Campbell et al v. Facebook Inc.

Filing 215

RESPONSE (re 207 MOTION to Compel Production of Configuration Tables ) Defendant Facebook, Inc.s Opposition to Plaintiffs Motion to Compel Production of Configuration Tables filed byFacebook Inc.. (Attachments: # 1 Poole Declaration in Support of Facebooks Opposition to Plaintiffs Motion to Compel Production of Configuration Tables)(Jessen, Joshua) (Filed on 8/19/2016)

Download PDF
1 2 3 4 5 6 7 8 9 10 11 12 13 GIBSON, DUNN & CRUTCHER LLP JOSHUA A. JESSEN, SBN 222831 JJessen@gibsondunn.com JEANA BISNAR MAUTE, SBN 290573 JBisnarMaute@gibsondunn.com PRIYANKA RAJAGOPALAN, SBN 278504 PRajagopalan@gibsondunn.com ASHLEY ROGERS, SBN 286252 ARogers@gibsondunn.com 1881 Page Mill Road Palo Alto, California 94304 Telephone: (650) 849-5300 Facsimile: (650) 849-5333 GIBSON, DUNN & CRUTCHER LLP CHRISTOPHER CHORBA, SBN 216692 CChorba@gibsondunn.com 333 South Grand Avenue Los Angeles, California 90071 Telephone: (213) 229-7000 Facsimile: (213) 229-7520 Attorneys for Defendant FACEBOOK, INC. 14 UNITED STATES DISTRICT COURT 15 NORTHERN DISTRICT OF CALIFORNIA 16 OAKLAND DIVISION 17 18 19 20 21 22 MATTHEW CAMPBELL and MICHAEL HURLEY, Plaintiffs, v. FACEBOOK, INC., Defendant. Case No. C 13-05996 PJH (SK) DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES” No Hearing Unless Requested By Court (Dkt. 203) The Honorable Phyllis J. Hamilton 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 TABLE OF CONTENTS 2 Page 3 I. INTRODUCTION AND SUMMARY OF ARGUMENT ....................................................... 1 4 II. FACTUAL AND PROCEDURAL BACKGROUND .............................................................. 3 III. LEGAL STANDARD ............................................................................................................... 4 IV. ARGUMENT ............................................................................................................................ 4 5 6 A. Plaintiffs’ Request For “Configuration Tables” Is An Improper Fishing Expedition ..................................................................................................................... 4 B. Plaintiffs’ Request for “Configuration Tables” Is Overbroad And Not Proportional ................................................................................................................... 6 C. Plaintiffs’ Assertion That “Facebook Has Relied On Configuration Table Data For Its Defenses Yet Refused to Produce that Data” Is False ....................................... 8 11 D. Granting Plaintiffs’ Request Could Materially Delay This Case .................................. 9 12 E. Facebook Has Already Agreed To Produce Proportional Discovery ......................... 10 7 8 9 10 13 V. CONCLUSION ....................................................................................................................... 10 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP i DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 2 I. INTRODUCTION AND SUMMARY OF ARGUMENT At the June 30, 2016 Case Management Conference, this Court made it abundantly clear to 3 Plaintiffs that any additional discovery was limited to the specific practices discussed in the Court’s 4 class certification order. Indeed, the Court took control of the three discovery disputes presently 5 before it to “ensure that the additional discovery is confined to the limitations necessitated by the 6 Court’s class certification ruling.” (Dkt. 203.) Plaintiffs’ Motion to Compel Production of 7 “Configuration Tables” is further evidence that Plaintiffs have not accepted the Court’s instruction. 8 The supposed basis for Plaintiffs’ Motion is that so-called “configuration tables demonstrate 9 how Facebook uses private message data.” (Dkt. 207 at 2 (capitalization omitted, emphasis added).) 10 Plaintiffs even go so far as to submit yet another declaration from their technical expert (Dr. Jennifer 11 Golbeck), in which she claims that she needs configuration tables “[i]n order to understand where 12 [data from URLs shared in messages] is stored and any subsequent use.” (Dkt. 207-2 ¶ 15 (emphasis 13 added).) These are stunning admissions in light of the fact that this Court certified a Rule 23(b)(2) 14 class based on the prior representations of Plaintiffs and Dr. Golbeck that they knew exactly how 15 Facebook was “using” URLs shared in Facebook messages. In connection with Plaintiffs’ Motion for 16 Class Certification, Dr. Golbeck submitted not one—but two—declarations (totaling almost 50 pages) 17 in which she devoted a lengthy section to “Facebook’s Uses of Intercepted Private Message Data.” 18 (Dkt. 199-2 at 15-26.) But now Dr. Golbeck tells the Court that she also “need[s] to review certain of 19 Facebook’s configuration tables in order to fully opine on the issues.” (Dkt. 207-2 ¶ 4.) 20 This Court’s certification order clearly framed the remaining issues in the case: one alleged 21 “interception” (the storage of a URL preview in the form of a share object or EntShare—the basis for 22 commonality (see Dkt. 192 at 15))—and three alleged anonymous and aggregate “uses” of that 23 information: (1) the counter next to the “Like” button social plugin, (2) “recommendations for other 24 users” in Facebook’s Recommendations plugin, and (3) the “sharing of user data with third parties” 25 through Facebook’s “Insights” product. (Id. at 3-5.) The Court certified a class based on these 26 specific alleged practices even though Plaintiffs’ then-operative complaint did not allege all of the 27 practices (or include Plaintiffs’ new class definition); instead, the Court certified a class and 28 generously permitted Plaintiffs to amend their complaint “to bring the complaint in line with the [new] Gibson, Dunn & Crutcher LLP 1 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 2 allegations, and the class definition, as presented on this motion for class certification.” (Id. at 6.) The Court explained at the recent Case Management Conference that it took this procedurally 3 unusual step for the sake of efficiency. Remarkably, Plaintiffs now seek to open up an entirely new 4 avenue of extensive discovery—a “text file dump” of “all configuration tables for all databases that 5 contain data derived from Private Message URL content”—in the apparent hope of finding entirely 6 new “uses” of message data to add to their case. But the Court already rejected this approach at the 7 Case Management Conference. (Dkt. 203.) After two and a half years of litigation and 18 months of 8 discovery that Plaintiffs acknowledged was “extensive,” the issues in this case have finally come into 9 focus, and Plaintiffs should not be permitted to go off on another frolic and detour. Plaintiffs’ Motion 10 should be denied for five primary reasons: 11 First, the Motion is an impermissible fishing expedition. The Court has already framed the 12 remaining issues in this case, and, despite Plaintiffs’ fact-free assertions to the contrary, Plaintiffs’ 13 requests are not tied to those issues. Plaintiffs should not be permitted to fish for new “uses.” 14 Second, Plaintiffs’ request for “configuration tables” is wildly overbroad and not proportional 15 to the needs of the case. Their general request for tables relating to “Private Message URL content” 16 sweeps in practices unrelated to this case (for example, Facebook’s storage of the message body 17 itself), and their request for specific tables and databases fares no better. The nine tables that Plaintiffs 18 identify by name consist of hundreds of millions of cells of irrelevant and sensitive information. 19 Third, Plaintiffs’ argument that “Facebook has relied on configuration table data for its 20 defenses yet refused to produce that data” (Mot. at 7) is false. In an effort to avoid wasteful motion 21 practice, Facebook produced data relating to practices that Plaintiffs challenged for the first time in 22 their Motion for Class Certification and which were not the basis for the court’s certification order nor 23 mentioned in Plaintiffs’ Second Amended Complaint. In any event, Plaintiffs now have this data. 24 Fourth, if their Motion is granted, and if history is any guide, Plaintiffs almost certainly will 25 seek to further delay the case schedule. After Facebook produces information from its databases, 26 additional discovery regarding that information would surely follow and likely would take another 27 several months. Plaintiffs have sought (unsuccessfully) to extend the schedule several times and can 28 Gibson, Dunn & Crutcher LLP 2 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 be expected to do so again. Plaintiffs have identified no corresponding benefit to justify this delay. 2 Fifth, Plaintiffs have received extensive discovery on all of the challenged practices, and 3 Facebook has agreed to provide more (including proportional additional document searches, 30(b)(6) 4 testimony, and interrogatories). If Plaintiffs identify discrete database information actually relevant to 5 the source code for their four challenged practices, Facebook has offered to consider producing such 6 information. But Plaintiffs should not be permitted wide-ranging discovery of entirely new avenues of 7 information not limited to the challenged practices, especially at this post-certification stage. II. 8 9 FACTUAL AND PROCEDURAL BACKGROUND As described in Facebook’s responses to Plaintiffs’ other discovery motions, Facebook has 10 produced an extensive amount of discovery in this case, including a significant number of documents, 11 extensive and detailed written discovery responses, dozens of hours of deposition testimony, and 12 extensive (and highly proprietary) source code that Plaintiffs’ experts and consultants have reviewed 13 for a cumulative total of 86 days. (Dkt. 201 at 5.) 14 On top of this admittedly “extensive” discovery, Plaintiffs now request that this Court “order 15 Facebook to produce all configuration tables for all databases that contain data derived from Private 16 Message URL content including, but not limited to, the three categories of configuration tables 17 specified [in their Motion], to be produced as a text file dump within two weeks of the entry of an 18 Order on this motion.” (Mot. at 10.) They explain that the “categories of configuration tables” they 19 seek are those related to six Facebook databases, which they group into three categories: (1) 20 “ ” “Hive,” “Scribe” (which is not actually a database), and “HBase,” (2) “ 21 “ ,” which Plaintiffs claim is a security-related database (but it is not). (Id. at 1.) In her 22 declaration, Dr. Golbeck identifies nine specific tables from those databases: “ 23 “h 24 “ 25 Plaintiffs contend that these so-called “configuration tables” show how “Facebook stores and uses data 26 intercepted from Private Messages” and “instructions for how that data is utilized.” (Dkt. 207 at 3.) 27 Plaintiffs’ assertions about the nature of configuration data at Facebook are incorrect. There ,” “ ”“ ”“ ”“ ” and “ ”“ ,” ,” .” (Dkt. 207-2 ¶¶ 18-25.) 28 Gibson, Dunn & Crutcher LLP ,” and (3) 3 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 are no distinct, identifiable “configuration tables” for the challenged practices. (Declaration of Neal 2 Poole (“Poole Decl.”) ¶¶ 4-6.) Nor has any Facebook witness in this case testified otherwise (contrary 3 to the misimpression Plaintiffs attempt to give the Court). Similarly, Plaintiffs’ assumptions and 4 assertions about the specific tables and databases they identify by name are also wrong. Those tables 5 and databases are massive, and they contain data that are almost entirely unrelated to the subject 6 matter of this case (let alone the disputed issues concerning the claims and defenses). (Id. ¶¶ 11-13.) 7 Nevertheless, Facebook offered to investigate configuration data related to any source code calls that 8 Plaintiffs identify that pertains to the practices at issue in this case. Plaintiffs refused and instead filed 9 their Motion. III. 10 11 LEGAL STANDARD The recent amendments to the Federal Rules of Civil Procedure “crystalize[] the concept of 12 reasonable limits on discovery through increased reliance on the common-sense concept of 13 proportionality”; they make clear that the “pretrial process must . . . eliminate unnecessary or wasteful 14 discovery” with a “careful and realistic assessment of actual need.” 2015 Year-End Report on the 15 Federal Judiciary available at http://www.supremecourt.gov/publicinfo/year-end/2015year- 16 endreport.pdf; see also Fed. R. Civ. P. 26, Adv. Comm. Note to 2015 Amdt. (acknowledging “the 17 problem of overdiscovery” and the need “to guard against redundant or disproportionate discovery”). 18 “No longer is it good enough to hope that the information sought [by Plaintiffs] might lead to the 19 discovery of admissible evidence. In fact, the old language to that effect [in Rule 26] is gone. Instead, 20 a party seeking discovery of relevant, non-privileged information must show, before anything else, 21 that the discovery sought is proportional to the needs of the case.” Gilead Scis., Inc. v. Merck & Co, 22 Inc., No. 13–04057–BLF, 2016 WL 146574, at *1 (N.D. Cal. Jan. 13, 2016). IV. 23 24 25 A. ARGUMENT Plaintiffs’ Request For “Configuration Tables” Is An Improper Fishing Expedition Plaintiffs’ primary argument is set forth in the very first substantive heading of their brief: 26 “Facebook’s Configuration Tables Demonstrate How Facebook Uses Private Message Data.” (Dkt. 27 207 at 2 (emphasis added).) As a factual matter, this is incorrect for several reasons (as detailed 28 Gibson, Dunn & Crutcher LLP 4 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 below), but more importantly, the Court has already limited Plaintiffs to the three “uses” framed by the 2 certification order. See supra 1-2. Plaintiffs’ brief contains no explanation—and there is none—for 3 why they need “configuration tables” to understand the three uses of URL message data remaining in 4 the case. Nor could Plaintiffs make such an argument with a straight face, given that they (and their 5 expert) purported to explain these three uses in considerable detail in Plaintiffs’ Motion for Class 6 Certification and supporting evidence. (Dkt. 199-1; Dkt. 199-2; Dkt. 184-17.) 1 7 Plaintiffs argue that they need certain “configuration tables” to “fully determine the 8 functionality of [certain] logging systems with respect to logging of URLs in Private Messages” (Dkt. 9 207 at 3 (emphasis added)), but this Court did not certify a class based on allegations of “logging” (a 10 standard practice for all major websites). On the contrary, the Court permitted a class to be certified 11 because it found commonality based on the “fact that Facebook creates a share object every time a 12 message is sent with a URL [preview].” (Dkt. 192 at 15 (emphasis added).) A share object (or 13 EntShare)—which is simply the storage of the URL preview that permits the preview to be displayed 14 to the message recipient—is not “logging.” And unlike the word “EntShare” (the alleged 15 interception), which appears throughout the operative complaint, the word “logging” appears neither 16 in the Court’s certification order nor in Plaintiffs’ Second Amended Complaint, and is not in the case. 17 Unhappy with their case—which is about anonymous and aggregate uses of data that ceased 18 many years ago—Plaintiffs now seek to go on a fishing expedition for additional “uses” of URLs in 19 messages. But the Court already made clear that “additional discovery is confined to the limitations 20 necessitated by the Court’s class certification ruling.” (Dkt. 203.) On this basis alone, the Motion 21 should be denied. 2 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP Plaintiffs’ argument that “configuration tables are equivalent to source code” and therefore “should have been produced” previously (Dkt. 207 at 3) is baseless. As one of Facebook’s Rule 30(b)(6) designees explained last fall, “[s]ource code is meant to be functional, meaning it performs operations.” (Adkins Depo. at 85:6-8.) Plaintiffs themselves previously acknowledged that “[t]he source code enables Plaintiffs to understand the processes Facebook employs for its messaging functionality.” (Dkt. 113 at 2.) Databases, in contrast, contain data and are distinct. 2 Furthermore, Plaintiffs’ Motion and expert declaration are replete with references to what they “may” find though additional discovery. (Mot. at 9 n.32; Dkt. 207-2 ¶¶ 11, 18, 20.) Such speculation does not come close to meeting Plaintiffs’ burden to demonstrate that the discovery they seek is relevant and proportional. “Speculation should never bait a relevancy hook.” Steel Erectors v. AIM 1 5 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 2 B. Plaintiffs’ Request for “Configuration Tables” Is Overbroad And Not Proportional Plaintiffs’ requests also are wildly overbroad. Plaintiffs ask this Court to order production of 3 information related to all databases with data derived from URLs in Facebook messages. But this 4 would sweep in databases pertaining to a host of practices not at issue. Well aware of the fishing 5 expedition that their Motion represents, Plaintiffs claim that they “seek only the production of those 6 configuration tables that relate to Plaintiffs’ claims . . . and thus the discovery satisfies [] 7 proportionality.” (Dkt. 207 at 4.) But Plaintiffs’ beliefs about the nature of configuration data at 8 Facebook are factually incorrect. As Facebook has repeatedly explained, “configuration table” is not a 9 sufficiently descriptive phrase to identify any specific material at Facebook, and configuration data 10 related to the challenged practices do not exist in a discrete table. (Poole Decl. ¶¶ 4-6.) Moreover, the 11 facts regarding the scope of Plaintiffs’ requests belie their assertions about proportionality. 12 “ ” “Hive,” “Scribe,” and HBase.” Plaintiffs claim to seek configuration data for 13 these databases that “will reflect any locations where Private Message content is logged by Facebook 14 as well as any further uses.” (Mot. at 5.) First, this is overbroad—Plaintiffs’ case is only about URLs 15 in messages for four specific practices. Second, to support their need for this data, they (1) pivot once 16 again to “logging” (i.e., the 17 declaration. But Plaintiffs have never offered a single piece of evidence showing that either 18 and scribeh_share_stats tables), and (2) cite to Dr. Golbeck’s or scribeh_share_stats had any relationship to any of the challenged practices during the 19 class period, and Dr. Golbeck merely speculates that Facebook “may have the link” in 20 (Dkt. 207-2 ¶ 18.) Moreover, Dr. Golbeck does not mention needing any information from Hive, 21 Scribe, or Hbase anywhere in her declaration, contrary to Plaintiffs’ suggestion. 22 “ ” Plaintiffs claim that tables in the 23 . database “contain , which is required to understand EntShare and 24 EntGlobalShare objects.” (Mot. at 6 (emphasis added).) Plaintiffs admit that these tables contain 25 information about , despite the fact that the challenged practices involve 26 27 Steel Int’l., 312 F.R.D. 673, 677 & n.5 (S.D. Ga. 2016) (denying motion to compel discovery “based solely on plaintiffs[’] pure speculation and in the face of existing discovery”). 28 Gibson, Dunn & Crutcher LLP 6 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 only one Object (EntShare) and no Associations. Plaintiffs do not “need” to understand for the practices at issue; the request is overbroad. 3 2 “ 3 ”: Plaintiffs claim that “ configuration tables are necessary to 4 understand the use Facebook’s security systems make of Private Message content” (Mot. at 6), but, 5 again, Plaintiffs do not restrict their request to security variables related to the challenged practices. 6 Moreover, Facebook does not intend to rely on information in this database in support of its “ordinary 7 course of business” argument, which is the only reason Plaintiffs provide for needing it. 4 8 9 The specific tables that Plaintiffs request from these databases contain massive volumes of irrelevant—and sensitive—information (Poole Decl. ¶¶ 4, 13-15), and the request is overbroad for this 10 independent reason. See, e.g., United States ex rel. Dani Shamesh, v. CA, Inc., No. 09-1600-ESH, 11 2016 WL 74394, at *7 (D.D.C. Jun. 1, 2016) (“[T]he party seeking discovery must demonstrate that 12 the information sought to be compelled is within the scope of discoverable information under Rule 13 26.”). Yet even the specific examples of tables and databases that Plaintiffs and their expert identified 14 (see, e.g., Mot. at 1; Dkt. 207-2 ¶¶ 18-25) are grossly overbroad and not proportional—consisting of 15 millions of cells of irrelevant and highly sensitive information from across Facebook’s system: 16 infrastructure, systems operations, security, front-end web design, all products, etc. 5 (Poole Decl. ¶¶ 17 11-13.) Moreover, several of the tables that Plaintiffs request do not contain configuration data at all. 18 (Id.) For the Court’s convenience, Appendix A identifies each of the requested tables, its excessive 19 scope, its estimated size, Plaintiffs’ supposed need for the data, and any actual nexus to any actual 20 element (which is, almost without exception, none). 21 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP Plaintiffs again cite Dr. Golbeck’s declaration, which states that data “could be stored as an association” (Dkt. 207-2 ¶ 20 (emphasis added))—she wants to fish through the database for associations that she has no basis to believe exist (and they do not (Dkt. 183-8 at 12)). 4 Plaintiffs also misrepresent the testimony of Facebook engineer Michael Adkins. Contrary to Plaintiffs’ statement, Mr. Adkins never testified that “it is not possible to determine functionality without” the “‘ configuration table in the database.” (Dkt. 207 at 4.) On the contrary, Mr. Adkins made no mention of “the table” at all. 5 This is extremely sensitive and valuable data—both for competitive and security reasons. For example, disclosure of the names of internal Facebook systems, processes, and table and database names increases the risk of successful spoofing or phishing campaigns by third-party hackers (i.e., unauthorized access achieved by impersonating authorized personnel). (See, e.g., Dkt. 211 at 3.) 3 7 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 C. 2 Plaintiffs’ Assertion That “Facebook Has Relied On Configuration Table Data For Its Defenses Yet Refused to Produce that Data” Is False 3 Plaintiffs accuse Facebook of “discovery misconduct,” asserting Facebook is making 4 “deliberate efforts to conceal the [supposedly] central role of the configuration tables.” (Dkt. 207 at 5 7-8.) On the contrary, what occurred here is that Plaintiffs introduced several entirely new theories of 6 liability in their Motion for Class Certification, which Facebook spent considerable time investigating 7 and rebutting. Facebook responded to those theories (concerning “ 8 in the January 14, 2016 Declaration of Alex Himel. (Dkt. 184-11 ¶¶ 44-57.) Neither theory was the 9 basis for this Court’s certification order, but, in the interests of trying to avoid this very motion, ” and “Nectar” logging) 10 Facebook produced data further substantiating two minor points in the Himel Declaration concerning 11 these practices. Facebook is not holding back any data on either of these (irrelevant) points. “ 12 ” Data. In response to one of Plaintiffs’ new theories, Facebook submitted 13 evidence that (i) before the start of the proposed class period, data regarding URL attachments sent 14 with Facebook messages may have been logged in a table called “ 15 deleted before the beginning of the class period. (Dkt. 185.) Facebook later learned that the table 16 existed (but was not logging new information) for a very brief period (22 days) during the proposed 17 class period. (Id.) The critical part of the statement remained true: this table only reflected pre-class 18 period data and therefore did not implicate any class members. (Id.) Nonetheless, although the point 19 was immaterial, Facebook filed an errata in the interests of complete transparency. (Id.) 6 Facebook 20 also produced the record it relied upon (from “metadata logs,” not a “configuration table” (Dkt. 209-5 21 at 203:7-22)). Plaintiffs now contend that Facebook’s response to unpled theories raised for the ,” and (ii) the table was 22 23 24 25 26 27 Proving that “no good deed goes unpunished,” Plaintiffs have repeatedly sought to use Facebook’s voluntary correction of the record on this immaterial point to tarnish Facebook, its employees, and its counsel. Purposely not recognizing the distinction between an unintentional mistake (what happened here) and a knowing falsehood, Plaintiffs have repeatedly referred to Mr. Himel’s initial statement as “false” and have argued that this small error on an immaterial point “demonstrates” that “neither Facebook nor its employees can be relied on to provide an accurate representation of Facebook’s internal data and systems.” The Court should disregard these baseless and unprofessional ad hominem attacks. 6 28 Gibson, Dunn & Crutcher LLP 8 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 first time in their Motion for Class Certification justifies further irrelevant discovery. The opposite 2 is true: Facebook produced the data, closing the door to further fishing. 7 “Nectar” Logging And “Scribe” Data. In their Motion for Class Certification, Plaintiffs 3 4 also challenged (for the first time) certain “Nectar” logging—which their expert admitted was done 5 (if at all) only before a message was sent (taking it squarely outside the realm of wiretapping law). 6 This theory is not in the case (neither the certification order nor the amended complaint mentions it), 7 but in response to Plaintiffs’ assertions at the time, Facebook noted in the Himel Declaration that 8 URL “share scrapes” were logged to Nectar between 0.01% and 1% of the time during the relevant 9 period. (Dkt 184-11 ¶ 52.) To avoid wasteful motion practice, Facebook produced later that data. 10 As for “scribeh_share_stats,” Plaintiffs raised that issue for the first time in connection with 11 their reply brief in support of class certification. (Dkt. 184-17 ¶ 28.) Facebook objected and sought 12 leave to address this new (and untimely) issue in a supplemental Himel declaration (Dkt. 170-1), but 13 the Court did not consider the new declaration. (Dkt. 192 at 11 n.5.) At all events, like alleged 14 logging to “ 15 Court’s class certification order. See supra 1-2. 16 ,” alleged logging to “scribeh_share_stats” is not in the case as framed by the Accordingly, Plaintiffs’ argument that “Facebook has relied on configuration table data for its 17 defenses yet refused to produce that data” (Mot. at 7) is simply false. 18 D. 19 Granting Plaintiffs’ Request Could Materially Delay This Case If the Court grants the requested discovery, Plaintiffs almost certainly would seek to further 20 delay the case schedule (as they have done several times). If Facebook produced this overbroad 21 information, the additional discovery regarding that information that would surely follow—new 22 interrogatories and document requests, depositions and perhaps new expert designations—likely 23 would take another several months, at a minimum. There is no corresponding benefit to justify this 24 delay. 25 26 27 28 Gibson, Dunn & Crutcher LLP Plaintiffs’ argument that Facebook “cherry-picked” for a “sliver of configuration data that … is helpful to its arguments” (Dkt. No. 207 at 8) is meritless. Facebook produced the record that was directly relevant to the issue raised by Plaintiffs. Plaintiffs’ argument that the document “appears to reflect Facebook’s lawyer’s editing” is likewise baseless. As is a standard and appropriate practice for producing material from internal systems, Facebook exported the information to a file, a practice expressly contemplated by the agreed-upon ESI Order. (Dkt. 191 at 2; Dkt. 74 at 3.) 7 9 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 E. Facebook Has Already Agreed To Produce Proportional Discovery 2 In response to Plaintiffs’ overbroad and burdensome request, Facebook has proposed 3 reasonable alternative discovery, including configuration data actually called for by the source code 4 for the four practices challenged in Plaintiffs’ complaint. This is the appropriate scope of any further 5 discovery on this topic. See, e.g., Sharma v. BMW of North America LLC, No. 13-02274-MMC, 2016 6 WL 1019668, at *7-8 (N.D. Cal. Mar. 15, 2016) (limiting discovery to vehicle components actually at 7 issue). As Facebook explained in the parties’ meet and confers, the most efficient way to identify 8 relevant configuration data is for Plaintiffs to identify specific calls in the source code. (Poole Decl. 9 ¶¶ 7-9.) If Facebook were to try to guess what configuration data Plaintiffs would find relevant, its 10 engineers would undertake the same process (although without the benefit of knowing what 11 information Plaintiffs are seeking)—reviewing the relevant code. 8 12 In addition, despite Plaintiffs’ access to sufficient technical information about the practices at 13 issue, Facebook has offered to provide another Rule 30(b)(6) deposition concerning those practices. 14 These proposals are more appropriately limited to the proper scope of discovery than Plaintiffs’ 15 overbroad requests for all “configuration tables” (for which they cannot articulate any concrete need) 16 and are far more proportional to the needs of the case. V. 17 18 CONCLUSION Plaintiffs seek to compel information that is irrelevant and disproportionate to the needs of this 19 injunctive-relief-only class action. Facebook has offered reasonable compromises that will keep the 20 case on schedule and allow for the ultimate merits determination that the parties both desire. 21 Facebook respectfully requests that the Court deny Plaintiffs’ Motion. 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP Plaintiffs’ argument that requiring them to identify the relevant code calls would be “equivalent to requiring that Plaintiffs identify every relevant document in Facebook’s possession before it produces them” (Mot. at 10) is meritless. Plaintiffs have received extensive discovery, including the relevant source code, and it is not unreasonable to expect them to identify the relevant calls from the code. Facebook need not turn over every document in the company for Plaintiffs’ review. 8 10 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) 1 Dated: August 19, 2016 Respectfully submitted, 2 GIBSON, DUNN & CRUTCHER LLP 3 By: 4 5 /s/ Joshua A. Jessen Attorneys for Defendant FACEBOOK, INC. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP 11 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES”; Case No. C 13-05996 PJH (SK) Appendix A: Overbreadth of Requested Tables Table 9 (not configuration data) (not configuration data) (not configuration data) (not configuration data) s Scope Data Size Plaintiffs’ Purported “Need” Facebook’s Responses Some of these tables “may have” a “link between [a] log call and the “ ” table” and “may” lead to “additional use being made of the user ID and URL.” (Golbeck Decl. ¶ 18.) The “ ” table did not play any role in any of the challenged practices during the class period, and in fact, nothing was logged to the ” table during the class period. (Dkt. 18511 ¶¶ 3-4.) 10 In any event, the request is excessively overbroad. “[D]ata could be stored as an association, and these tables may contain information showing how the association is created.” (Golbeck Decl. ¶ 20) Plaintiffs need to know “if user IDs and URLs sent in private messages are being stored together and where.” (Id.) No relevant Associations are created during any of the challenged practices, and Plaintiffs already know that the EntShare reflects both the sender user ID and the URL sent in the message. In any event, the request is overbroad. There is only one Object at issue (EntShare). The site variable, which configures , is relevant to Facebook’s defense that its acquisition of EntShares occurs in the ordinary course of business. (Id. ¶ 25) This is the first time that Plaintiffs have identified this request for , and they did not meet and confer regarding this request. But there is no nexus to justify production of the remainder of the overbroad request for the whole table. For information in the “Table,” “Scope,” and “Data Size” columns, see Poole Decl. ¶¶ 11-13. Dr. Golbeck also mentions the scribeh_share_stats table, but does not even attempt to make any connection between it and any of Plaintiffs’ requests. Moreover, scribeh_share_stats is not related to any of the practices at issue and was mentioned for the first time in Dr. Golbeck’s rebuttal report in support of Plaintiffs’ reply in support of their motion for certification. (Dkt. 184-17 ¶ 28.) 9 10 DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF “CONFIGURATION TABLES” Case No. C 13-05996 PJH (SK)

Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.


Why Is My Information Online?