Free Reply to Response to Motion - District Court of California - California

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 1 of 29

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

MATTHEW D. POWERS (Bar No. 104795) E-Mail: [email protected] EDWARD R. REINES (Bar No. 135960) E-Mail: [email protected] JEFFREY G. HOMRIG (Bar No. 215890) E-Mail: [email protected] JILL J. HO (Bar No. 236349) E-Mail: [email protected] WEIL, GOTSHAL & MANGES LLP Silicon Valley Office 201 Redwood Shores Parkway Redwood Shores, CA 94065 Telephone: (650) 802-3000 Facsimile: (650) 802-3100 Attorneys for Plaintiff-Counterclaim Defendant NETAPP, INC. UNITED STATES DISTRICT COURT NORTHERN DISTRICT OF CALIFORNIA SAN FRANCISCO DIVISION NETWORK APPLIANCE, INC., Plaintiff-Counterclaim Defendant, v. SUN MICROSYSTEMS, INC., Defendant-Counterclaim Plaintiff. Case No. C-07-06053 EDL NETAPP, INC.'S REPLY TO SUN MICROSYSTEMS, INC.'S RESPONSIVE CLAIM CONSTRUCTION BRIEF

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 2 of 29

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

TABLE OF CONTENTS Page I. INTRODUCTION .............................................................................................................. 1 II. ARGUMENT ...................................................................................................................... 1 A. U.S. Patent No. 5,925,106.......................................................................................................... 1 1. "Domain Name" (Claims 1, 9, 14, 22, 27, and 35) .................................................... 2 a. Sun's "Preferred Embodiment" Argument Is Misdirected ......................... 2 b. Sun Misapplies The Doctrine Of Claim Differentiation ............................. 2 c. Sun Ignores The Meaning Of "Domain Name" As Expressed In The Specification And The File History ..................................................... 3 2. "Server Identification Data" (Claims 1, 4, 9-10, 14, 17-18, 23-24, 27, 30, 35-36) ...................................................................................................................... 3 a. NetApp's Construction Is Supported By The Intrinsic Evidence .............. 4 b. Sun's Attempt To Cure Its Overreaching Construction Is Ineffective........ 5 B. U.S. Patent No. 5,459,857 .......................................................................................................... 6 1. "Responsive to Writing..." / "Responsive to Receipt..." (Claims 6 and 11) ............ 6 a. NetApp's Construction Gives Effect To The Plain Meaning Of The Claims As Written; Sun's Construction Simply Rewrites Them................ 6 b. Sun's Purported Justification Is Wrong As A Matter Of Law .................... 8 c. Sun's Construction Selectively Ignores The Intrinsic Evidence And Redefines Specification Terms While Improperly Picking And Choosing Limitations To Import................................................................. 9 C. U.S. Patent No. 5,749,095 ........................................................................................................ 13 1. "Completing [a] Write Operation..." (Claims 1, 11, and 17) ............................... 13 a. Sun's Argument That The Claims Are "Clear On [Their] Face To [A] Person Of Ordinary Skill In The Art" Is Irrelevant ............................ 13 b. The "Provided To Subsequent Read Operations . . . To The Same Address" Requirement of Sun's Alternative Construction Lacks Support In Either The Claim Language Or The Specification Of The '095 Patent ......................................................................................... 14 c. The "Is Or Will Be Coherent" Requirement Of Sun's Alternative Construction Is Unsupportable.................................................................. 15 d. Sun's Criticism Of NetApp's Proposed Construction Is Logically Flawed ....................................................................................................... 15 e. Sun Misapplies The "Rules of Claim Construction" ................................ 16 D. U.S. Patent No. 6,873,630........................................................................................................ 17 1. "Portion" / "Element" (Claims 3, 5, 8, 13-15, 21-23, 45-48, 50, 52, 73, 76, 79, 89, 113, 114, 116, 117) ................................................................................... 17 a. The Claim Terms Require Construction ................................................... 17 b. The '630 Patent Claims Should Be Limited To Data Striping Below The MAC Layer ........................................................................................ 18
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

i

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 3 of 29

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 III.

TABLE OF CONTENTS (continued) Page c. Sun's Claim Differentiation Arguments Fail ............................................ 21 CONCLUSION ................................................................................................................. 23

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

ii

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 4 of 29

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

TABLE OF AUTHORITIES Page FEDERAL CASES Andersen Corp. v. Fiber Composites, LLC, 474 F.3d 1361, 1370 (Fed. Cir. 2007)..................................................................................... 16 Applied Materials, Inc. v. Advanced Semiconductor Materials America, Inc., 98 F.3d 1563 (Fed. Cir. 1996)................................................................................................... 5 AstraZeneca AB, Aktiebolaget Hassle, KBI-E, Inc. v. Mutual Pharmaceutical Co. Inc., 384 F.3d 1333 (Fed. Cir. 2004)............................................................................................... 19 Becton Dickinson & Co. v. C.R. Bard, Inc., 922 F.2d 792 (Fed Cir. 1990).................................................................................................... 6 CVI/Beta Ventures, Inc. v. Tura LP, 112 F.3d 1146 (Fed. Cir. 1997)................................................................................................. 4 Curtiss-Wright Flow Control Corp. v. Velan, Inc., 438 F.3d 1374 (Fed. Cir. 2006)......................................................................................... 16, 21 Helmsderfer v. Bobrick Washroom Equipment, Inc., 527 F.3d 1379 (Fed. Cir. 2008)................................................................................................. 8 Karlin Technology Inc. v. Surgical Dynamics, Inc., 177 F.3d 968 (Fed. Cir. 1999)................................................................................................... 4 O2 Micro International Ltd. v. Beyond Innovation Technology Co., 521 F.3d 1351 (Fed. Cir. 2008).......................................................................................... 17-18 On Demand Machine Corp. v. Ingram Industrial Inc., 442 F.3d 1331 (Fed. Cir. 2006)......................................................................................... 19, 20 Ormco Corp. v. Align Technology, Inc., 498 F.3d 1307, 1316 (Fed. Cir. 2007)....................................................................................... 4 Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005)............................................................................. 4, 5, 6, 13, 14 SRAM Corp. v. AD-II Engineering, Inc., 465 F.3d 1351 (Fed. Cir. 2006)................................................................................................. 7 Sinorgchem Co. Shandong v. International Trade Commission, 511 F.3d 1132 (Fed. Cir. 2008)............................................................................................... 16

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

iii

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 5 of 29

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 A.

I. INTRODUCTION As demonstrated below, Sun has failed to present any compelling reason why NetApp's proposed constructions for United States Patent Nos. 5,925,106 ("the '106 patent"), 5,459,857 ("the '857 patent"), 5,749,095 ("the '095 patent"), and 6,873,630 ("the '630 patent") should not be adopted by the Court. II. ARGUMENT U.S. PATENT NO. 5,925,106

As illustrated by the parties' opening and responsive briefs, the fundamental dispute with respect to the '106 patent is whether the claim terms should be construed to approximate what was purportedly invented or broadened beyond what the named inventors represented to the public and the Patent Office as their invention. NetApp's proposed

constructions should be adopted because they stay true to the nature of the alleged invention, which, as described by the inventors themselves, is narrowly aimed at helping inexperienced users of the Internet so they do not become disoriented when viewing websites bearing either non-descriptive or mis-descriptive domain names. Walter Decl. Exh. A at 1:58-2:3.1 Sun contends, on the other hand, that it is entitled to an exceedingly broad reading of its claims. For example, Sun argues that "the invention can be used on any type of network and by any type of client to retrieve information about a server accessed through any type of domain name." See Sun's Responsive Cl. Constr. Br. [hereinafter "Sun Opp."] at 4. But this aggressive contention is overreaching, as Sun's brief itself acknowledges. Indeed, Sun concedes that it "has no objection to the Court stating in its claim construction order that an IP address does not itself constitute either limitation" (i.e., "server identification data" or "descriptive information about a server") and likewise "does not object to the Court stating in its claim construction order

All references to "Walter Decl." in this brief refer to the Declaration of Derek C. Walter in Support of NetApp's Opening Claim Construction Brief, filed on July 8, 2008. Likewise, references to NetApp's expert declarations refer to those filed with the opening brief.
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

1

1

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 6 of 29

1 2 3 4 5 6 7 8 9 10 11

that the domain name, hypertext links, URLs, web page titles and bookmarks do not by themselves satisfy" those same limitations. Id. at 12, 13. Sun's effort to preserve its proposed constructions by carving out what it expressly disclaimed during prosecution results in unworkable constructions that continue to depart from the true nature of the purported invention, and should be rejected. 1. "Domain Name" (Claims 1, 9, 14, 22, 27, and 35) Term "domain name" NetApp's Construction Sun's Construction A third-party-approved name of A name that has a numerical IP a website on the Internet, i.e. a address associated with it. registered domain name. Sun's "Preferred Embodiment" Argument Is Misdirected

a.

Sun makes much of the fact that the '106 patent merely describes a "preferred embodiment utilizing WWW browser and WWW server applications." Walter Decl. Exh. A at 12 1:36-37; see Sun Opp. 3-6. This misses the point. NetApp is not asking the Court to construe 13 "network" ­ a different claim term ­ to exclude private networks. Of course, it would be illogical 14 to implement the invention in a private network of a company, where the owner of the server(s) is 15 known to be the company itself, because the intended purpose of the purported invention is to 16 assist inexperienced users of the Internet in identifying the server of a viewed web page. But 17 even if it were so used, the claims of the '106 patent would require that a "domain name" be used 18 to obtain "server identification data," i.e., a registered domain name as opposed to an unregistered 19 name on a local or private network. 20 21 22 adoption of NetApp's proposed construction of "domain name" would not result in the '106 23 patent reciting two claims with identical scope. 24 Instead, Sun argues that the invention should not be limited to the use of WWW, 25 URLs and HTTP because those limitations are added by various dependent claims. Yet, the 26 Internet is broader than the World Wide Web, see generally Reno v. Am. Civil Liberties Union, 27 521 U.S. 844, 849-53 (1997), and encompasses other public network applications existing at the 28
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

b.

Sun Misapplies The Doctrine Of Claim Differentiation The

Sun's claim differentiation arguments should be accorded no weight.

2

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 7 of 29

1 2 3 4 5 6

time the '106 patent was filed, such as Telnet. See Almeroth Decl. at 10. Thus, understanding that the independent claims of the '106 patent require the use of registered domain names would not render them identical in scope to dependent claims that further recite WWW, URLs or HTTP. c. Sun Ignores The Meaning Of "Domain Name" As Expressed In The Specification And The File History

As understood by a person of skill in the art, a "domain name" in the context of the '106 patent is a third-party-approved name of a website on the Internet, i.e., a registered domain 7 name. See Almeroth Decl. at 5-9. The specification itself confirms that "[d]omain names are 8 requested by the maintainers of the website and are approved by a third party." Walter Decl. Exh. 9 A at 1:56-57. 10 documents that describe domain names as being registered with a third-party Internet domain 11 name registry. Id. at 1:27-44; see also NetApp's Opening Cl. Constr. Br. at 6 n.2. 12 Likewise, during prosecution, the patentee expressly stated that "[t]he `Domain 13 Name' term is used in accordance with standard usage in the field." Walter Decl. Exh. E. at 6. 14 As NetApp pointed out in its opening brief, the patentee even distinguished domain names from 15 the additional information about the server that is retrieved on the basis that the latter, unlike 16 domain names, "need not be unique nor approved by a third party." Id. at 7. This intrinsic 17 evidence combined with all the extrinsic evidence presented by both sides strongly support 18 19 20 21 22 23 24 25 26 27 28
2

The specification also incorporates by reference various Internet standards

adoption of NetApp's proposed construction.2 2. "Server Identification Data" (Claims 1, 4, 9-10, 14, 17-18, 23-24, 27, 30, 35-36) Term "server identification data" NetApp's Construction Human-friendly information identifying a specific web server designed not to be intimidating to inexperienced users of the World Wide Web. Sun's Construction Information that uniquely identifies one server from other servers and can be seen by a user.

Sun's proposed construction of "server identification data" is similarly strained. Sun argues for a broad interpretation, yet backtracks in the same breath, conceding that it would

Sun points to NetApp's use of the phrase "NIS domain name" in its manuals, see Sun Opp. at 8, but the use of this phrase as opposed to "domain name" by itself, further confirms that "domain name" has a standard meaning from which "NIS domain name" needed to be distinguished. 3
Case No. C-07-06053 EDL

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 8 of 29

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

have "no objection" to the Court limiting its construction to avoid reading on prior art and "moot" NetApp's argument. See Sun Opp. 12-13. Sun cannot have it both ways. a. NetApp's Construction Is Supported By The Intrinsic Evidence

Sun summarily argues that there is no basis for importing the limitations "humanfriendly" and "designed not to be intimidating." Sun Opp. at 10. But it cannot ­ and, indeed, does not attempt to ­ reconcile its position with the stated goal of the invention: to assist inexperienced users of the Internet in identifying the origin of a website by "presenting humanfriendly server identification information to a user." See Walter Decl. Exh. A at 1:16-17.3 Indeed, during prosecution, the patentee reiterated that the invention was intended to solve the problem of users being confronted with "non- or mis-descriptive server domain names by

presenting server-specific descriptive information to the user that identifies and/or describes the server." See Walter Decl. Exh. E at 8. The narrow goal of the patent is significant because it is well-established that "the problem the inventor was attempting to solve, as discerned from the specification and the prosecution history, is a relevant consideration" in claim construction. CVI/Beta Ventures, Inc. v. Tura LP, 112 F.3d 1146, 1160 (Fed. Cir. 1997). The sole purpose of the '106 patent is to assist novice users of the Internet (or more specifically, to help inexperienced individuals navigating websites on the World Wide Web). Against that backdrop, "server identification data" must (1) be human-friendly; (2) not be intimidating to an inexperienced user; and (3) identify the server hosting the viewed website. See Ormco Corp. v. Align Technology, Inc., 498 F.3d 1307, 1316 (Fed. Cir. 2007) ("Accordingly, to attribute to the claims a meaning broader than any indicated in the patents and their prosecution history would be to ignore the totality of the facts of the case and exalt slogans over real meaning."). While there is admittedly a high bar for limiting the scope of the patent to what the patentee says is "the present invention," see Karlin Tech. Inc. v. Surgical Dynamics, Inc., 177 F.3d 968, 973 (Fed. Cir. 1999), here that standard is met because the alleged improvement over the prior art was minimal, at best. Indeed, the only disclosed embodiment accomplishes the goal of assisting disoriented novices of the Internet by simply adding the name of the website owner in
3

Emphasis added throughout unless otherwise noted. 4
Case No. C-07-06053 EDL

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 9 of 29

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

parentheses next to the URL. Compare Walter Decl. Exh. A Figs. 3B & 3C; see also id. at 2:1116. Moreover, viewing the claims in the context of the state of the art at the time the '106 patent was filed confirms that "server identification data" requires something more than the previously known means of uniquely identifying a server, such as IP addresses. See Phillips v. AWH Corp., 415 F.3d 1303, 1333 (Fed. Cir. 2005) (noting that the differences between the claims and the prior art should be ascertained "to better define what the inventor holds out as the invention"); Applied Materials, Inc. v. Advanced Semiconductor Materials Am., Inc., 98 F.3d 1563, 1573 (Fed. Cir. 1996) (affirming the district court's construction that "correctly placed the term `cold purge process' in the context of the state of the art when the '389 invention was made"). Here, the specification explains that domain names are often "not descriptive or

misdescriptive of the site" and "IP addresses carry no descriptive value whatsoever." Walter Decl. Exh. A at 1:60-64. The patentee also expressly stated during prosecution that "an IP address that is returned by a name server" would fail to qualify as "server identification information." Walter Decl. Exh. E at 6. As explained in NetApp's opening brief, the type of information provided to identify the server must therefore be more descriptive to an inexperienced user of the Internet than an IP address. Finally, although Sun argues that the phrases "human friendly" and "designed not to be intimidating" would fail to provide guidance to a jury, this is untrue. Indeed, a jury is fully capable of determining what types of server-identifying information would be intimidating to a layperson. b. Sun's Attempt To Cure Its Overreaching Construction Is Ineffective

Sun's responsive brief fails to address the fact that Sun's proposed construction of "server identification data" would encompass even an IP address, which was expressly disclaimed by the patentee, because it uniquely identifies one server from another server and can be seen by a user. Instead, Sun attempts to "moot" NetApp's argument by inviting the Court to carve out from its proposed definition various types of data (not just IP addresses) that would read on prior art. This is exactly the type of judicial redrafting of claims that is not permitted. See generally
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

5

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 10 of 29

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

Phillips, 415 F.3d at 1327; Becton Dickinson & Co. v. C.R. Bard, Inc., 922 F.2d 792, 799 & n.6 (Fed Cir. 1990). Not only does this intricate attempt to draft around the inventors' admissions during prosecution result in an unworkable construction, it is wholly unnecessary if the Court stays true to the nature of the purported invention, as understood by those of ordinary skill in the art and the inventors themselves. Because it reflects this understanding, NetApp's construction should be adopted.

B.

U.S. PATENT NO. 5,459,857

NetApp's proposed constructions track the plain meaning of the disputed claim language. In stark contrast, Sun asks the Court to rewrite the asserted claims by importing limitations selectively from the specification. Sun offers but one justification for this aggressive position: that the '857 patent discloses only one embodiment, and that the asserted claims would not cover this embodiment under NetApp's construction. But this sole justification is both factually and legally deficient, and in no way warrants rewriting the plain language of the claims. As a result, the Court should adopt NetApp's proposed constructions. 1. "Responsive to Writing..." / "Responsive to Receipt..." (Claims 6 and 11) Term "In response to writing a data record to said one redundancy group" / "Responsive to writing a data record to one of said redundancy groups" "In response to the receipt of a stream of data records from said data processor" / "Responsive to the receipt of a stream of data records from said data processor" a. NetApp's Construction After and in reaction to the writing of a data record to a single redundancy group. Sun's Construction In response to writing the data record to memory associated with the one redundancy group. / Responsive to writing the data record to memory associated with one of the redundancy groups. Sun contends this phrase does not require construction because the phrase is clear on its face.

After and in reaction to the receipt of data records from a processor.

NetApp's Construction Gives Effect To The Plain Meaning Of The Claims As Written; Sun's Construction Simply Rewrites Them

NetApp's construction gives effect to the plain language of the claims which, as NetApp showed in its opening brief, requires that transmitting of data records occur only after
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

6

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

writing them, and that selecting space to write them occur only after they have been received. In contrast, rather than clarify or interpret the meaning of the words in the claims, Sun simply adds new words. Indeed, as it admits freely, Sun would have the Court rewrite the asserted claims by replacing the words "redundancy group" with the words "memory associated with a redundancy group." transmitting, in response to writing a data record to memory associated with said one redundancy group, said written data record to the other of said data storage subsystems...

Walter Decl. Exh. B at 14:12-14 (Sun's proposed modifications supplied). The "memory associated with" a redundancy group that Sun argues should be added to the claim is cache memory. See Sun Opp. at 17. This is significant, because the cache memory is completely separate from, and performs a different function than, the redundancy group. By replacing the requirement that the data record be written to the redundancy group with the proposed requirement that the data record be written to the cache memory, Sun changes the order of steps required under the claims. The plain language of the claims requires the data records to be transmitted to the second storage system after they are written to the single redundancy group. But, under Sun's construction, data records can be transmitted before they are written. See Sun Opp. at 20. As a result, another way of illustrating Sun's proposed change is by deleting the words "in response to" from the claim: transmitting, in response to writing a data record to said one redundancy group, said written data record to the other of said data storage subsystems... Walter Decl. Exh. B at 14:12-14 (Sun's proposed modifications supplied). This result is improper. Put simply, courts do not rewrite claims. See, e.g., SRAM

23 Corp. v. AD-II Engineering, Inc., 465 F.3d 1351, 1359 (Fed. Cir. 2006)(explaining that courts, 24 "are powerless to rewrite the claims and must construe the language of the claim at issue based on 25 the words used.")(citations omitted). NetApp's constructions should be adopted because they 26 give full effect to the plain language of the claims as written. 27 28
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

7

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 12 of 29

1 2 3 4 5

b.

Sun's Purported Justification Is Wrong As A Matter Of Law

Sun's sole argument in support of its request to have its claims rewritten hinges on the premise that every single claim of the '857 patent must somehow cover the "sole embodiment" allegedly disclosed in the specification. Sun argues that, "[t]he '857 patent

describes a single embodiment for writing duplicate data records through the data storage 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 with a construction that did "not cover the preferred embodiment or the other illustrated 21 22 23 24 25 26 27 28
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

subsystem," and refers to the "preferred and only embodiment," or similar phrases no fewer than eight times throughout its brief. Sun Opp. at 14, 17-19. The rule Sun wishes to invoke is far more nuanced. It is designed to reflect the common sense observation that a patent drafter would not generally draft a set of claims such that none of the claims would cover the preferred or only embodiment disclosed in the specification. But this rule, such as it is, cannot be woodenly applied to every situation. In fact, there is simply no requirement that every single claim of a patent cover a preferred embodiment, whether or not it is the only embodiment. Indeed, Sun points to no such rule. Where the set of claims as a whole covers the single embodiment, there can be no dispute that it is permissible for certain claims to exclude it. See Helmsderfer v. Bobrick

Washroom Equipment, Inc., 527 F.3d 1379, 1383 (Fed. Cir. 2008) ("It is often the case that different claims are directed to and cover different disclosed embodiments."). Helmsderfer dealt

embodiments." See, id. at 1383. The Court went on to explain that, "this does not mean that these embodiments are all excluded from the scope of the invention, but rather that they are excluded from the scope of these particular claims." Id. Thus, the law recognized that where a specification discloses one embodiment, and that embodiment is covered by certain claims, the drafter may elect to include additional claims that do not cover it. And, taking Sun's factual assertions at face value, that is precisely the

8

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 13 of 29

1 2 3 4

situation here.4 For example, dependent claim 26 actually requires the use of a cache memory, the very element that Sun's argument would require the court to read into all the claims of the patent. See Walter Decl. Exh B at 18:28-45 (claiming, "The data storage system of claim 24 wherein each of said data storage controls contains a cache memory..."). Likewise, claim 24 is

5 6 7 8 9 10 11 12 internally inconsistent. 13 disclosure, Sun's construction would redefine terms already defined in the specification, such as 14 "cache memory," and "redundancy group." 15 Sun's construction equates writing to a cache memory with writing to a 16 redundancy group, reasoning that, because a cache memory is "associated with the redundancy 17 group," ­ a vague statement that could mean anything ­ writing a data record to one redundancy 18 group is satisfied by writing a data record to a cache. See, e.g., Sun Opp. at 17 (explaining that, 19 "cache memory... is associated with the redundancy group."). This argument is at odds with the 20 evidence from the specification, which defines a redundancy group as follows: "A redundancy 21 group consists of N+M disk drives. The redundancy group is also called a logical volume or a 22 23 24 25 26 27 28 Contrary to Sun's assertions, it is not at all clear that the specification requires one and only one order of transmitting and writing. At 4:39-54, an order of data record write operations is disclosed, without any mention of transmitting data to the redundant storage control unit. See Walter Decl. Exh. B. In addition, as discussed below, infra, at 11-12, the passage Sun cites for the "one and only" order of performed steps, actually discloses a different order from that in the figure it purportedly describes. These facts show that there is not a single embodiment in the '857 patent, and that if there is one, it does not require one strict ordering of the steps of a data record write operation. Sun's reliance on a single passage from 11:12-12:9 of the '857 patent is therefore misleading, because it purports to show a single possible embodiment where at best it is merely an option disclosed in more detail than other contemplated options. On these facts, the idea that every claim should cover exactly what the preferred embodiment supposedly discloses, is simply wrong.
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

drafted in a manner fully consistent with the order asserted by Sun to be the preferred embodiment, reciting first a means for transmitting, and next a means for writing. See Walter Decl. Exh B at 7:55-18:23. In sum, Sun's argument is simply wrong as a matter of law. c. Sun's Construction Selectively Ignores The Intrinsic Evidence And Redefines Specification Terms While Improperly Picking And Choosing Limitations To Import

In addition to advocating a rewrite of the actual claim language, Sun's position is Although it purports to modify the claim language to match the

4

9

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 14 of 29

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

logical device."

Walter Decl. Exh. B at 7:18-20.5

Thus, not only does the patent define

redundancy group to mean disks and not cache memory, but it even goes so far as to name two other things redundancy groups might be called. If the patentees had meant to equate cache with redundancy group they should have written, "The redundancy group is also called a cache memory." They wrote no such thing. In addition, the cache memory is defined as a part of the storage control unit and not as a part of the redundancy groups. See, e.g., id. at Fig. 2 (showing the cache 512 as part of the storage control unit 105, which is separate from the redundancy groups also shown in the same figure). The distinction between redundancy group and cache is made extremely clear in the following passage from the specification: "No data stored in a redundancy group is modified. A virtual track is staged from a redundancy group into cache. The host then modifies some, perhaps all of the records on the virtual track. Then... the modified virtual track is selected to be destaged to a redundancy group." Id. at 3:65-4:4. It is therefore essential to the way the purported invention functions that the cache and the redundancy group not be the same thing, and that therefore writing to the one is not the same as writing to the other. Moreover the same passages Sun cites to support this conflation of cache memory and the redundancy group demonstrate that writing to cache and writing to redundancy group are completely separate operations. See, e.g., Walter Decl. Exh. B at 11:57-12:9 (explaining that "Upon receipt of the data record, storage cluster 501 forwards the data record.... to store the data record... in cache memory 512," and that later on in the process, "storage control unit 107 writes the data record from its cache memory into a selected redundancy group of the disk drives."). This passage not only demonstrates that cache memory and redundancy group are not the same, but that, in fact, the cache is "associated with" the storage control unit of which it is a part and not with a particular redundancy group. Sun's analysis of front-end versus back-end processes represents another attempt to rewrite the claims so as to contradict definitions from the specification. Sun explains that "the Other locations where the specification defines "redundancy group" to mean disks are found e.g. at 2:5-11, and 3:2-9, among others.
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

5

10

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 15 of 29

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

front-end operations, such as the write commands from the host processor to the data storage subsystems, are completely decoupled from the back-end operations, such as aggregating groups of data records in cache and transferring them to disk in bulk." Sun Opp. at 17. This argument is self-defeating because the specification defines the channel commands, which put the data record in cache memory as front-end, whereas it defines writing to the redundancy group as backend. See Walter Decl. Exh. B at 6:55-57 (equating front-end with channel transfer operations and backend with device transfer operations); id. at 9:57-10:58 (defining the "channel commands" used by a host processor to communicate with the DASD-emulating storage system); id. 6:5-31 (describing the process of writing to redundancy groups through "backend channels 108"). Therefore, according to Sun's logic, the "backend" write to the redundancy group is not complete until the data record leaves the front-end cache and enters the backend redundancy group. Thus the very front-end/back-end distinction Sun cites for the proposition that the completion of a write to the redundancy group means writing to cache, actually distinguishes the redundancy group from the cache as separate entities requiring separate write operations. Sun also uses a passage from the specification, describing when the "write complete" message is sent to the host processor, to justify its position that completion of the write to the redundancy group from the host processor's perspective is all that matters. See Sun Opp. at 15 (citing passages from the specification describing a sequence in which the host processor is given a "write complete" message before the data record is written to the redundancy group). In fact, the specification provides at least two conflicting options regarding when the "write complete" message is sent to the host processor. The passage cited by Sun, as discussed above, discloses sending "write complete" to the host processor before writing to the redundancy group. See Walter Decl. Exh. B at 11:62-12:7. But Figure 6, the figure purportedly described by the cited passage of the specification, discloses the "write complete" message being sent to the host processor after the data record is written out to disk. According to Sun's own characterization of the "write complete" message as defining when the write to a redundancy group is complete, the specification actually discloses two embodiments, one requiring write to cache, and one requiring write to disks. The fact that the specification discloses both options, even in two locations
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

11

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 16 of 29

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

purporting to describe the same thing, demonstrates that when the host processor receives a "write complete" message has nothing to do with how the patent defines "cache memory" and "redundancy group." Sun's construction would therefore require the Court to redefine both "redundancy group" and "cache memory" giving them a different meaning for the purpose of these claims than they have in the specification. This is exactly what Sun has been telling the Court it cannot do. Sun's construction depends on the idea that the "single embodiment" in the specification must be covered by all the claims. But this purported "single embodiment" includes multiple

"redundancy groups" and a "cache memory" that are not the same thing. By asking the Court to equate these separate things, Sun is in fact seeking claims that do not cover the very embodiment it says must be covered by all the claims. Sun's construction purports to respect the "sole embodiment" of the specification by changing the definitions provided by the specification in order to protect an order of operations that the specification does not require. In order to avoid performing steps in an order which the claims require, Sun would completely redefine the nature of the steps ­ turning them into something they never were, and the specification requires them not to be. NetApp's proposed construction will assist the jury by clarifying the meaning of "in response to" / "responsive to," that writing a data record to a redundancy group means writing the data record to the disks that make up that redundancy group, and that a "redundancy group" refers to a single redundancy group. NetApp's construction also gives effect to the requirement that the receipt of data records precede the selection of space for writing them. For these reasons, the Court should adopt NetApp's proposed construction.

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

12

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 17 of 29

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 "Completing [a] Write Operation With Respect to [a] Processor"

C.

U.S. PATENT NO. 5,749,095

1. "Completing [a] Write Operation..." (Claims 1, 11, and 17) Term "Completing [a] Write Operation Within [a] Local Processing Node" NetApp's Construction Transferring the write data from an initiating processor to a system interface. Sun's Construction Sun contends this phrase does not require construction because the phrase is clear on its face. However, if the Court decides the construe this phrase, the phrase should be construed to mean: "data for the write operation (1) is provided to subsequent read operations within the local processing node to the same address as the write operation and (2) is or will be coherent within distributed shared memory." Sun contends this phrase does not require construction because the phrase is clear on its face. However, if the Court decides the construe this phrase, the phrase should be construed to mean: "data for the write operation (1) is provided to subsequent read operations by a processor to the same address as the write operation and (2) is or will be coherent within distributed shared memory."

Transferring the write data from an initiating processor to a system interface.

As explained in NetApp's opening brief, when properly construed in view of the written description of the '095 patent, a write operation is complete when the data is transferred to the system interface. In contrast, Sun's position is at odds with the intrinsic evidence and the law. As a result, the Court should adopt NetApp's proposed constructions. a. Sun's Argument That The Claims Are "Clear On [Their] Face To [A] Person Of Ordinary Skill In The Art" Is Irrelevant

Sun first proposes that the Court provide the disputed claim language to the jury without construction because "[t]he meaning of completing a write operation is clear to a person of ordinary skill in the art." Sun Opp. at 26 (emphasis added). As support for this argument, Sun cites the declaration of its expert (Dr. Donald Alpert), providing his understanding of the claim language. Id. Sun, however, cites no legal support for the proposition that claim construction is unnecessary if the claim language would be understood by a skilled artisan, nor is any support to be found in the case law. To be sure, the Federal Circuit has stated that for some "commonly

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

13

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 18 of 29

1 2 3 4 5 6 7 8 9

understood words" claim construction may be resolved by lay judges using "the application of the widely accepted meaning of [those] words." Phillips v. AWH Corp., 415 F.3d 1303, 1314 (Fed. Cir. 2005). Sun, however, does not argue that the phrase at issue here is "commonly understood" such that it could be applied by a jury, but instead merely that its expert purports to comprehend the phrase's meaning. Sun's argument, therefore, is without merit. b. The "Provided To Subsequent Read Operations . . . To The Same Address" Requirement of Sun's Alternative Construction Lacks Support In Either The Claim Language Or The Specification Of The '095 Patent

Falling back from its argument that the claim language is supposedly "clear on its face," Sun proposes two separate two-part definitions for the claim terms recited in claims 1, 11, 10 and 17. Part (1) of Sun's proposed constructions define write completion as "[when] data for the 11 write operation . . . is provided to subsequent read operations . . . to the same address as the write 12 operation." The language of the claims, however, provide no support for this construction. In 13 fact, the word "read" does not appear in any of the claims of the '095 patent, nor is the ability to 14 read data stated or even implied to be a necessary predicate to the completion of a write. 15 The only supporting evidence that Sun cites consists of extrinsic evidence: (1) the 16 opinion of its expert and (2) two articles published years before the filing date of the '095 patent 17 by individuals having no relation to the '095 patent's named inventor. The Federal Circuit has 18 repeatedly cautioned against reliance upon extrinsic evidence to improperly alter the meaning 19 provided by the intrinsic evidence. See, e.g., Phillips, 415 F.3d at 1318. Here, Sun's proposed 20 construction ignores the intrinsic record. 21 specification of the '095 patent establishes the centrality of the system interface to the claimed 22 invention and provides clear support for its proposed construction ­ that write completion occurs 23 when the data is transferred to the system interface. For example, the patent teaches: "[d]uring 24 step 322, the data is received and stored by system interface 24. The write operation is thereby 25 complete with respect to the initiating processor 16." See Walter Decl. Exh. C at 28:48-51 26 (emphasis added); see also id. at 29:47-49 ("Since the fast write stream transactions are 27 completed from processors 16 by storing the transaction into SMP in queue 94 [of the system 28
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

As explained in NetApp's opening brief, the

14

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 19 of 29

1 2 3 4 5

interface shown in FIG. 3]"). The opinion of Sun's expert combined with vague references to unrelated articles cannot alter the proper construction as provided by the intrinsic record. c. The "Is Or Will Be Coherent" Requirement Of Sun's Alternative Construction Is Unsupportable

Part (2) of Sun's proposed constructions define write completion as "[when the data] is or will be coherent within distributed shared memory." Again, there is nothing in the 6 context of the claim language to support Sun's construction. 7 expressly recite the timing relationship between write completion and coherency operation, 8 providing that one class of writes complete prior to coherency and the other class of writes 9 complete subsequent to coherency. Sun's "is or will be coherent" proposal is, at best, redundant 10 of the express language of the claims, and is at worst, vague and confusing as to temporal 11 relationship between coherency and write completion. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 d. Sun's Criticism Of NetApp's Proposed Construction Is Logically Flawed To the contrary, the claims

Sun next criticizes NetApp's proposed construction based on the argument that "[t]he specification is replete with examples of write operations that are completed `subsequent to' the completion of the coherency activity." Sun Opp. at 27 (emphasis added). In effect, Sun urges that by requiring transferal to the system interface, NetApp's construction of write completion is inconsistent with the disclosed invention. As discussed above, the specification of the '095 patent expressly equates write completion with transferal to the system interface. Indeed, Sun admits that asserted claim 11 includes "an express limitation providing that the `system interface' `completes' the write operation." Sun Opp. at 29. Thus, both the specification and claims support NetApp's construction. Rather than cite to actual support in the specification, Sun instead uses flawed logic in an attempt to manufacture such support. Sun argues that "[a] write operation may be completed without transferring data to other caches in the system by ensuring that any outdated copies are `invalidated.'" Id. at 28. The logical flaw in Sun's argument is that it assumes without any support that a write is completed before its data is transferred, i.e., Sun's argument relies on its own conclusion as a premise. This is confusion, not a logical argument. When Sun's flawed
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

15

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 20 of 29

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

premise is removed, NetApp's proposed construction is entirely consistent with the language of the claims ­ a write request is complete when its data is transferred to the system interface. Using this definition, Sun's example of a write that is stored locally for a period of time rather than being transferred immediately to the system interface simply is not yet completed. e. Sun Misapplies The "Rules of Claim Construction"

Finally, Sun purports to apply a number of "rules of claim construction," most of which amount to no more than variations on the doctrine of claim differentiation. That doctrine creates a presumption that "each claim in a patent has a different scope." Sinorgchem Co. Shandong v. Int'l Trade Comm'n, 511 F.3d 1132, 1139 (Fed. Cir. 2008). As above, see supra at 2-3, Sun misapplies this doctrine. As to Sun's comparisons of independent claims 1, 11, and 17, NetApp's proposed construction does not render any of those claims of identical scope, and therefore, does not implicate the doctrine. Instead, Sun argues that, because the claims use different language to describe a specific aspect of the disclosed invention, the relevant portions of the claim must be construed to have a different scope. The Federal Circuit has made clear, however, that "that two claims with different terminology can define the exact same subject matter." Curtiss-Wright Flow Control Corp. v. Velan, Inc., 438 F.3d 1374, 1380 (Fed. Cir. 2006). Further, even if NetApp's proposed construction did result in multiple claims having identical scope, that would not end the inquiry. Instead, it would still be necessary to look to the intrinsic evidence to determine the proper claim scope. Andersen Corp. v. Fiber

Composites, LLC, 474 F.3d 1361, 1370 (Fed. Cir. 2007). Here, the specification unambiguously equates write completion with transferal to the system interface, which would negate Sun's purported "presumption." Id. Finally, as to Sun's argument that NetApp's construction reads limitations into the claims, Sun itself seeks to construe write completion on the basis of when the written data can be read. That construction has no support in the claim language, thereby violating the very rule Sun purports to enforce by grafting a read requirement into a claim that includes no such language.

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

16

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 21 of 29

1 2 3 4 5 6 7 8 9 10 11 12 "Element [of a] Communication"

D.

U.S. PATENT NO. 6,873,630

1. "Portion" / "Element" (Claims 3, 5, 8, 13-15, 21-23, 45-48, 50, 52, 73, 76, 79, 89, 113, 114, 116, 117) Term "Portion [of a] Communication" NetApp's Construction The fraction or portion of a frame carried by one channel. A portion (e.g., a byte) of a miniframe that is individually encoded for transmission across one of a plurality of logical channels, where a mini-frame is a fraction or portion of a communication received from or sent to a media access control layer from a physical layer device and that is carried by one channel. Sun's Construction Sun contends this phrase does not require construction because the phrase is clear on its face. Sun contends this phrase does not require construction because the phrase is clear on its face.

The '630 patent pertains strictly to a technique for striping data communications below the MAC layer. Indeed, the patent disclaims striping above the MAC layer, explaining the 13 shortcomings of such an approach and distinguishing it as prior art. See Walter Decl. Exh. D at 14 8:22-41. By requiring "portion [of a] communication" to be a fraction of a "frame," which the 15 patent defines as being the unit of information received from a MAC layer, NetApp's 16 construction reflects these teachings that the purported invention is directed to striping below the 17 MAC layer. 18 Consequently, the Court should adopt NetApp's proposed construction. 19 20 21 communication" and "element of a communication" require no construction. Indeed, Sun even 22 admits the word "element" can have different meanings from one claim to the next. See Sun Opp. 23 at 43 ("NetApp's proposed construction conflates at least two different uses of `elements' in the 24 claims."). If so, how would a jury know which "plain and ordinary meaning" to use? 25 More importantly, the Federal Circuit has explained that even seemingly simple 26 words must be construed when reliance on their ordinary meaning fails to resolve the parties' 27 claim scope dispute. See O2 Micro International Ltd. v. Beyond Innovation Technology Co., 521 28
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

Sun's position does not comport with the intrinsic evidence or the law.

a.

The Claim Terms Require Construction

Sun's own arguments belie the notion that the vague claim terms "portion of a

17

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 22 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13

F.3d 1351, 1361-1362 (Fed. Cir. 2008) (district court erred by failing to construe "only if" to determine whether that term permitted a commonly understood exception). The issue here-- whether the claims cover data striping occurring above the MAC layer--cannot be resolved by reference to the terms' so-called "ordinary" meanings. NetApp's construction for "portion [of a] communication" makes clear that the claims of the '630 are limited to data striping that is done on data frames that are already packaged by the MAC layer; Sun disagrees with this view. See Sun Opp. at 41 ("NetApp's addition of the italicized language [requiring a portion of a communication to be received from or sent to a MAC layer] is wrong[.]"). This important claim interpretation dispute simply cannot be resolved by the jury by reference to vague claim terms such as "portion of a communication." It should thus be resolved by the Court. b. The '630 Patent Claims Should Be Limited To Data Striping Below The MAC Layer

Sun holds fast to its position that the claims cover data striping above the MAC layer. See Sun Opp. at 37 ("[S]ome (but not all) claims require the communication be received 14 from a MAC module[.]"). NetApp's proposed construction corrects this misunderstanding and 15 clarifies that apportionment of the "communication"--whether it be a data stream or a single 16 17 layer, into mini-frames. Acampora Decl. at 3-4. 18 Sun's position to the contrary is untenable, as the patent disparages, distinguishes, 19 and plainly disavows above-the-MAC data striping: 20 21 22 23 24
6

frame6--is accomplished by subdividing each individual "frame," already packaged by the MAC

As one skilled in the art will appreciate, striping data across multiple channels may be performed at different levels of a network protocol stack. For example, if implemented above the MAC layer (e.g., as with 802.3 link aggregation) multiple network "flows" or "conversations" must be distributed and collected and almost all of the network interface hardware . . . must be duplicated. In addition, the speed of an individual flow during such "flow striping" is limited to the speed of an individual channel. Contrary to Sun's repeated contentions, see, e.g., Sun Opp. at 33-38, 40-41, NetApp has never disputed that the claimed methods cover the transmission of multiple frames. The intent of NetApp's proposed construction, rather, is to make clear that apportionment of the "communication"--whether it be a data stream or a single frame--is accomplished by subdividing each individual frame, already packaged by the MAC layer into mini-frames. In other words, NetApp's contention is that "portion of a communication" comprises fractions of "frames," but not necessarily fractions of "a" frame. Thus, Sun's quibble regarding the word "a" misses the point entirely.
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

25 26 27 28

18

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 23 of 29

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 · In contrast, one or more embodiments of the invention discussed herein perform striping of network data at a lower level in the network protocol stack. . . . Because striping is done with the contents of individual MAC frames or packets in these embodiments, only Physical layer resources need to be duplicated. See Walter Decl. Exh. D at 8:22-41. First, this passage admits above-the-MAC data striping as prior art. Techniques such as link aggregation were not only known at the time, but were already being standardized by the IEEE as part of the Ethernet protocol. Second, this passage disparages above-the-MAC striping by explaining that such techniques (1) do not increase data bandwidth because they still limit transmission to a single logical channel, and (2) require duplication of equipment in higher levels of the network protocol stack. Finally, the language leaves little doubt that the patent distinguishes below-the-MAC data traffic ("frames" and "mini-frames") from above-the-MAC data traffic ("flows" and "conversations"), and disclaims the latter ("In contrast, one or more embodiments of the invention discussed herein perform striping of network data at a lower level in the network protocol stack."). "Where the general summary or description of the invention describes a feature of the invention . . . and criticizes other products . . . that lack that same feature, this operates as a clear disavowal of these other products." AstraZeneca AB, Aktiebolaget Hassle, KBI-E, Inc. v. Mutual Pharmaceutical Co. Inc., 384 F.3d 1333, 1339-1340 (Fed. Cir. 2004); see also On Demand Machine Corp. v. Ingram Industries, Inc., 442 F.3d 1331, 1340 (Fed. Cir. 2006) ("When the scope of the invention is clearly stated in the specification, and is described as the advantage and distinction of the invention, it is not necessary to disavow explicitly a different scope."). Sun's request that the patent be understood to encompass such disclaimed features should thus be rejected. Indeed, as NetApp explained in its opening brief, every affirmative teaching in the specification confirms this disavowal of claim scope and re-affirms that the invention is strictly limited to below-the-MAC data striping. For example: · The abstract describes the alleged invention as an "Ethernet" (a layer 2 protocol) architecture that works by manipulating "frames," which is only possible below the MAC layer. Walter Decl. Exh. D at Abstract. Every figure in the specification depicts the distributor / collector module, responsible for striping and reassembling data, to be below the MAC layer. See e.g., id. at Fig. 1.

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

19

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 24 of 29

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

·

The specification describes the invention in terms of "frames" and "miniframes"--referring to the latter 45 times--confirming that the patent's teachings are confined below the MAC layer. See, e.g., id. at 5:48-51, 5:51-53. Against the backdrop of these clear teachings in the specification, Sun even agreed

to a number of constructions that comport with the understanding that the '630 patent pertains strictly to below-the-MAC data striping. For instance, Sun has agreed that the "distributor" divides "Ethernet frame[s] received from a MAC module." See Nathan Decl. Exh. A. Likewise, Sun has agreed that the "collector" module "reassembles an Ethernet frame to be provided to a MAC layer." Id. Having agreed to these constructions, Sun's position that some claims may cover above-the-MAC striping is puzzling indeed. Nevertheless, claims simply cannot be broader in scope than the invention that is set forth in the specification. See On Demand, 442 F.3d at 1340. Having disclaimed above-the-MAC data striping and having affirmatively focused on the opposite, the '630 patent is simply not entitled to claim above-the-MAC striping. NetApp's construction of the claim term "element [of a] communication" reinforces the below-the-MAC limitation by making clear that an "element" is a further subdivision of a "frame." The specification firmly supports this. Indeed, figures 5A to 5D--the very figures that Sun's expert relies on--makes clear that "elements" are in fact the bytes in a mini-frame. These figures (a portion of Fig. 5B appears below) show four separate streams of data, each of which is depicted as a solid line, that result from striping the data stream shown in Figure 4. Id. at 13:14-17.

Id. at Fig. 5B. In the bottom row in Figure 5B, each hexagon, except for the idle codes, represents one byte. Id. at 13:39-46. The mini-frames are separated by the idle codes. Id. at 13:47-50. Thus, the first mini-frame in Figure 5B comprises the series of hexagons (i.e., bytes) from "SMD" ("Start of Mini-frame Delimiter") to "EMD" ("End of Mini-frame Delimiter"). Id. at 11:6-10. The row immediately above, depicting the same data stream, shows numbered bytes
NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

20

Case No. C-07-06053 EDL

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 25 of 29

1 2 3 4 5 6 7 8 9 10 11 12

from the original data stream cycling through the figures in a round-robin fashion--preamble byte P1 is in Figure 5A, preamble byte P2 is in Figure 5B, . . . and so on. According to these figures, then, it is clear that each individual byte corresponds to an "element" described in 6:67-7:6, because it is these individual bytes that are being "distributed among multiple logical channels . . . on a round robin basis." The specification confirms the drawings, teaching that the elements are distributed among multiple channels to form a mini-frame on each channel: In short[,] individual frame elements (e.g., bytes) are distributed among multiple logical channels . . . on a round-robin basis. Each channel thus carries one "miniframe" or "mini-packet," the contents of which will be reunited with those of the other mini-frames at the receiving entity. Id. at 6:67-7:6. d. Sun's Claim Differentiation Arguments Fail

In opposing NetApp's constructions, Sun repeatedly raises claim differentiation, stating, for instance, that NetApp's construction of "portion [of a] communication" "eviscerates 13 any distinction" in the meanings of different claims because some dependent claims actually use 14 the word "frame" instead of "communication." But claim differentiation is by no means a rigid 15 rule, and "practice has long recognized that claims may be multiplied . . . to define the metes and 16 17 18 19 20 21 22 23 24 25 26 27
7

bounds of the invention in a variety of different ways.7" Id. The Federal Circuit has explained: Different claims with different words can, of course, define different subject matter within the ambit of the invention. On the other hand, claim drafters can also use different terms to define the exact same subject matter. Indeed this court has acknowledged that two claims with different terminology can define the exact same subject matter. . . . In this context, this court has cautioned that claim differentiation is a guide, not a rigid rule. Curtiss-Wright Flow Control Corp. v. Velan, Inc., 438 F.3d 1374, 1380-1381 (Fed. Cir. 2006) (internal citations omitted). And, one thing is crystal clear: claim differentiation cannot broaden a claim beyond its proper scope. See id. at 1381 (vacating a district court's claim construction because its reliance on claim differentiation contradicted the meaning of the claim and was inconsistent with the overall context of the invention).

28

Indeed, this appears to be precisely the strategy adopted by the `630 patent, which, though disclosing only a single embodiment to explain the invention, employs 125 claims to claim it. 21
Case No. C-07-06053 EDL

NetApp's Reply to Sun's Responsive Claim Construction Brief
SV1:\297394\09\6dgy09!.DOC\65166.0004

Case 3:07-cv-06053-EDL

Document 128

Filed 08/01/2008

Page 26 of 29

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

Again, Sun's reliance on claim differentiation is simply misplaced. With respect to claim 2, Sun argues that NetApp's construction renders it superfluous because it ostensibly limits the term "communication" in its parent claim to an "Ethernet frame." Sun Opp. at 35. But Sun is wrong. Not all frames are Ethernet frames. Indeed, the '630 patent itself defines the word "frame" to be broader than just Ethernet frames. Walter Decl. Exh. D at 5:47-50. As such, equating "communication" with "frame" does not render any claim superfluous. Claim 1

includes within its scope all frame-based layer 2 protocols such as token ring, whereas claim 2 is limited to just Ethernet. This interpretation makes perfect sense in light of the background section of the patent, which lists several layer 2 architectures before acknowledging Ethernet as the most advantageous of the bunch. Id. at 1:54-2:9. For similar reasons, Sun's claim differentiation attacks against NetApp's "element [of a] communication" construction are without merit. No dependent claim is made completely superfluous by NetApp's construction. True, dependent claim 53 uses the term "mini-frame." Id. at 21:56-61. But its parent claim 25 does not use the disputed