCLUSION For the foregoing reasons, and the reasons discussed in Broadcom’s opening brief, Broadcom respectfully requests that the Court enter judgment in favor of Broadcom that Emulex’s BE3 and QT2025 products infringe claim 1 of the ’194 patent, and that its SOC442 product infringes claim 5 of the ’124 patent. Dated: December 2, 2011 WILMER CUTLER PICKERING HALE AND DORR LLP By: /s/ Louis W. Tompros Louis W. Tompros Attorneys for Plaintiff BROADCOM CORPORATION Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 20 of 21 Page ID #:25085 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 18 BROADCOM’S REPLY BRIEF ON JMOL OF INFRINGEMENT OF ’194 AND ’124 PATENTS Broadcom Corporation v. Emulex Corporation, Case Nos. SACV09-1058 JVS (ANx) and SACV 10-03963-JVS (ANx) PROOF OF SERVICE Fed. R. Civ. P. 5(b) & L.R. 5-3 I, Jason H. Liss, hereby certify that a copy of the foregoing Plaintiff Broadcom Corporation’s Reply Brief in Support of Its Motion for Judgment as a Matter of Law of Infringement of Claim 1 of the ’194 Patent (BE3 and QT2025), and Claim 5 of the ’124 Patent (SOC442) was served upon the following parties as indicated below on this 2nd day of December, 2011. For Emulex Corporation: smg@jmbm.com broadcom-emulex@fr.com broadcomemulexservice@tklaw.com (email addresses for service on Emulex per agreement of the parties) Via Hand Delivery Via Overnight Courier (1 copy) Via Facsimile Via Electronic Mail (1 copy) I certify under penalty of perjury that the foregoing is true and correct. /s/ Jason H. Liss Jason H. Liss ACTIVEUS 91509590v10 Case 8:09-cv-01058-JVS-AN Document 987 Filed 12/02/11 Page 21 of 21 Page ID #:25086
