DataTern, Inc. v. Blazent, Inc.

• Court: District Court, D. Massachusetts
• Decided: June 5, 2018
• Docket: 1:11-cv-11970
• Status: Unknown

Opinion

UNITED STATES DISTRICT COURT DISTRICT OF MASSACHUSETTS

_______________________________________ ) Civil Action Nos. DATATERN, INC., ) 11-11970-FDS (Lead) ) 11-12220-FDS (Consolidated) Plaintiff, ) 11-12024 ) 11-12025 v. ) 11-12026 ) 11-12223 MICROSTRATEGY INC. et al., ) 11-12225 ) 11-12227 Defendants. ) _______________________________________)

ORDER ON DEFENDANTS’ MOTIONS FOR AN EXCEPTIONAL CASE DETERMINATION AND ATTORNEYS’ FEES SAYLOR, J. This was a case for patent infringement. Plaintiff Datatern, Inc. sued twenty-three defendants in twenty-two separate lawsuits in this District for infringement of U.S. Patent 6,101,502, titled “Object Model Mapping and Runtime Engine for Employing Relational Database with Object Oriented Software.” After almost six years, plaintiff’s counsel withdrew from the representation. As plaintiff was unable to retain substitute counsel within a reasonable time, the case was dismissed for failure to prosecute. Defendant MicroStrategy and the six other remaining defendants have filed motions to recover their attorneys’ fees and expert fees pursuant to 35 U.S.C. § 285. For the reasons set forth below, those motions will be granted in part and denied in part. I. Background A. Factual Background DataTern is the owner of U.S. Patent No. 6,101,502 (the “’502 patent”). The inventors of the ’502 patent filed the utility application on September 25, 1998, claiming priority to provisional application number 60/069,157 (filed on December 9, 1997), and provisional application number 60/059,939 (filed on Sept. 26, 1997). The U.S. Patent and Trademark Office issued the patent on August 8, 2000.1 The invention claimed in the ’502 patent facilitates interaction between two popular systems for organizing computerized data: object-oriented software applications and relational

databases. (’502 patent, col. 1 ll. 22-24). Object-oriented software applications encapsulate information in a collection of discrete “objects” that correspond to “classes,” which define the type of object. (11-cv-11970, ECF 130-1 ¶¶ 20, 22). For example, an object-oriented software application for a human-resources department might contain the class “employee,” which corresponds to objects representing particular employees such as “Jane Brown.” The object might contain attributes concerning Jane’s employment, such as her wage rate and scheduled hours. Relational databases organize information into rows and columns, with each column representing an attribute and each row representing an instance of those attributes. (Id. ¶ 29). To use the same example, the database would display a table with columns containing information

about employees’ wages and hours, and rows representing a particular employee, such as Jane. The so-called “object-relational mismatch” arises because of different assumptions and approaches underlying the two systems. (Id. ¶ 35). The ’502 patent addresses the mismatch by generating intermediaries to translate between the systems, making the interaction easier and more reliable. (Id. ¶¶ 42-43). Representative Claim 1 reads: A method for interfacing an object oriented software application with a

1 In 2007, the PTO reexamined the ’502 patent following a request by a third party for ex parte reexamination. The PTO ultimately confirmed the patentability of the original Claims 1-18 and allowed new Claims 19-44 in 2009. relational database, comprising the steps of: selecting an object model; generating a map of at least some relationships between schema in the database and the selected object model; employing the map to create at least one interface object associated with an object corresponding to a class associated with the object oriented software application; and utilizing a runtime engine which invokes said at least one interface object with the object oriented application to access data from the relational database. ’502 patent col. 7 l. 51-col. 8 l.3. DataTern contended that MicroStrategy’s Business Intelligence Platform infringes the ’502 patent, and that the infringement extended to MicroStrategy’s customers, at least some of whom are consolidated defendants in this case. B. Procedural Background On November 7 and 8, 2011, DataTern filed eight lawsuits against eight different defendants, all customers of MicroStrategy, alleging infringement of the ’502 patent. All of those cases except that against Blazent, Inc., were voluntarily dismissed by DataTern after six weeks. Shortly thereafter, on November 15, 2011, DataTern filed nine more lawsuits against ten more customers of MicroStrategy, again alleging infringement of the ’502 patent. Five of those cases—those not assigned to Judge Stearns—were voluntarily dismissed by DataTern within a few months. Then, on December 14, 2011, DataTern filed similar complaints against four other MicroStrategy customers and MicroStrategy itself. It marked all these cases as “related” to one of the November 15 suits before Judge Stearns, and all five were thus assigned to Judge Stearns. On February 23, 2012, MicroStrategy filed a motion to intervene and stay in the then- pending cases against its customers before Judge Stearns, on the theory that the case against it would be all but dispositive as to the cases against its customers. (e.g., No. 11-cv-12024, ECF 12). On February 24, 2012, Judge Stearns entered an order consolidating those nine cases and naming case No. 11-cv-12220, that against MicroStrategy, as the lead case. He then denied the

motion to intervene and stay as moot. At that time, the case naming Blazent as the defendant was not yet consolidated. The Court entered a scheduling order on March 21, 2012, and gave DataTern a deadline of June 1, 2012, to indicate whether it intended to pursue claims of infringement against MicroStrategy’s customers separate from their use of MicroStrategy’s software. A few months later, on April 26, 2012, DataTern filed a notice with the Court explaining that it did not intend to pursue claims of patent infringement against MicroStrategy’s customers that were unrelated to their use of MicroStrategy’s software, and moved to voluntarily dismiss those claims without prejudice pursuant to Fed. R. Civ. P. 41(a). (No. 11-cv-12220, ECF 42).

Five of MicroStrategy’s customers opposed the dismissal and cross-moved for summary judgment, arguing that because DataTern had not filed infringement contentions on the customer claims in accordance with the Court’s scheduling order, it had waived those claims and they should be dismissed with prejudice. (No. 11-cv-12220, ECF 45). That motion also requested that all claims against MicroStrategy’s customers should be stayed until the claims between MicroStrategy and DataTern were resolved. On May 10, 2012, MicroStrategy moved for judgment on the pleadings on the ground that the patent was drawn to patent-ineligible subject matter. (11-cv-12220, ECF 43). On May 25, 2012, MicroStrategy moved to compel DataTern to produce adequate infringement contentions. Judge Stearns granted that motion in part and denied it in part, requiring DataTern to provide amended contentions that “clearly identif[ied] the accused products (or combinations of products) that is alleged to infringe each asserted claim, and where each asserted claim element is found within each accused product or combination and supporting evidence” but explaining that “[a]t this preliminary stage, it is not necessary for plaintiff to

provide highly detailed or ultimately successful contentions” and that “[p]laintiff is not required to know or disclose all possible theories of infringement at this point, and may uncover and disclose additional theories of infringement in the course of discovery.” (No. 11-cv-12220, Electronic Order, July 2, 2012). DataTern filed amended contentions on July 20, 2012. (No. 11- cv-12220, ECF 72). On July 5, 2012, Judge Stearns denied the customers’ motion for summary judgment and allowed the dismissal of the claims without prejudice. As to the motion to stay, he explained that “the court will not entertain inefficient and duplicative multiple proceedings, and will stay the cases against the customer-defendants if and only if every customer-defendant agrees to be

bound by the adjudication as to liability as to MicroStrategy.” (11-cv-12220, Electronic Order, July 5, 2012). In the next few weeks, all the customer-defendants agreed to be so bound, and the case as to them was stayed on July 17. Judge Stearns denied MicroStrategy’s motion for judgment on the pleadings on July 31, 2012, without prejudice to its renewal after claim construction, explaining that “the patentability issue in this case turns on the significance of certain claim elements” and that it would be prudent to construe the claims first. (11-cv-12220, Electronic Order, July 31, 2012). That same day, the case was randomly reassigned to the undersigned judge. (11-cv- 12220, ECF 78). MicroStrategy made its source code available to DataTern on August 10, 2012. (See 11- cv-11970, ECF 156-2). Pursuant to the parties’ joint scheduling statement, DataTern had proposed that it would supplement its infringement contentions within 90 days of that production with specific citations to the source code. (11-cv-12220, ECF 27 at 8). Although that part of the joint scheduling statement had not been explicitly adopted by the Court, DataTern represented

that it needed the source code to provide more accurate infringement contentions and that it would so supplement its contentions. (11-cv-11970, ECFs 156-1, 156-4). Meanwhile, the ’502 patent was also the subject of litigation in the United States District Court for the Southern District of New York. See Microsoft Corporation v. Datatern, Inc. (11- cv-02365-KBF) and SAP AG and SAP of America v. DataTern, Inc. (11-cv-02648-KBF). That litigation involved DataTern, but not MicroStrategy. On August 24, 2012, the New York court issued an order on claim construction in which it construed the terms “object model” and “to create at least one interface object,” among others in the ’502 patent. It construed the term “object model” to mean “a template with a predetermined standardized structure both relating to

an object-oriented software application and including object classes and inheritance relationships among classes.” Microsoft Corp. v. DataTern, Inc., Nos. 11-cv-02365, 11-cv-02648, 2012 WL 3682915, at *14 (S.D.N.Y. Aug. 24, 2012). It construed the term “to create at least one interface object” to mean “to generate code for at least one class and instantiate an object from that class, where the object is not part of or generated by the object oriented application and is used to access the database.” Id. At that point, in September 2012, MicroStrategy threatened to request sanctions against DataTern if it did not immediately dismiss the case the prejudice. In response, DataTern filed a motion to stay the litigation in Massachusetts pending the entry of final judgment in the New York cases. (11-cv-12220, ECF 83). The scheduling order in place at that time would have required that opening claim-construction briefs be filed by October 1, 2012. DataTern proposed that that exercise was unnecessary in light of the fact that the terms construed by the New York court were sufficient to show noninfringement. MicroStrategy contended that the case was baseless, its continuance was prejudicial to its business, and that claim construction should move

forward so that the Federal Circuit could address all the disputed claim terms at once (as there were disputed terms the New York court had not reached). (11-cv-12220, ECF 86). The Court stayed the case on October 5, 2012, pending the issuance of final judgment in New York. At that hearing, DataTern disputed its obligation to supplement its infringement contentions within 90 days of the source code being produced, contending that its July 20, 2012 contentions were adequate to allow the parties to proceed to claim construction and that citations to source code were not normally provided until the expert reports. (11-cv-12220, ECF 98 at 14- 18). The Court noted that the scheduling order did not require such a supplementation, and permitted DataTern to inspect the source code during the stay. (Id. at 18-19, 32-33). The Court

deferred ruling on whether DataTern was required to file amended infringement contentions including citations to the source code, explaining that it was “not prepared to make a final binding decision right now” although it might be “the first sentence of [MicroStrategy’s] first motion” after the lift of the stay. (Id. at 36-37). The New York court issued a final judgment in the consolidated cases before it on December 26, 2012. (11-cv-12220, ECF 105 at 4). On January 4, 2013, the Court granted a joint motion to consolidate the cases consolidated under 11-cv-12220 with the case against Blazent; at that point, case number 11-cv- 11970-FDS became the lead case. Blazent also agreed to be bound by the adjudication as to liability against MicroStrategy. (11-cv-11970, ECFs 25, 26). On January 24, 2013, DataTern appealed the judgment of the New York court to the Federal Circuit. On February 4, 2013, DataTern filed a motion for judgment in favor of MicroStrategy as a matter of law in this Court, conceding that if the New York court had correctly construed the

term “to create at least one interface object,” the accused MicroStrategy product could not be held to infringe the ’502 patent. Specifically, under that construction, the accused product here would not meet the claim limitation to “create at least one interface object” because it does not “generate code for at least one class and instantiate an object from that class.” (No. 11-cv- 11970, ECF 30). Although DataTern had offered to stipulate to judgment in MicroStrategy’s favor, MicroStrategy did not accept that stipulation. Instead, MicroStrategy filed a motion for summary judgment of noninfringement and attorneys’ fees, detailing five separate grounds— over and above DataTern’s concession—on which it believed it deserved summary judgment. (No. 11-cv-11970, ECF 32).

On February 7, 2013, the Court granted summary judgment to MicroStrategy based on the concession by DataTern that MicroStrategy could not be held to have infringed the ’502 patent based on the New York court’s construction of “create at least one interface object.” (No. 11-cv-11970, ECF 39). It declined to address the other grounds MicroStrategy advanced, and deferred consideration of the fees question until after any decision by the Federal Circuit. DataTern then appealed the judgment in this case. On May 5, 2014, the Federal Circuit upheld the New York court’s judgment of non- infringement. Microsoft Corp. v. DataTern, Inc., 755 F.3d 899, 909 (Fed. Cir. 2014). In so doing, it upheld the New York district court’s determination that the term “object model” required classes. Id. It did not, however, review the district court’s construction of “to create at least one interface object,” as that proved unnecessary to render its decision. Id. at 908 n.5. On December 19, 2014, the Federal Circuit vacated this Court’s order of summary judgment, finding that the New York court’s construction of the term “to create at least one interface object” (which this Court had essentially adopted) was incorrect. DataTern, Inc. v.

Epicor Software Corp., 599 F. App’x 948, 954-55 (Fed. Cir. 2014). It construed the term “to create at least one interface object” to mean “to instantiate at least one interface object from a class.” Id. The Federal Circuit remanded the case to this Court for further proceedings. Id. at 955. On February 11, 2015, MicroStrategy filed two separate motions for summary judgment, one on the basis of invalidity for non-patentable subject matter and the other on the basis of non- infringement. (11-cv-11970, ECFs 49, 53). On September 4, 2015, the Court denied both motions, explaining that the ’502 patent was not drawn to patent-ineligible subject matter and that the motion for summary judgment was premature because it depended on the construction of

the term “class.” (11-cv-11970, ECF 101). Although the Court was troubled by the fact that DataTern had taken inconsistent positions as to the construction of the term “class” in this litigation and the New York litigation (in which it had stipulated to a different construction than the one it advanced here), it declined to hold that DataTern was estopped because the matter had not been fully briefed. On February 19, 2016, MicroStrategy filed a motion to require DataTern to post a $2.5 million bond to cover its expected attorneys’ fees in the event that MicroStrategy prevailed and was awarded fees. (11-cv-11970, ECF 103). The Court denied that motion without prejudice on March 9, 2016, on the ground that the “[l]ittle has changed since the Court denied summary judgment in September 2015: the parties still have not begun claim construction and the judicial estoppel argument on the term ‘class’ remains undeveloped.” (11-cv-11970, ECF 107 at 9). On July 29, 2016, MicroStrategy filed a motion to compel production of documents from prior litigation involving the ’502 patent. (11-cv-11970, ECF 132). On August 5, 2016,

DataTern filed a motion to compel production of MicroStrategy’s source code, which MicroStrategy argued it should not have to do because DataTern had had an opportunity to view the source code back in 2012 and forfeited it. (No. 11-cv-11970, ECFs 140, 156). On September 7, 2016, Magistrate Judge Dein granted in part and denied in part MicroStrategy’s motion, ordering MicroStrategy to obtain as many of the documents as possible from publicly available sources and accepting DataTern’s representations as to the non-existence of certain statements, but requiring DataTern to provide hard copies of anything not publicly available and allowing MicroStrategy to file a further motion to compel if there were additional documents it thought it deserved. Judge Dein also granted DataTern’s motion to compel MicroStrategy to

produce its source code. (11-cv-11970, ECF 169). On September 26, 2016, the Court held a Markman hearing on the seven disputed claim terms. The Court issued its Markman order on February 7, 2017. (11-cv-11970, ECF 204). The Court ultimately agreed with MicroStrategy’s construction of “class,” but, the matter having been fully briefed, found that MicroStrategy had failed to demonstrate that DataTern would derive an unfair advantage from its position being adopted by the Court, and therefore DataTern was not judicially estopped from asserting its position on the proper construction of the term “class.” Following that order, DataTern filed amended infringement contentions in March 2017. (11-cv-11970, ECF 206). Those infringement contentions did not include citations to MicroStrategy’s source code. In April 2017, MicroStrategy filed amended invalidity and noninfringement contentions. (11-cv-11973, ECF 210). It then filed a motion to compel adequate infringement contentions from DataTern in June 2017. (11-cv-11973, ECF 212). Before the Court could rule on that motion, DataTern’s counsel file a motion to withdraw

on August 22, 2017. Therein, DataTern’s counsel explained (obliquely, to avoid violating the attorney-client privilege), that their representation agreement “provides that the client will consent to [counsel]’s withdrawal upon the occurrence of certain circumstances that include material disagreement over the conduct of the case, non-payment of expert fees and non-payment of other disbursements” and that one or another of those circumstances had come to pass. (Docket No. 227 at 1-2). On August 28, 2017, the Court granted the motion to withdraw effective September 18, 2017, in order to give DataTern the opportunity to find substitute counsel. The Court later extended that deadline to October 20, 2017. DataTern was not able to find substitute counsel, and, because a corporation is not permitted to proceed pro se, the Court

dismissed the action for failure to prosecute on that date. MicroStrategy requested leave to file another motion for summary judgment, which the Court denied. (11-cv-11970, ECF 241). Subsequently, MicroStrategy filed this motion for fees. The other six remaining defendants have filed follow-on motions for fees incurred prior to the cases being consolidated. II. Standard of Review “The court in exceptional cases may award reasonable attorney fees to the prevailing party” in a patent dispute. 35 U.S.C. § 285. In order to be a “prevailing party,” a party must have received “relief on the merits.” Highway Equip. Co. v. FECO, Ltd., 469 F.3d 1027, 1033- 34 (Fed. Cir. 2006). The “‘touchstone of the prevailing party inquiry must be the material alteration of the legal relationship of the parties.’” CRST Van Expedited, Inc. v. Equal Emp’t Opportunity Comm’n, 136 S. Ct. 1642, 1646 (2016) (quoting Tex. State Teachers Ass’n v. Garland Indep. Sch. Dist., 489 U.S. 782, 789 (1989)). That change “must be marked by ‘judicial imprimatur’”; in other words, the party must have prevailed “because he has received a ‘judicially sanctioned change in the legal relationship of the parties.’” Id. (quoting Buckhannon

Bd. & Care Home, Inc. v. W. Va. Dep’t of Health & Human Res., 532 U.S. 598, 604-05 (2001)); see also Highway Equip. Co., 469 F.3d at 1033. “[A]n ‘exceptional case’ is simply one that stands out from others with respect to the substantive strength of a party’s litigating position (considering both the governing law and the facts of the case) or the unreasonable manner in which the case was litigated. District courts may determine whether a case is ‘exceptional’ in the case-by-case exercise of their discretion, considering the totality of the circumstances.” Octane Fitness, LLC v. ICON Health & Fitness, Inc., 134 S. Ct. 1749, 1756 (2014). While there is no precise formula for determining whether a case is exceptional, considerations include “‘frivolousness, motivation, objective

unreasonableness (both in the factual and legal components of the case) and the need in particular circumstances to advance considerations of compensation and deterrence.’” Id. at 1756 & n.6 (quoting Fogerty v. Fantasy, Inc., 510 U.S. 517, 534 n.19 (1994)). A party need only show that a case is exceptional by a preponderance of the evidence. Id. at 1758. Even following Octane Fitness, the fact that a case is deemed “exceptional” does not automatically entitle a prevailing party to fees for the entire litigation. Homeland Housewares, LLC v. Sorensen Research, 581 F. App’x 877 (Fed. Cir. 2014) (affirming an award of partial fees roughly connected with litigation misconduct, but declining to require the district court to strictly limit the award to fees incurred in responding to specific acts of misconduct because “it is the ‘totality of the circumstances,’ and not just discrete acts of litigation conduct, that justify the court’s award of fees”).2 Rather, the determination of the amount of “reasonable” attorneys’ fees is a matter within the Court’s discretion. Highmark Inc. v. Allcare Health Mgmt. Sys., Inc., 134 S. Ct. 1744, 1749 (2014) (holding that “an appellate court should apply an abuse-of-discretion standard in reviewing all aspects of a district court’s § 285 determination”).

Expert fees are not subject to recovery under 35 U.S.C. § 285. Rather, “a district court may invoke its inherent power to impose sanctions in the form of reasonable expert fees” in cases “with ‘a finding of fraud or abuse of the judicial process’” or “in cases involving bad faith that cannot be otherwise reached by rules or statutes.” Takeda Chem. Indus., Ltd. v. Mylan Labs., Inc., 549 F.3d 1381, 1391 (Fed. Cir. 2008). III. Analysis A. Prevailing Party Although there appears to be no precedent directly on point, defendants are surely the prevailing parties in this lawsuit. This case was dismissed involuntarily and with prejudice, meaning that plaintiff will never be able to recover from defendants on these claims.3 That is

2 The following cases, all of which precede Octane Fitness, generally approve the awarding partial fees related to the misconduct when the basis for finding a case exceptional is litigation misconduct: Cartner v. Alamo Grp., Inc., 561 F. App’x 958, 963 (Fed. Cir. 2014) (“A fee award under § 285 may only ‘compensate a party for the extra legal effort to counteract the . . . misconduct.’” (quoting Highmark, Inc. v. Allcare Health Mgmt. Sys., Inc., 687 F.3d 1300, 1316 (Fed. Cir. 2012), vacated on other grounds by 134 S. Ct. 1744 (2014))); Rambus Inc. v. Infineon Techs. AG, 318 F.3d 1081, 1106 (Fed. Cir. 2003) (“In cases deemed exceptional only on the basis of litigation misconduct, however, the amount of the award must bear some relation to the extent of the misconduct.”); Special Devices, Inc. v. OEA, Inc., 269 F.3d 1340, 1344 (Fed. Cir. 2001) (“[T]he amount of the attorney fees depends on the extent to which the case is exceptional.”); Integrated Tech. Corp. v. Rudolph Techs., Inc., 629 F. App’x 972, 976 (Fed. Cir. 2015) (“On remand, the district court must award reasonable attorneys’ fees commensurate with [defendant’s] misconduct.”); Beckman Instruments, Inc. v. LKB Produckter AB, 892 F.2d 1547, 1553 (Fed. Cir. 1989) (“Since any injustice present in this case is based upon [defendant’s] bad faith and misconduct during litigation, the penalty imposed must in some way be related to bad faith and misconduct.”). 3 It appears that one reason for plaintiff’s counsel’s withdrawal may have been non-payment of fees. (No. 11-cv-11970, ECF 227). To the extent that DataTern is unable to pay its own attorneys, of course, the exercise of determining whether it ought to be required to pay defendants’ fees is likely to be a futile one. clearly an “actual, court-ordered alteration in the legal relationship in the parties,” Chapman Law Firm Co. v. Greenleaf Constr. Co., 490 F.3d 934, 939 (Fed. Cir. 2007), which “materially alter[ed] the legal relationship between the parties by modifying one party’s behavior in a way that ‘directly benefits’ the opposing party,” Shum v. Intel Corp., 629 F.3d 1360, 1367 (Fed. Cir. 2010) (quoting Farrar v. Hobby, 506 U.S. 103, 111-13 (1992)). Dismissals with prejudice for

failure to prosecute are considered decisions “on the merits” for the purposes of claim preclusion. Int’l Nutrition Co. v. Horphag Research, Ltd., 220 F.3d 1325, 1329 (Fed. Cir. 2000). There is no obvious reason for such a disposition to operate differently in the context of a fee award. B. Exceptional Case Defendants assert four broad categories of conduct that they allege make this case “exceptional”: (1) that DataTern’s business model is to extract nuisance-value settlements and avoid testing the merits of its claims; (2) that DataTern had no good-faith theory of infringement and prolonged the litigation in an attempt to force a settlement; (3) that DataTern failed to conduct an adequate pre-suit investigation into known flaws in the ’502 patent’s inventorship;

and (4) that DataTern engaged in blatant judge-shopping. Plaintiff, in turn, accuses defendants of filing unnecessary and frivolous motions in this lawsuit and retaliating in another proceeding. These assertions will be first taken in turn, and then assessed together in the totality of the circumstances. 1. DataTern’s Business Model “[A] pattern of litigation abuses characterized by the repeated filing of patent infringement actions for the sole purpose of forcing settlements, with no intention of testing the merits of one’s claims, is relevant to a district court’s exceptional case determination under § 285.” SFA Sys., LLC v. Newegg Inc., 793 F.3d 1344, 1350 (Fed. Cir. 2015). Factors to consider in determining whether a particular suit was brought to force a settlement include the number of actions brought by the patentee, whether the settlement amounts are low in comparison to the costs of litigation, and other evidence suggesting that the patentee had no intention of testing the merits of its claims. See id. at 1350-51; Monolithic Power Sys., Inc. v. O2 Micro Int’l Ltd., 726 F.3d 1359, 1367-68 (Fed. Cir. 2013); Eon-Net LP v. Flagstar Bancorp, 653

F.3d 1314, 1327-28 (Fed. Cir. 2011). But “the fact that [plaintiff] has filed several lawsuits against numerous defendants is insufficient to render [a] case exceptional. In many cases, patent infringement is widespread and the patent owner may be forced to revert to widespread litigation against several infringing parties to enforce its intellectual property rights.” SFA Sys., 793 F.3d at 1347 (quoting the district court, which was affirmed). To support their contention that plaintiff has abused the litigation system in filing suits only for nuisance value settlements, defendants point to (1) the fact that plaintiff filed four suits in the Eastern District of Texas against dozens of defendants, none of which reached the claim- construction stage, and (2) the value of plaintiff’s previous licensing agreements, which vary, but all are less than the amount typically required to defend a patent lawsuit.4

Plaintiff has, indeed, sued many defendants for patent infringement. The Court will accept, for the purposes of this motion, that plaintiff is a non-practicing entity (although it appears to be an IP holding company for affiliated technology companies); that it has little to lose in filing suit; that it has never won a judgment on the merits (although it has won several intermediate rulings in its favor); and that it has received settlement funds from approximately four dozen defendants in the nuisance-value range. See Eon-Net, 653 F.3d at 1327-28. But the evidence does not support a finding by this Court that it had no intention of testing the merits of

4 They range from under $100,000 to, in one case, more than $1 million. (11-cv-11970, ECF 254-2). its claims in this lawsuit, which, after all, is the suit for which defendants are requesting fees. Defendants contend that plaintiff’s strategy was to delay and draw out this litigation unnecessarily, and that this is evidence that it never intended to test the merits of its claim. However, the Court does not find that this case suffered from undue delay caused by plaintiff. It is true that it took six years from the time DataTern filed its first suits against MicroStrategy’s

customers to reach a judgment in the case. But at least two of those years were dead time waiting for the appeal at the Federal Circuit. It was defendants who requested that the Court adopt the same construction as the New York court had adopted; they can hardly complain that this caused them to win, and the Court can hardly punish plaintiff for accepting that fact and offering to stipulate to non-infringement in order to pursue its appeal. A substantial portion of the remaining time prior to claim construction was spent on defendants’ motions—including the May 2012 motions for summary judgment and judgment on the pleadings (two months from filing to resolution) and two February 2015 motions for summary judgment (seven months from filing to resolution), all of which defendants lost.5 Although the filing of those motions clearly

shows that defendants would have preferred a speedy resolution, plaintiff’s oppositions cannot be considered unreasonable delay calculated to draw out the litigation. Approximately nine months elapsed between the Court’s decision on defendants’ February 2015 summary judgment motions and the beginning of the claim-construction briefing. Discovery was taking place during this period, and while there were a handful of disputes and motions to compel, each side prevailed on some, and the Court finds nothing exceptional in either party’s behavior during that time. Certainly there is nothing to suggest that plaintiff was deliberately causing delay.

5 Some of the delay, of course, was caused by the Court itself when rendering decisions on those motions. The claim-construction process began in July 2016 with the submission of the parties’ opening Markman briefs; the Markman hearing was held in September of that year, and the Court’s decision, which essentially adopted the constructions defendants desired, was issued in February 2017. Defendants did not renew their motions for summary judgment following that order, despite the fact that, according to them, claim construction was all that stood between

them and judgment in their favor. Instead, in June 2017, they filed a motion to compel adequate infringement contentions. Before that motion had been resolved, plaintiff’s counsel withdrew and the case ended. In sum, this process has not been speedy, but the Court finds little to lay in particular at plaintiff’s door (at least beyond the initial strategy of filing multiple suits and forcing others to do the work of consolidation, which is addressed below). On the whole, it cannot fairly be said that plaintiff showed no interest in litigating the merits of its claims. While plaintiff did dismiss several of the cases against MicroStrategy’s customers before they had even been served, let alone before the merits of the case could be tested, it has pursued the litigation against

MicroStrategy for many years, up to the Federal Circuit and back. It did not voluntarily dismiss the case in the face of multiple case-dispositive motions by MicroStrategy, as one might expect from a plaintiff determined to shield a weak patent from scrutiny for use in later settlement strong-arming; rather, it prevailed on them. See Adjustacam, LLC v. Newegg, Inc., 861 F.3d 1353 (Fed. Cir. 2017). It only gave up the litigation after it was apparently unable to continue the fight. Contrary to defendants’ insinuations, the Court does not find that plaintiff was attempting to evade a prevailing-party determination by firing its attorneys. Instead, it appears that plaintiff intended for this litigation to go forward and was attempting to find substitute counsel. Furthermore, as discussed more fully below, plaintiff had a reasonable basis to believe that MicroStrategy’s product infringed, at least at the beginning of the case. Checkpoint Sys., Inc. v. All-Tag Sec. S.A., 858 F.3d 1371, 1375 (Fed. Cir. 2017) (“Motivation to implement the statutory patent right by bringing suit based on a reasonable belief in infringement is not an improper motive.”). The Court therefore does not conclude that these lawsuits were filed with “the sole purpose of forcing settlements.”

2. Unreasonable Litigating Position Next, defendants contend that plaintiff never had a coherent theory of infringement, and therefore the substantive strength of its litigating position was exceptionally poor. Patentees have an obligation to evaluate the strength of their case continually as the litigation goes on. Taurus IP, LLC v. DaimlerChrysler Corp., 726 F.3d 1306, 1328 (Fed. Cir. 2013) (“[A] party cannot assert baseless infringement claims and must continually assess the soundness of pending infringement claims, especially after an adverse claim construction.”). It is certainly possible that a lawsuit that was baseless from the beginning could nevertheless require a great deal of litigation and expense before that becomes clear to the Court. But it is also possible that a lawsuit which had a reasonable basis in the beginning becomes insupportable only after some

discovery or claim construction has taken place. The Court therefore undertakes to evaluate the strength of plaintiff’s case both before and after claim construction. a. Before Claim Construction in This Case Defendants have been vocal from the start about their dissatisfaction with plaintiff’s infringement contentions. As early as May 2012, they filed a motion to compel infringement contentions, arguing that the contentions plaintiff had served were inadequate. (No. 11-cv- 12220, ECF 52). The Court granted that motion in part and denied it in part, explaining that “[a]t this preliminary stage, it is not necessary for plaintiff to provide highly detailed or ultimately successful contentions; however, defendants are entitled to know precisely which products (or combinations of products) are alleged to infringe each asserted claim, and how each asserted claim limitation is allegedly met by each accused product.” (No. 11-cv-12220, Electronic Order, July 2, 2012). Plaintiff filed amended contentions on July 20, 2012. MicroStrategy contends that those contentions were still inadequate, but that it responded with their invalidity and non- infringement contentions anyway on August 10 in the hopes that, 90 days after it produced its

source code, plaintiff would further amend their infringement contentions with specific citations to it. (See 11-cv-11970, ECF 213 at 4). Shortly thereafter, the case was stayed pending the outcome of the Federal Circuit. Despite having indicated that it did intend to update its infringement contentions to include citations to MicroStrategy’s source code within 90 days of its being produced, as the parties’ proposed joint scheduling order had provided, DataTern contended at the hearing on the motion to stay that it was not required to do so because that portion of the parties’ proposal was never adopted by the Court. (11-cv-12220, ECF 98 at 14- 18). The Court agreed, and DataTern did not inspect the source code at that time. (Id. at 18-19, 32-33).6

After this case was remanded from the Federal Circuit, MicroStrategy filed two motions for summary judgment in February 2015, the briefing schedule for which extended into May 2015. (No. 11-cv-11970, ECF 65). After the Court’s ruling on those motions in September 2015, and motion practice concerning MicroStrategy’s motion for a bond in February and March 2016, a new case management schedule was set in April 2016, resetting the Markman briefing to

6 Defendants dispute whether plaintiff was obligated to inspect the source code at that time. The Court’s order explained that “[t]he stay shall not, however, prohibit DataTern from inspecting the software source code made available by MicroStrategy on such reasonable terms as the parties may agree,” and that “all filing deadlines are suspended pending further order of the Court.” (No. 11-cv-12220, ECF 96; see also No. 11-cv-12220, ECF 98 at 32-33 (“I will permit DataTern, if it chooses, to inspect the source code now. It says it doesn’t need to, but I would not view that as a violation of the stay. It would be DataTern imposing costs on itself voluntarily if it chose to do that to help expedite matters.”)). Thus, DataTern had the option to inspect the source code during that time, but was not required to. begin in September. (No. 11-cv-11970, ECF 115). That schedule provided that the “[d]eadline for amending infringement and invalidity Contentions” was “30 days after Court issues claim construction decision.” (No. 11-cv-11970, ECF 114). It seems that plaintiff may have been disingenuous about its intentions to supplement its infringement contentions with citations to the source code. However, it is also clear that plaintiff

did not have an obligation to do so, at least prior to claim construction. The question is really whether the infringement contentions plaintiff did file failed to adequately disclose a reasonable litigating position as to particular claim elements, and were instead either facially frivolous or calculated to confuse so that defendants would be unable to defend themselves. At least by the time of the July 20, 2012 amended infringement contentions, the Court cannot conclude that plaintiff’s infringement contentions were facially frivolous. Plaintiff proffered a 43-page, element-by-element chart with explanations of its positions and citations to text and images from publicly available descriptions of how MicroStrategy’s product worked. It is true that they are not a model of clarity, and sometimes name more than one structure as

satisfying a claim element (for example, both third parties’ own object-oriented applications and the applications in the “Report” layer of MicroStrategy’s Business Intelligence Platform appear to be accused as the “object oriented software application”). (11-cv-12220, ECF 72-1 at 1-3, 8). But they are adequate to show a reasonable investigation of the merits of infringement and a facially reasonable basis for bringing suit. Beyond conclusory statements that the contentions are incomprehensible, defendants complain that “in various places DataTern accused nine or more different program structures as allegedly satisfying the ‘interface object’ element” and “accused the same structures of also being many other elements.” (11-cv-11970, ECF 235 at 10).7 But a few of these “different” structures appear to be different words for the same structure, and the Court can hardly expect that the publicly available documentation plaintiff relied on to support its contentions would use language with a perfect one-to-one correspondence between the claimed structures and the alleged infringing structures.8 However, contentions that appear adequate on their face might prove baseless in

particulars. Over the course of the litigation, defendants have put forth various theories as to why they should be granted summary judgment of noninfringement on particular claim elements, and the Court will consider each in turn to determine whether plaintiff’s position with respect to that element was objectively baseless. In their motion for fees, defendants focus on plaintiff’s inability to show that the accused products meet the “object model” limitation of the claims, which was also the focus of their most recent (February 2015) motion for summary judgment. That claim term was construed by the New York court, against DataTern, to require “classes,” and that construction was upheld on appeal. Microsoft, 755 F.3d at 909. DataTern never stipulated to noninfringement in this

litigation based on that construction of “object model”—its stipulation in this case was limited to the term “to create at least one interface object.” It had stipulated to a construction of the term “class” in the New York litigation, but contended in this litigation that that term required

7 Specifically: “(1) ‘Metadata Objects,’ (2) unidentified ‘assembled objects,’ (3) unidentified ‘[d]ynamic, object-oriented metadata objects,’ (4) other unidentified ‘similar sets of associated objects’ ‘in the “Develop” layer of the MicroStrategy Business Intelligence platform,’ (5) ‘custom SQL code,’ (6) ‘optimized, mulit-pass SQL queries,’ (8) ‘the [MicroStrategy] communication framework,’ and (9) an unidentified ‘set of data that represents the relationships between the items in the object model to the physical layout of the relational database.’” (11-cv- 11970, ECF 213 at 7). Note that there are actually only eight “different structures” listed. 8 For example, “As the central contact point to the metadata, Intelligence Server dynamically assembles the metadata objects to create optimized, multi-pass SQL queries for every major relational database.” (11-cv-12220, ECF 72-1 at 3) (quoting MicroStrategy’s own materials). construction, and this Court held that it was not estopped from taking that position.9 In its opposition to defendants’ February 11, 2015 motion for summary judgment on that ground, plaintiff presented an expert declaration supporting its construction of “class” and its position that MicroStrategy’s platform contained “classes” under that definition. The Court found this persuasive enough to deny summary judgment until the term “class” could be construed. (11-cv-

11970, ECF 101). For the reasons given in the memorandum denying summary judgment, the Court will not now hold that such a position was unreasonable or baseless. Medtronic Navigation, Inc. v. BrainLAB Medizinische Computersysteme GmbH, 603 F.3d 943, 954 (Fed. Cir. 2010) (“Absent misrepresentation to the court, a party is entitled to rely on a court’s denial of summary judgment and JMOL . . . as an indication that the party’s claims were objectively reasonable and suitable for resolution at trial.”). In addition to the “object model” ground also addressed in the February 11, 2015 motion for summary judgment, MicroStrategy asserted five other grounds on which it contended it deserved summary judgment in its February 4, 2013 motion—two based on the New York

court’s claim constructions, and three based on plaintiff’s infringement contentions. That motion was never opposed, as the Court stayed the litigation two days after it was filed and defendants withdrew it shortly after the remand. (11-cv-11970, ECFs 38, 62). The Court granted this motion only on the “object model” ground, without considering any of the other grounds. And the only one of those grounds DataTern seeks to relitigate in its fees motion is its argument with respect to the “code generator” limitation in claim 10. Nevertheless, it seems necessary to determine at this juncture whether that motion establishes that plaintiff had no colorable

9 Indeed, the Federal Circuit held that DataTern had agreed to “a narrower construction of classes than that required by the ’502 patent,” because the patent only required classes to include “attributes,” not “attributes and behaviors.” Microsoft Corp. v. DataTern, Inc., 755 F.3d 899, 909 (Fed. Cir. 2014). infringement argument prior to claim construction in this case. Defendants first contended that their accused products do not “create an interface object” according to the New York court’s claim construction, because that construction required generation of software code for a class, which MicroStrategy’s product did not do. That construction, however, was reversed by the Federal Circuit, and so, to the extent plaintiff’s

infringement position depended on its proposed construction, it was certainly reasonable. DataTern, Inc. v. Epicor Software Corp., 599 F. App’x 948 (Fed. Cir. 2014). Defendants also argued that their accused products did not satisfy the “runtime engine” element, which was construed by the New York court to be “software that (i) the object oriented software application depends on to run, (ii) must be running to execute the object oriented software application, (iii) uses the map in its processing, and (iv) is not part of the object oriented software application.” Microsoft, 2012 WL 3682915, at *14. Defendants simply argued that plaintiff had not put forth any evidence as to any of these requirements, and therefore they deserved summary judgment. (11-cv-11970, ECF 32 at 29). That construction was not

addressed by the Federal Circuit, and this Court did not adopt it until February 7, 2017. Pre- claim-construction infringement contentions surely cannot be expected to include evidence supporting post-claim-construction requirements, and accordingly the inability to satisfy the “runtime engine” limitation does not render plaintiff’s infringement position unreasonable from the beginning. The contentions identified structure allegedly corresponding to the “runtime engine” and performing the function of the runtime engine recited in the claims in the context of MicroStrategy’s product. That was a reasonable basis on which to bring suit as to that claim element. Defendants further contended that plaintiff never had a reasonable basis to believe that the accused products had an interface object that was “associated with an object corresponding to a class associated with the object oriented software program.” According to defendants, that term requires the interface object in MicroStrategy’s product to be associated with “other objects found in a different software application,” and plaintiff provided no evidence of such a relationship. (11-cv-11970, ECF 32 at 31). In plaintiff’s July 20, 2012 contentions, that

“different software application” could be either a “software application in the ‘Report’ layer of the MicroStrategy Business Intelligence Platform” or a “client application” that might be “created by MicroStrategy Incorporated’s licensees, customers, and/or users.” (11-cv-12220, ECF 72-1 at 7-8). The contentions also state: The interaction between the interface object instantiated by the MicroStrategy Business Intelligence platform and the application object instantiated by the application in the ‘Report’ layer of the MicroStrategy Business Intelligence platform, indicated by the cited evidence, shows the required association. . . . . For example, when the MicroStrategy Business Intelligence platform receives a request from the ARC Document Retrieval Service, the MicroStrategy Intelligence Server component instantiates an interface object. The interaction between the interface object instantiated by the MicroStrategy Business Intelligence platform and the client application object instantiated by the client application, indicated by the cited evidence, shows the required association. See Exhibit B.” (Id.). The closest the “cited evidence” comes to showing an association between “Report layer” applications and interface objects is to say that “report processing is performed by Intelligence Server” which “is linked to the other platform components using a highly optimized communication framework,” which in turn “handles the information flow between different platform components and provides secure user access, efficient report delivery, and complete data interactivity” and manages requests “uniformly from any requesting source” including Report layer applications and third party applications. (Id. at 8). No Exhibit B was docketed, although it is referred to elsewhere in contentions in sections describing the relationship between MicroStrategy’s software and the ARC Document Retrieval Service. (Id. at 11). Absent any argument one way or the other, the Court will assume that Exhibit B exists and simply was not filed on the docket. Plaintiff never had an opportunity to oppose defendants’ motion for summary judgment, and while it is not entitled to factual inferences drawn in its favor

at the fee-award stage, it is defendants’ burden to show that the case is exceptional. The infringement contentions recognize that an association must be shown, and purport to show that association in Exhibit B. Coupled with the other cited evidence, the Court concludes that defendants have not met their burden, with respect to this ground for noninfringement, to show that plaintiff’s position was unreasonable. Defendants further contend—both in the February 4, 2013 motion and in their current motion for fees—that plaintiff cannot show that the accused products include a “code generator” as required by independent claim 10 of the ’502 patent. Claim 10 recites “a code generator that employs said map to create at least one interface object.” ’502 patent, col. 8 ll. 32-33. Plaintiff

stipulated that MicroStrategy’s Business Intelligence platform does not “generate code for at least one class and instantiate an object from that class,” as was required by the then-governing construction of “to create at least one interface object” in claim 1. See DataTern, Inc. v. Epicor Software Corp., 599 F. App’x at 949-50. That stipulation is not, as defendants suggest, an admission that MicroStrategy’s product does not generate code that creates an interface object— rather, it is an admission that MicroStrategy does not both generate code for at least one class and instantiate an object from that class. The Federal Circuit noted that “[e]ven claim 10 does not expressly require generating code for a class.” 599 F. App’x at 952 n.2. And plaintiff’s infringement contentions cite to additional structures and portions of MicroStrategy’s publicly available documents for this element of claim 10 as compared to the structures they cited for the corresponding element in claim 1. (11-cv-12220, ECF 72-1 at 7-10, 24-26). Therefore, the Court does not view plaintiff’s stipulation regarding claim 1 to be conclusive on that point and will not find plaintiff’s position to be baseless for that reason. Finally, defendants contended that plaintiff never explained how the “runtime engine”

“invokes said at least one interface object with the object oriented application to access data from the relational database,” because (according to plaintiff’s infringement theory) the “object oriented application” has no role in accessing the relational database. (11-cv-11970, ECF 32 at 33). The Court agrees that, according to plaintiff’s infringement contentions, the “object oriented application” appears to be limited to generating the request to the runtime engine and receiving the report with the data. (See 11-cv-12220, ECF 72-1 at 10-11). But it is not unreasonable to interpret “invokes at least one interface object with the object oriented application” to encompass a scenario where the object oriented application sends a request, the details of which determine which interface object is invoked by the runtime engine. Therefore,

defendants’ argument as to this ground for summary judgment does not convince the Court that plaintiff’s infringement suit was baseless from the beginning. And to the extent that this argument has merit, the Court notes that it was made in one paragraph of a motion that was never opposed or ruled on; surely it is improper to rely on it for a finding that DataTern’s infringement allegations were baseless from the outset. Furthermore, although the Court will consider the totality of the circumstances and evaluate the case as a whole, it is worth noting that it already determined in March 2016 that case was not exceptional enough to warrant a bond because “little ha[d] changed since the Court denied summary judgment in September 2015: the parties still have not begun claim construction and the judicial estopped argument on the term ‘class’ remains undeveloped.” (11- cv-11970, ECF 107). And the Court declined to address defendants’ argument that the case was exceptional back in 2013 immediately before the case was stayed. (11-cv-11970, ECF 32). b. After Claim Construction in This Case “[A] case can be found exceptional when a party prolongs litigation in bad faith.” Taurus

IP, 726 F.3d at 1328. Prior to claim construction, it may well be that plaintiff had to rely on publicly available information that did not precisely identify components of MicroStrategy’s software by name in ways that clearly matched up with elements of the claims, but that nonetheless could plausibly be interpreted to show that various parts of the software together performed each and every element of the claims. And, prior to claim construction, plaintiff had no particular obligation to annotate its infringement contentions with citations to the source code. But after claim construction and a thorough inspection of the source code, plaintiff was required to meaningfully address the Court’s constructions and make its contentions more specific. In particular, after this Court construed “class” to mean “a definition that specifies attributes and behavior of objects, and from which objects can be instantiated,” plaintiff was

required to show a reasonable theory as to how the MicroStrategy platform could meet the limitation “to create at least one interface object,” which the Federal Circuit had construed to mean “to instantiate at least one interface object from a class.” (11-cv-11970, ECF 204 at 37); DataTern, Inc. v. Epicor Software Corp., 599 F. App’x at 954. Plaintiff added the following paragraphs to its contentions to address this: The set of data that represents the relationships between the items in the object model to the physical layout of the relational database is used to dynamically generate custom SQL code, which is stored in an object associated with an instance of the current Report being run. Each item of metadata (such as a Report) in the metadata repository (i.e., object model) contains a set of associated attributes and behaviors. For example, the MicroStrategy platform employs a ReportInstance object that stores information relating to a report definition, the schema, and generated SQL code. See, e.g., . There is a ReportInstance object associated with each instance of a report being executed. See, e.g., id. (“We use a report instance to hold all of the information associated with one particular execution of a report. Each time a report is executed we create a new report instance.”) . . . . “Objects are simply assembled directly from the metadata and resolved when the report is run.” MicroStrategy Architecture Review, p.36. (11-cv-11970, ECF 206-1 at 10-11). While that may be adequate to show a good-faith basis to believe that the map is used to instantiate an interface object (a “ReportInstance”) from a class (a “Report”), there is nothing but plaintiff’s say-so to suggest that the class has “attributes and behaviors.” Although plaintiff was willing to stipulate to this definition of “class” in the New York litigation, it refused to do so here even in the face of a judicial-estoppel argument, suggesting that at least plaintiff believed a broader definition was required to cover the MicroStrategy Business Intelligence platform, which is the only accused product in this lawsuit. After all that, and a thorough inspection of MicroStrategy’s source code, plaintiff should have been able to articulate exactly what the required “behaviors” were and where they could be found. Simply stating that the class “contains a set of associated attributes and behaviors” is not enough. Furthermore, after claim construction, plaintiff should have been able to show that the “runtime engine” is “software that (i) the object oriented software application depends on to run, (ii) must be running to execute the object oriented software application, (iii) uses the map in its processing, and (iv) is not part of the object oriented software application.” Although plaintiff did address this construction in its March 9, 2017 contentions, it again did not offer more than conclusory statements, despite its inspection of MicroStrategy’s source code.10 Thus, while not baseless at the outset, plaintiff’s case appears to have become unreasonable following the Court’s construction of “class.” 3. Inventorship of the ’502 Patent Defendants next argue that plaintiff failed to conduct an adequate investigation before

filing suit as to whether the correct inventors were listed on its patent. Specifically, defendants complain that plaintiff failed to interview one Max Arai, who was listed as an inventor on the two provisional patent applications but not the utility patent application. Despite having marshalled considerable evidence on the topic, defendants insist, in their reply, that they are not asking the Court to make a determination as to the correct inventorship of the patent. (11-cv-11970, ECF 275 at 9). And the Court will not do so. Defendants did not raise this issue in any of their prior motions, including their motions for summary judgment, even though they have known since at least 2012 that Arai was on the provisional patent but not the issued one, and that Arai refused to sign his oath of inventorship in connection with the provisional. Nor did defendants seek to depose Arai. Defendants themselves admit that

“[e]v[e]n a cursory review of the ’502 Patent file history reveals serious flaws in its inventorship.” (11-cv-11970, ECF 235 at 15). The Court is unwilling to delve deeply into Arai’s declaration and the documents he provided to defendants; among other things, plaintiff never had an opportunity to cross-examine him and those documents were produced to plaintiff only 27 days before plaintiff’s counsel sought to withdraw. Under the circumstances, it is

10 Plaintiff added, without citation: “When a Report is in the process of running, generated SQL stored in an interface object is submitted to a SQL Query Engine, which is responsible for accessing the relational database and returning the results of the query. The particular object-oriented application (such as described with respect to the preamble) depends on this Query Engine to run, and thus the Query Engine must be running to execute the object-oriented application. The Query Engine is not part of the object-oriented application, which may, for example, include user-provided elements in the metadata repository.” (11-cv-11970, ECF 206-1 at 13). enough to say that based on the materials before the Court, the listed inventorship is not so clearly incorrect that the matter is beyond reasonable dispute. Rather, the only question is whether plaintiff acted unreasonably in failing to make an adequate investigation of the inventorship of the ’502 patent before filing suit.11 It is undisputed that plaintiff never interviewed Max Arai. In arguing that it should have,

defendants point to (1) the fact that the inventors were changed; (2) documents produced by plaintiff showing that Arai and two others left off the utility application refused to assign their rights to Ontos, DataTern’s predecessor in interest; and (3) a 2009 lawsuit in which DataTern sued its former attorneys at Foley & Lardner LLP for breach of contract, in which it alleged, among other things, that “Foley failed to interview a provisional inventor (Max Arai) on U.S. Patent No. 6,101,502.” (11-cv-11970, ECF 254-31 ¶ 34(f)).12 In its answer to the complaint, Foley & Lardner asserted that: 30. Foley & Lardner pointed out to [DataTern’s predecessor in interest] that, as a result of the difference in the names of the inventors between the provisional and non-provisional applications, the defendant [in that case, Red Hat, Inc.] was likely to assert improper inventorship as a defense to an infringement action. 31. During the course of the litigation, Foley & Lardner interviewed several, but not all, of the inventors, their supervisors at the time the applications were filed, and another party who was not a named inventor on any of the applications. Foley & Lardner determined at that time that there was no basis to conclude that

11 Defendants cite to several cases where fees were awarded because a patentee attempted to defraud the Patent Office or made misrepresentations about the inventorship of the patent. Advanced Magnetic Closures, Inc. v. Rome Fastener Corp., 607 F.3d 817, 832 (Fed. Cir. 2010); PerSeptive Biosystems, Inc. v. Pharmacia Biotech, Inc., 225 F.3d 1315, 1321 (Fed. Cir. 2000). But, as defendants acknowledge in their reply, the question for this Court is not whether the inventorship is erroneous, and the Court makes no finding as to whether plaintiff perpetrated a fraud on the Patent Office. Had defendants argued that plaintiff had engaged in inequitable conduct, it seems that the Court would have jurisdiction to entertain that claim in connection with a fees motion and could hold that the patent was unenforceable on that basis. Monsanto Co. v. Bayer Bioscience N.V., 514 F.3d 1229, 1242-43 (Fed. Cir. 2008). 12 The relevant breach-of-contract claim alleges that “Foley breached the Contract to provide the first rate, top quality, thorough and professional IP legal services promised in at least the following respects: . . . .” (11-cv- 11970, ECF 254-31 ¶ 34). the inventorship on the ’502 patent was incorrect and that there was no basis to support a defense that the named inventors intended to perpetuate a fraud concerning inventorship. (11-cv-11970, ECF 254-32 ¶¶ 30-31). That suit later settled. See DataTern, Inc. v. Foley & Lardner, LLP, No. 2:09-cv-00038-TJW (E.D. Tex. filed Jan. 29, 2009). Defendants argue that all of this, taken together, is clear evidence that plaintiff knew that inventorship was a possible weakness of the patent—indeed, to the extent that plaintiff thought the failure to interview Arai was an error worth suing over—and that therefore its failure to interview Arai before asserting claims under the patents makes this an exceptional case. Plaintiff counters that it reasonably relied on its prior counsel’s determination that there was no basis to conclude that the named inventors were incorrect. Plaintiff points to the depositions of three people taken in connection with the New York litigation: Kenneth Lord, who, like Arai, was named on the provisional applications but not the utility application (11-cv- 11970, ECF 270-1 at 143); Gabriel Oancea, who was named on the utility application but not the provisionals (11-cv-11970, ECF 270-1 at 239); and Robert Heubner, who was named on both the provisionals and the utility application (11-cv-11970, ECF 270-2 at 219). Those depositions—

along with those of Jon Coleman and Robert Donald, who were named only on the utility application—were taken (by the alleged infringers in the New York cases) in August 2012. (11- cv-11970, ECF 254-36).13 Defendants do not point to anything in those depositions that suggest an interview of Arai was necessary or that there was no reasonable basis for believing the listed inventors to be correct. The Court is skeptical that Foley’s answer to a lawsuit in which it was adverse to

13 Plaintiff also points to a post-suit declaration by Mark Eisner, who was the supervisor of Arai and others listed as inventors on the provisional applications but not on the utility application. (11-cv-11970, ECF 265). DataTern can be considered a “statement of counsel” on which DataTern was entitled to rely. See Andrew Corp. v. Gabriel Elecs., Inc., 785 F. Supp. 1041, 1055 (D. Me. 1992). DataTern’s own amended complaint specifically lists the failure to interview Arai as an actionable error. Nevertheless, in the Court’s view, an interview of Arai was not required for plaintiff to have a reasonable belief that the inventorship was correct. Foley & Lardner’s statement shows that it

had performed some investigation that it considered adequate. Moreover, it appears that in August 2012, prior to the lawsuits in this case being filed, inventorship had been explored with other individuals named as inventors on either the provisional or utility applications in the New York litigation. While consistency is certainly desirable, it does not appear, and defendants do not argue, that plaintiff should be estopped from taking the position that it had a reasonable basis for concluding that the inventorship is correct because it included a failure to interview Arai among a laundry list of claims in another suit. And there is no requirement that a patentee, having been assigned a patent claiming priority to 1997, must track down every person ever alleged to be an inventor and interview them; a reasonable investigation is sufficient.

Here, prior to filing suit against defendants in this case, plaintiff or its counsel had interviewed several inventors and was familiar with the events surrounding the conception and reduction to practice of the invention contained in the ’502 patent. Under the circumstances, that seems sufficiently reasonable to avoid the imposition of attorney’s fees. 4. Judge-Shopping Finally, defendants contend that plaintiff’s behavior at the beginning of this litigation compels the conclusion that it was engaged in judge-shopping behavior. On November 7 and 8, 2011, plaintiff filed eight separate actions in this district against parties who turned out to be MicroStrategy’s customers. (Nos. 11-cv-11970, 11-cv-11975, 11-cv-11976, 11-cv-11977, 11- cv-11980, 11-cv-11982, 11-cv-11983, 11-cv-11984). None of those cases were marked as related, and they were randomly assigned to various judges in this district. None of them were assigned to Judge Stearns. On November 15, 2011, plaintiff filed nine more separate actions in this district, again against parties who turned out to be MicroStrategy’s customers. (Nos. 11-cv-12022, 11-cv- 12023, 11-cv-12024, 11-cv-12025, 11-cv-12026, 11-cv-11027, 11-cv-12028, 11-cv-12029, 11-

cv-12030). Again, none of these cases were marked as related, and they were also assigned randomly. This time, however, four of them were assigned to Judge Stearns. Beginning days later, on November 18, 2011, and December 12, 2011—before any of the defendants had been served—plaintiff voluntarily dismissed every case that had not been assigned to Judge Stearns except Nos. 11-cv-11970, which had originally been assigned to a magistrate judge, and 11-cv-12027, which had been assigned to Judge O’Toole.14 Then, on December 14, 2011, plaintiff filed five more cases in this district against four more of MicroStrategy’s customers and against MicroStrategy itself. (Nos. 11-cv-12220, 11-cv- 12223, 11-cv-12224, 11-cv-12225, and 11-cv-12227). This time, plaintiff marked these cases as

related to each other and related to No. 11-cv-12024, which had been assigned to Judge Stearns. Accordingly, those cases were also assigned to Judge Stearns. Defendants assert that this was blatant judge-shopping. In its defense, plaintiff states only the following: When we began this suit, we began it against individuals who we thought were independent software vendors. It turned out that they were resellers. We joined MicroStrategy then in December of 2011 and ultimately agreed to a stay of the claims against the resellers so long as they agreed to be bound with respect to the

14 No.11-cv-11970 was later assigned to Judge Woodlock, and then reassigned to the undersigned judge. The parties later filed a joint motion to consolidate 11-cv-11970, the lowest-numbered case of all the proceedings, with 11-cv-12220, the case against MicroStrategy. No. 11-cv-12027 was voluntarily dismissed by DataTern on April 10, 2012, after the defendant there, InQuira Inc., had been served. MicroStrategy suit. At every turn of this case, we have attempted to do this in the most efficient fashion as we are attempting to do it right now. (No. 11-cv-12220, ECF 98 at 26:25-27:7). The Court finds that suggestion not entirely credible. Plaintiff would have the Court believe that it spent $400 in filing fees suing each of seventeen defendants it thought were independent software vendors, only find out a few weeks later that it had made a mistake and that they were all actually selling the same product, originally made by MicroStrategy. It seems that any amount of pre-suit investigation would have alerted plaintiff to the fact that those defendants were resellers (and it did not take plaintiff long to find out after the suits were filed, nor was any discovery or even responsive pleading required). But even assuming that plaintiff made the decision to file seventeen lawsuits in eight days without bothering to determine whether the cases were related beforehand, that does not explain why

plaintiff proceeded to dismiss all the suits (except two) not before Judge Stearns, keep all the cases that were before Judge Steans, and mark its five new cases as related not to the lowest- numbered remaining case (that against Blazent), which at the time was assigned to Magistrate Judge Dein, but to the lowest-numbered remaining case before Judge Stearns. The Court therefore finds that plaintiff deliberately filed two groups of cases, against parties as to which it never intended to proceed, in an attempt to secure a judge it thought would be favorable to it. It then waited to file its case against MicroStrategy itself until it could mark that case as related to one in front of its preferred judge. That conduct, under the circumstances, is serious misbehavior.

5. Defendants’ Behavior By way of offset or mitigation, plaintiff complains that defendants filed a retaliatory ex parte reexamination request against M2M Imaging, a medical imaging company. M2M is a portfolio company of Amphion Innovations PLC, which wholly owns DataTern. (11-cv-11970, ECF 266). A Fish and Richardson attorney who identified himself as “Attorney for Microstrategy Inc.” sent M2M the draft request on February 8, 2012, along with an email saying he intended to file it with Patent Office during the week of February 13, 2012, and encouraging M2M to forward the draft request to Amphion’s CEO. (11-cv-11970, ECF 266-1). MicroStrategy does not appear to be a direct business competitor to M2M. (See 11-cv-11970,

ECF 266). Plaintiff interpreted this as an act of retaliation for DataTern’s suit, undertaken just two months after the litigation against MicroStrategy was filed. Defendants respond that their reexamination was meritorious and largely successful, in that it resulted in the cancellation of three claims and the narrowing of many others, and that they have a constitutionally protected right to file such requests. See Baker Driveaway Co. v. Bankhead Enters., Inc., 478 F. Supp. 857, 859 (E.D. Mich. 1979). While perhaps aggressive, the Court concludes that this is not outside the realm of the expected, and does not on its own cause this case to stand out from others. Plaintiff also complains that defendants filed unnecessary and wasteful motions,

including refusing to accept the voluntary dismissal of claims against MicroStrategy’s customers, refusing to accept a narrow stipulation of infringement, filing a second motion for invalidity under 35 U.S.C. § 101 prior to claim construction after the court had already denied a similar motion in favor of waiting for claim construction (although claim construction had occurred in the New York litigation), and filing yet another motion for summary judgment even after the case was effectively over following the withdrawal of counsel. Defendants’ multiple motions were largely unsuccessful, either because they were denied outright or the consideration of the issues raised was deferred. While § 101 invalidity arguments are routinely raised prior to claim construction, the multiple motions for summary judgment here were excessive. The Court will take that behavior into account in the totality of the circumstances. See Motorola, Inc. v. Interdigital Tech. Corp., 121 F.3d 1461, 1468 (Fed. Cir. 1997) (explaining that the court may take both parties’ behavior into account). 6. Totality of the Circumstances Considering the totality of the circumstances, the Court concludes that this case is

“exceptional,” within the meaning of § 285. First, plaintiff engaged in a blatant and unapologetic attempt to manipulate the random assignment of cases. “Every court considering attempts to manipulate the random assignment of judges has considered it to constitute a disruption of the orderly administration of justice.” In re Bell South Corp., 334 F.3d 941, 959-60 (11th Cir. 2003) (citing cases); see Vaqueria Tres Monjitas, Inc. v. Rivera Cubano, 341 F. Supp. 2d 69 (D.P.R. 2004); see also In re Cargill, Inc., 66 F. 3d 1256, 1262-64 (1st Cir. 1995) (warning that entertaining a petition for mandamus where there is an appearance of judge shopping “run[s] a risk of eroding public confidence in the courts by seeming to reward a litigant for its gamesmanship”). Furthermore, plaintiff’s case became objectively baseless following the claim- construction ruling. And its choice to submit infringement contentions that merely paid lip

service to the Court’s Markman order, even after having had a meaningful opportunity to review MicroStrategy’s source code, shows that it was planning to continue the litigation in bad faith. The Court does not, however, conclude that this litigation was originally filed in bad faith or for an improper purpose. Plaintiff’s litigation position was not objectively unreasonable or frivolous for the greater part of the course of litigation. And part of the delay in reaching resolution in this case was due to defendants’ multiple premature or unsuccessful motions for summary disposition. Therefore, while some award of fees is appropriate, an award of fees for the entire litigation clearly would not. C. Award of Fees 1. Attorneys’ Fees to MicroStrategy For the foregoing reasons, the Court will award MicroStrategy (1) fees reasonably related to plaintiff’s judge-shopping activities—generally, fees incurred prior to these cases being consolidated and stayed against the customers—and (2) fees reasonably incurred after the Court’s claim-construction decision.

2. Attorneys’ Fees to MicroStrategy’s Customers Furthermore, the Court will grant the customers’ motions for attorneys’ fees to the extent they arise out of plaintiff’s judge-shopping activities. As discussed above, most of MicroStrategy’s customers had their claims dismissed within weeks, before they had even been served. The customers remaining in this lawsuit appear to remain only because they were unlucky enough to have their cases assigned to Judge Stearns. And they incurred the greater part of their fees prior to July 2012, when the case was stayed against them and they agreed to be bound by the determination of liability as to MicroStrategy. It seems that much of that cost could have been avoided had plaintiff clarified sooner that it had no claims against those defendants that were not related to MicroStrategy’s software. Therefore, those fees, as a general

matter, are directly related to plaintiff’s misconduct, and will be awarded upon application. However, there appears to be at least one exception. Certain customers (Epicor Software Corp.; Airlines Reporting Corp.; Carl Warren & Co., Inc.; Teradata Corp.; and Lancet Software Development, Inc.) did not simply accept the fact that DataTern had dropped the other claims against them and instead filed a summary judgment motion. (11-cv-12220, ECF 44). Accordingly, any fees incurred in connection with the filing of that motion will not be included as part of any award. 3. Expert Fees Defendants further request that the Court “invoke its inherent power to impose sanctions in the form of reasonable expert fees” in this case. Such an award is appropriate in cases “with ‘a finding of fraud or abuse of the judicial process’” or “in cases involving bad faith that cannot be otherwise reached by rules or statutes.” Takeda Chem. Indus., Ltd. v. Mylan Labs., Inc., 549

F.3d 1381, 1391 (Fed. Cir. 2008). This is not such a case. To the extent that there is bad faith, it can be adequately reached by the fees awarded under § 285. Because the Court has not concluded that plaintiff’s litigation position was unreasonable in its entirety, defendants’ expert fees were not incurred in connection with the behavior that made this case exceptional, and plaintiff will not be required to pay defendants’ expert fees. 4. Plaintiff’s Fees and Costs in Connection with Defendants’ Motions for Fees Finally, plaintiff requests that it be awarded its costs and attorneys’ fees incurred in responding to defendants’ motion for fees. Because the Court is granting defendants’ motions, at least in part, they were not frivolous, and plaintiff is not entitled to its fees for responding. IV. Conclusion For the foregoing reasons, the following motions for an “exceptional” case determination under 35 U.S.C. § 285 are GRANTED in part and DENIED in part: 1. MicroStrategy Inc.’s Motion for an Exceptional Case Determination and Attorneys’ Fees (11-cv-11970, ECF 244); 2. Airlines Reporting Corporation’s Motion for an Exceptional Case Determination and Attorneys’ Fees (11-cv-11970, ECF 246);

3. Teradata Corporation’s Motion for an Exceptional Case Determination and Reasonable Attorneys’ Fees and Costs (11-cv-11970, ECF 248); 4. Defendant Epicor Software Corporation’s Motion for an Exceptional Case Determination and Attorneys’ Fees (11-cv-11970, ECF 249 and 11-cv-12220, ECF 142); 5. Defendant Premier Inc.’s Motion for an Exceptional Case Determination and Attorneys’ Fees (11-cv-12220, ECF 139);

6. Carl Warren and Company Incorporated’s Motion for an Exceptional Case Determination and Attorneys’ Fees Under 35 U.S.C. § 285 (11-cv-12220, ECF 141); 7. Defendant Lancet Software Development, Inc.’s Motion for an Exceptional Case Determination and Attorneys’ Fees (11-cv-12220, ECF 145). Each party seeking an award of fees shall file an application for fees, with affidavits and a detailed breakdown of time and costs, within 21 days of this order. The fees sought must be consistent with this order and reasonably related to the categories of activities for which the Court has indicated it will allow fees. Plaintiff may object or otherwise respond to such

applications within 14 days. So Ordered. /s/ F. Dennis Saylor F.Dennis Saylor, IV Dated: June 5, 2018 United States District Judge