Oracle Corporation et al v. SAP AG et al

Filing 102

Joint Discovery Conference Statement by Oracle International Corporation, Oracle Corporation, Oracle USA Inc. (Alinder, Zachary) (Filed on 6/24/2008) Modified on 6/25/2008 (far, COURT STAFF).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

BINGHAM McCUTCHEN LLP DONN P. PICKETT (SBN 72257) GEOFFREY M. HOWARD (SBN 157468) HOLLY A. HOUSE (SBN 136045) ZACHARY J. ALINDER (SBN 209009) BREE HANN (SBN 215695) Three Embarcadero Center San Francisco, CA 94111-4067 Telephone: [phone] Facsimile: [phone] [e-mail] [e-mail] [e-mail] [e-mail] [e-mail] DORIAN DALEY (SBN 129049) JENNIFER GLOSS (SBN 154227) 500 Oracle Parkway M/S 5op7 Redwood City, CA 94070 Telephone: [phone] Facsimile: [phone] [e-mail] [e-mail] Attorneys for Plaintiffs Oracle Corporation, Oracle USA, Inc., and Oracle International Corporation

Robert A. Mittelstaedt (SBN 060359) Jason McDonell (SBN 115084) Elaine Wallace (SBN 197882) JONES DAY 555 California Street, 26th Floor San Francisco, CA 94104 Telephone: [phone] Facsimile: [phone] [e-mail] [e-mail] [e-mail] Tharan Gregory Lanier (SBN 138784) Jane L. Froyd (SBN 220776) JONES DAY 1755 Embarcadero Road Palo Alto, CA 94303 Telephone: [phone] Facsimile: [phone] [e-mail] [e-mail] Scott W. Cowan (Admitted Pro Hac Vice) Joshua L. Fuchs (Admitted Pro Hac Vice) JONES DAY 717 Texas, Suite 3300 Houston, TX 77002 Telephone: [phone] Facsimile: [phone] [e-mail] [e-mail] Attorneys for Defendants SAP AG, SAP AMERICA, INC., and TOMORROWNOW, INC.

UNITED STATES DISTRICT COURT NORTHERN DISTRICT OF CALIFORNIA SAN FRANCISCO DIVISION ORACLE CORPORATION, et al., Plaintifs, v. SAP AG, et al., Defendants. Case No. 07-CV-1658 PJH (EDL) JOINT DISCOVERY CONFERENCE STATEMENT Date: July 1, 2008 Time: 11:00 a.m. Courtroom: E, 15th Floor Judge: Hon. Elizabeth D. Laporte

JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Pursuant to the Court's May 29, 2008 Order Following Discovery Conference, Plaintiffs Oracle Corporation, Oracle USA, Inc., and Oracle International Corporation (collectively, "Oracle") and Defendants SAP AG, SAP America, Inc., and TomorrowNow, Inc. (collectively, "Defendants," and with Oracle, the "Parties") hereby submit this Joint Discovery Conference Statement. Although the Parties have had multiple meet and confer sessions, and have made some progress on discovery, there are several issues related to the scope of discovery that remain unresolved and which will require formal briefing and resolution by the Court. The Parties agree that the Court should schedule ninety minutes on Tuesday, July 1 to discuss these discovery issues, as well as to hear argument on Defendants' pending appeal of two of Judge Legge's previous discovery rulings. 1. Discovery Limitations

Consistent with discussion between the Parties and the Court at the May 28, 2008 Discovery Conference, the Parties have continued to meet and confer, through detailed correspondence and in hours long telephonic conferences on June 6, June 12, June 20, and June 23, 2008, in an effort to reduce the volume and enhance the usefulness of each side's respective document productions. The overall status of those discussions is as follows: · While the Parties have agreed that restrictions on the number of custodians make sense in this case, they have not been able to come to agreement on the appropriate number of custodians for each side or on the specific arrangements that would allow flexibility beyond custodian limits (e.g., targeted searches, the mechanism and limits on adding custodians beyond the initial limits, whether and how inferences can be drawn or extrapolations made in lieu of full discovery). · On the issue of search terms, Oracle has developed a list and contends that testing confirms the list captures virtually all responsive, non-privileged documents. Defendants, who contend they are working with larger document volumes, have not yet been able to develop a list that captures virtually all responsive, nonprivileged documents, but the Parties continue to meet and confer in an effort to reach a compromise on this issue. 1
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

·

The Parties have been negotiating over which categories of financial and damages documents should be produced and prioritized so that the Parties can begin calculating damages in a meaningful way. Unfortunately, the type and scope of such documents remains in dispute.

·

Defendants have implemented a method for remote server access and production (the proposed "Data Warehouse Agreement") for a number of servers and computers maintained by Defendants, in an attempt to reduce the burden of production of these materials. Defendants provided Oracle with a detailed protocol and related procedures to gain that remote access on June 18, 2008 and the Parties expect to finalize the logistics of that arrangement so that Oracle will begin its review of the Data Warehouse in mid-July.

The following sections discuss the Parties' respective positions on (1) Custodian Limits, (2) Targeted Searches, (3) Search Terms, (4) Use of Extrapolation or Inference, and (5) Resolution of Damages Causation Evidence, along with a Joint Statement as to the Data Warehouse Agreement and each Party's report on Likely Discovery Motions. Because the Parties have not reached agreement, formal briefing to and decision by the Court on these issues is required. Per the Court's instruction at the last conference should agreement not be reached, the Parties will ask the Court to set an expedited hearing and briefing schedule at the upcoming Discovery Status Conference. a. Oracle's Position

To support Defendants' request that they should be excused from searching for and producing admittedly relevant documents, at the last Discovery Conference, Defendants provided the Court with unsubstantiated dollar figures regarding their discovery efforts. Oracle is seeking detail behind those figures 1 but, regardless, cost of discovery is only one factor a court must On June 19, Defendants provided a limited summary without any back-up, and asserted that review of their most highly involved custodians (principally board members and the SAP managers responsible for the SAP TN relationship and strategy) ­ none of whom has yet been produced ­ would incur even higher average costs. Given the importance of that number to the Court's evaluation of proper discovery limits, Oracle is seeking detailed back-up in advance of the expedited briefing on discovery limits. See also, e.g., Columbia Pictures Indus., Inc. v. Brunnell, 2007 WL 491696 (C.D. Cal. 2007) (vague assertions of "undue burden" are not sufficient, especially when a party fails to submit any evidentiary declarations supporting such objections). 2 JOINT DISC. CONF. STATEMENT
Case No. 07-CV-1658 PJH (EDL)
1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

consider in making a decision under Rule 26(b)(2)(C) to preclude a party from obtaining admittedly relevant discovery. Rule 26(b)(2)(C)(iii) requires a balancing of the burden or expense of discovery against the likely benefit, "considering the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the action and the importance of the discovery in resolving issues." The Court must consider all these factors and weigh them before deciding which limits to impose and how to balance those limits, including through the use of inference and extrapolation. Oracle believes its proposed custodian limits are reasonable, taking into account all of the Rule 26(b)(2)(C) factors. It provides a brief summary of its position here and will provide a more in-depth analysis of these factors in the formal briefing to the Court. Needs of Case: The burden of discovery on Defendants is a result of the astonishing scope of their wrongful conduct here ­ potentially millions of illegal software downloads, thousands of infringing software environments, and many hundreds of stolen customers ­ a business model premised on misappropriation that permeates the entire SAP TN subsidiary and taints every customer gain. Defendants committed these wrongs with permission from their highest-level executive board of directors and from the CEOs of the respective entities. Only Defendants and certain third parties hold the information that will reveal the scope of what Defendants have done, how they covered it up, and how they have benefited. Thus it is not surprising that the evidence in Defendants' possession is (and should be) disproportionate to Oracle's. Amount in Controversy. Because Defendants have not provided Oracle with critical information relevant to liability and resulting damages, Oracle does not yet know its damages with precision. Indeed, every key SAP custodial document production and every deposition of Defendants reveals more infringement of Oracle's intellectual property -- and more complicity. What the long-awaited access to SAP TN's Data Warehouse will reveal remains to be seen. But, even so, it appears Oracle's damages are, at a minimum, well into the several hundreds of millions of dollars and likely are at least a billion dollars.

3

JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
2

Objecting Party's Resources. SAP is one of the largest software companies in the world. SAP's market cap is approximately $63 billion and it reported net income of 1.9 billion (approximately $3.2 billion) in 2007. These resources undermine Defendants' claims of financial hardship from providing adequate discovery. Importance of Issues. Oracle has brought serious claims of intellectual property theft, including claims for violation of criminal statutes, infringement of numerous copyrights, and a cover-up by SAP's executive board of directors. The United States Department of Justice has opened a related investigation and sought information from SAP. At stake are the rules of fair play between the two leaders in an industry estimated to generate $190 billion in revenue this year. (1) Custodian Limits

Defendants' Limit Is Arbitrary And They Will Not Provide Adequate Safety Nets. Defendants admitted before the May 28, 2008 Discovery Conference that approximately 3,000 of their custodians may possess potentially relevant information. At the Discovery Conference, Oracle proposed an initial limit of 209 Defendant custodians (46 from the SAP Defendants and 83 from SAP TN), less than 1% of that total number, with safety nets of up to 40 custodians to accommodate the new claims and scope of allegations in the upcoming Second Amended Complaint and up to 40 to cover sales-related customer issues. 2 Defendants came into the last Discovery Conference with a hard custodian limit of 115 total custodians divided among the three Defendants and have not budged since. They do not and cannot confirm that production from this number of custodians will be adequate to expose the scope of their liability nor the extent of Oracle's damages. In contrast with Defendants' refusal to change position, Oracle has attempted to The SAC includes allegations, based on documents and testimony to date, that the SAP executive board of directors knew about SAP TN's illegal business model, but approved the purchase of SAP TN anyway and allowed it to continue operating illegally to get a public relations win and convert Oracle customers. It also includes allegations that SAP engaged in its illegal activities after the filing of litigation (evidence which is critical to Oracle's intent and punitive damages claims), did so related to other Oracle products (such as Siebel), and that SAP TN has been stealing Oracle's software as far back as at least 2002. None of these areas has been the subject of focused discovery to date. Oracle discovered information about them as a byproduct of its initial discovery focused on the original downloading allegations. 4
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

accommodate Defendants' desire for production from fewer custodians by proposing a variety of compromises, trade-offs, and safety nets, most of which Defendants have refused. Specifically, Oracle has proposed (a) extrapolation and/or inference from the more limited custodial production and information set (particularly as to number of infringing environments and infringing derivative works distributed by Defendants to their customers); (b) the elimination of sales custodian searches in exchange for agreement that Defendants are estopped from contending that Oracle cannot prove the causation element of its liability claims (since Defendants are the ones who have insisted from the beginning of the case that the sales custodians contain this critical lack of causation evidence), or alternatively, agreement from Defendants to provide comparable damages causation evidence to what Oracle has produced and will produce on top of their custodial production; (c) a limited number (10) of robust targeted searches; (d) a reasonable number (25) of additional custodians that Oracle can automatically name and receive farther along in discovery along with the ability to seek others should Oracle show good cause to the Court; and (e) the ability to select and prioritize Defendants' custodial productions. Thus, in contrast to Defendants, Oracle has approached the issue of custodial limits with creativity and flexibility, reducing its initial request for 209 Defendant custodians to 140 provided safety nets are put in place. Oracle explains the need for each of its safety nets and provides Defendants' responses below. The Parties Have Agreed Oracle Can Select Custodians and Get Production Quickly. Before the issue of arbitrary custodian limits surfaced, Defendants had already self-selected and reviewed a large percentage of the total 115 custodian limit that they now propose ­ mostly individuals from SAP TN. This concerned Oracle because, based on its experience to date, Oracle rightly feared that Defendants would use up their custodial production slots on the wrong targets and produce custodians in a sequence designed to avoid detection of relevant information (indeed, their production strategy to date had shielded the SAP corporate parents from scrutiny, exactly at the time that their own involvement and cover-up has just come to light). In response to this concern, Defendants stated at the last Discovery Conference that they would welcome Oracle's choice of custodians who made up the eventual number of custodians. 5
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Oracle took Defendants up on their offer, providing counsel with a specific list of nine custodians (namely those board members and high level executives that discovery to date indicate are most involved and knowledgeable) and requesting that Defendants immediately prioritize production from them. Based on Defendants' representations to the Court that they were producing an average of 10 custodians a month, Oracle also served notices of those individuals' depositions allowing for such a scheduled rate of production followed by review of produced documents. Initially, Defendants balked. They provided a list of eleven other custodians already in their production queue for June ­ none of whom Oracle had selected and some of whom appear only marginally relevant. Oracle asked them to stop production from all but one and to focus on the custodians on Oracle's list. Defendants agreed. However, they promised production for just these priority custodians only by late fall, and completion of their depositions thus cannot occur until Thanksgiving at the earliest. This calls into question Defendants' ability to produce the remaining custodians and targeted searches. Oracle needs more certainty that it will receive documents in a timely fashion. It will ask the Court to order Defendants to produce, at minimum, at the 10 custodian per month rate Defendants asserted it was meeting at the last Discovery Conference. This assurance is the only way that Oracle can be sure it can explore the most meaningful witnesses in the foreseeable future and get full production of the limited custodians in adequate time before the discovery cut-off. The Number of Custodians Produced by Defendants Should Exceed that Produced By Oracle. The relevant information in this case largely resides with Defendants: they did the misdeeds and they reaped the rewards. Moreover, there are three Defendants: the two corporate SAP parents and the SAP TN subsidiary ­ with three separate headquarters, numerous other international offices, different servers, different managers, and different SAP AG and SAP TN boards of directors (SAP America has no board). In contrast, Oracle is effectively one company. 3 Oracle's relevant information and custodians are essentially limited to its copyright registrations,
3

Oracle International Corporation is the holding company for Oracle's intellectual property, which Oracle USA licenses to customers, and both of which Oracle Corporation controls. Accordingly, in contrast to the three Defendants, Oracle has one headquarters in Redwood City, California, and is one customer-facing entity, e.g. with one human resources department and one sales and marketing team. 6
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

its relevant customer licenses, and the revenue streams reasonably associated with its customers who left for SAP TN (though Defendants seek much more). 4 Oracle needs to spend the limited time allotted to discovery reviewing Defendants' relevant documents and taking Defendants' depositions, not producing irrelevant information. Thus, whatever custodial limits are imposed on Defendants, Oracle requests that it be ordered to produce between one third and one half of Defendants' total. (2) Targeted Searches

The Parties agree that certain types of information should be specifically looked for from likely sources, such as a centralized database or from the person most likely to have such information ­ rather than assuming such information would be revealed in custodial productions. Oracle views this as an important safety net against the limited number of custodial productions. Oracle has suggested each Party be able to propound 10 targeted searches, with the ability for good cause shown, to seek the Court's order for more. In mid-May, two weeks before the last Discovery Conference, Oracle provided Defendants with an initial list of targeted searches it wanted assurance from Defendants they were providing in addition to custodial searches: (1) board materials; (2) financial information; (3) customer contracts for customers (and communications with them) claimed as wins in Defendants' Safe Passage program; (4) Defendants' reports on customer wins, losses and those at risk (analogous reports to reports Oracle found by targeted search and already produced to
4

For instance, Defendants have sought all of Oracle's documents concerning its competition with SAP and all of Oracle's financial documents (at the general ledger level) ­ discovery far afield from the products and issues in this case. See, e.g., Defendants' RFP No. 91 (seeking all documents relating Oracle's strategy to offer a broad product line "to compete against SAP AG"); Defendants' RFP No. 81 (seeking all documents related to "Project Fusion"); Defendants' First 30(b)(6) Notice of Deposition of Plaintiff Oracle Corp. Topic 6 (seeking to depose a witness on Oracle's general "Sales, marketing, or competitive intelligence . . . relating to SAP and/or TN"). Defendants have also sought all of Oracle's financial documents, at the general ledger level, to support every aspect Oracle business, including all "new software license revenue and expenses," "On Demand revenue and expenses," all "departments" of the PeopleSoft organization, and Oracle's "executive" department. See, e.g., June 6, 2008 Letter from Mr. McDonell to Mr. Howard. And, Defendants have repeatedly requested information about Oracle products and services completely (and admittedly) unrelated to the products at issue in this case. See, e.g., Defendants' RFP No. 67 ("Documents sufficient to show Oracle's revenues, costs, and profit margins for products other than those referred to in the Complaint or at issue in this litigation, and services relating to such products."). 7
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Defendants); (5) documents related to the acquisition of SAP TN (including business cases/financial projections/risk analysis/board reports); (6) relevant policy and procedure documents; and (7) documents related to Project Blue and the issues considered as part of it. On June 19, Defendants stated in meet and confer correspondence that they were willing to perform some of these targeted searches, but not all. For instance, they refuse to look for specific lost customer-related communications as part of a targeted search but instead seek to charge against their custodian limit any searches for specific lost customers; they refuse to produce all acquisition or Project Blue documents but want to rely on the production of such documents from custodians. Meet and confer continues on the specifics of what Defendants will do and which of these categories is not appropriate for targeted search. Oracle has already performed many robust targeted searches in responding to Defendants' discovery requests (including from approximately 30 to 40 additional custodians and about 10 centralized information sources). In early June Oracle asked for and on June 19 it received, Defendants' proposed additional targeted search list, namely (1) copyright information, (2) IP due diligence in connection with Oracle's acquisition of PeopleSoft, (2) PeopleSoft IP due diligence, (3) all third party "At Risk" reports, (4) all relevant policy and procedure documents, (5) all communications between SAP TN and Oracle, (6) customer files, (7) sales and customer information comparable to those Defendants are agreeing to produce through targeted searches. While the Parties are meeting and conferring on specifics, and Oracle does not agree, e.g., to produce topic number (5), which is one of the subjects of Defendants' current appeal, or to produce "complete" customer files for topic number (6), which is not a targeted search but a request for review of all custodians related to those customers, 5 Oracle has agreed to the concept of appropriate and reciprocal targeted searches but only if they complement the limited custodial production (and not take up slots within it). The timing of production of these targeted searches needs to be resolved. Oracle believes
5

Additionally, other than some minor clean-up, Oracle has already given Defendants complete "customer files" from the central repositories that Oracle maintains, and, therefore, has already completed the collection, review, and production of documents included in Defendants' sixth topic, to the extent it is a targeted search. 8
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

that early production and evaluation of Defendants' targeted search results will impact Oracle's ability to assess the adequacy of Defendants' limited custodial production. Thus, Oracle will request the Court order a date certain for the Parties' production of targeted search materials that does not impact the production of custodial searches. (3) Search Terms

For months, Oracle has argued that the Parties can and should use search terms to limit the burden of document review in this case without unduly sacrificing relevance. Oracle's arguments are based on the set of search terms related to its sales custodians it developed during review which, when applied, limited their documents by approximately 20%. Before doing so, Oracle validated these terms against the custodians, which had already been reviewed on a document-bydocument basis, and, after careful refinement, the terms hit 100% of the non-privileged documents that had been tagged for production. Since then, the Parties have met and conferred several times to discuss the possible use of search terms across all documents. In contrast to Oracle's sales custodian list of 548 terms, Defendants initially proposed a list of 72 terms. Defendants conceded that their list did not hit all of the responsive, non-privileged documents. Accordingly, through painstaking analysis and validation and with specific reference to Defendants' list and their Requests for Production, Oracle developed a new list of search terms to be applied across the documents for all Parties. Oracle then validated this list against six fully reviewed representative Oracle custodians. These search terms eliminated from 13 to 47% (for an average of 23%) of the total documents for these custodians without missing any non-privileged documents that had been marked for production. 6 Defendants claimed that Oracle's list did not limit their production other than by a few percent. Oracle then refined the list further, validating against its production first, and then proposed the new list to Defendants. On June 6, 2008, Defendants responded that this new list was insufficient, because it only limited their production by 4 to 28%. Thereafter, on June 12, Oracle provided comparable search term statistics and data to Defendants, and even coached Defendants' counsel on how to better refine their search terms
6

The Parties have agreed that documents that cannot be searched in this fashion (e.g., images, voicemails, etc.) shall be pulled out and separately reviewed for production. 9 JOINT DISC. CONF. STATEMENT
Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

in an attempt to get the same 100% hit rate Oracle has achieved. 7 On the meet and confer call on June 23, 2008, Defendants stated that their further efforts to refine the list had increased their hit rate (though still not to 100%), but decreased the amount of excluded documents from 18% to 13%. Oracle agreed to work with Defendants to further refine this search term list to increase the hit rate on responsive documents, while also reducing the total number of documents to review. The Parties agreed to the continue this process for at least another week to ensure that every effort is made to arrive at a validated list of search terms for both sides. Until Defendants can validate a set of search terms that hit all truly relevant, nonprivileged documents marked for production in a representative set of custodians, they should be required to continue their review on a document-by-document basis ­ though their complaints about cost should be considered against this inability to develop and validate a set of search terms. 8 On the other hand, Oracle has already validated these search terms against its production and found that the terms represent an effective and reasonable method of limiting its production, hitting every non-privileged document marked for production. 9 Regardless of whether Perhaps one reason Defendants have not been able to capture all produced documents when using their search terms is that they are vastly over-producing irrelevant documents. For instance, Defendants recently produced numerous emails about games they are playing at work (e.g., TNOR02152326), travel websites (e.g., TN-OR01576660, TN-OR01576627), irrelevant work travel (e.g., TN-OR00869131), inquiries about each other's health (e.g., TN-OR02168485), ESPN.com webpages (e.g., TN-OR00932600), pictures from a personal myspace.com website (e.g., TNOR00932409), a video of a street performer doing tricks (e.g., TN-OR00925348), and a humorous discussion of an employee's genitalia (e.g., TN-OR02152350). Amazingly, all of the above sample documents were designated either Highly Confidential or Confidential under the protective order in this case, demonstrating review and decision-making by an attorney. This over-production may also explain the high cost of their production.
8 7

While the Parties agree that the terms should be run in both English and German to ensure that the terms also hit relevant German documents, because the terms were developed in English only, the Parties will also need to validate the terms translated in German before applying the terms to any German custodians.

9

Oracle will continue to provide Defendants with information about its search term protocol, results, and analysis ­ indeed, it has in many cases provided more detailed information than have Defendants. Oracle must rely on Defendants' representations of numbers of responsive documents hit and missed as much as Defendants must rely on Oracle's. 10
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Defendants can validate these or another similar set of search terms, Oracle's validation process is complete and it will apply these search terms across its entire review set going forward. (4) Use of Extrapolation or Inference

At the Court's suggestion during the last two discovery conferences, Oracle has researched the use of inferences and/or extrapolation to help reduce the discovery burden for the Parties. The use of these tools can reduce discovery burdens and simplify the presentation of enormous quantities of evidence for trial. See, e.g., Fed. Jud. Ctr., Manual for Complex Litig. (Fourth) §§ 11.422, 11.493 (2004) (when it is necessary to limit discovery, "statistical sampling techniques [may be useful] to measure whether the results of the discovery fairly represent what unrestricted discovery would have been expected to produce," and noting that "[t]he use of acceptable sampling techniques in lieu of discovery and presentation of voluminous data from the entire population, may produce substantial savings in time and expense"); Smith v. Lowe's Home Centers, Inc., 236 F.R.D. 354, 357-58 (S.D. Ohio 2006) (finding that "limiting discovery to a statistically significant representative sampling . . . will both reasonably minimize the otherwise extraordinary burden imposed on the plaintiffs and their counsel and yet afford the defendant a reasonable opportunity to explore, discover and establish an evidentiary basis for its defenses"). The use of inferences could significantly reduce the discovery burden and ease presentation of evidence at trial in at least one important area: Defendants' use of environments to create infringing derivative works/distribution of derivative works. Defendants have thousands of these environments on their servers. The burden and difficulty of reviewing, producing, and presenting at trial the terabytes of this software evidence contained on Defendants' servers could be significantly reduced by extrapolating from a sample set of environments and updates created with those environments. This extrapolation evidence could also reduce the need to go out to each customer who received these infringing products, by creating a baseline derived from a sample set of customers. Oracle has not yet had any discovery of Defendants' actual infringing software which limits its ability to commit to a statistical approach. Defendants' proposal to provide Oracle with remote access to the enormous quantity of electronic evidence (mostly Oracle's own software, as described during the Technology Tutorial 11
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

to the Court) on Defendants' systems is an offer that can only be characterized as limited and/or unknown. Contrary to Defendants' claims, remote access to Defendants' Data Warehouse is not "production" of those materials. Accordingly, Oracle has reserved its rights to demand full production of forensic images of those servers and computers. Moreover, access to these servers has not even started, and as Defendants concede, production of the materials from those systems won't even begin until after (a) Oracle can review them, and (b) Defendants can re-review them to confirm that they agree that the requested materials should be produced. Only then, likely many months out given the quantity of data, will Defendants even begin to produce this evidence. Defendants cannot both limit the review and production of these highly relevant materials and also refuse to extrapolate from the limited subset that they do agree to produce. For its part, Oracle intends to push forward with the remote access review of Defendants' Data Warehouse as quickly as possible, given the many constraints, and expects that this process will assist in determining the optimum sampling or extrapolation approach. (5) Resolution of Damages Causation Evidence

Another issue that needs to be addressed in assessing Defendants' request for limited custodian production is how to secure to Oracle the same type of damages causation evidence that it has provided to Defendants. To date, Oracle has produced sales-related evidence from individual sales custodians (including 10 sales personnel, 1 support sales manager, and other executive-status sales personnel), as well as from three central repositories containing license and support sales documents (such as "OKS," a support contract management database used at Oracle to create, manage, and update service agreements and renewals, and to store documents and correspondence related to those agreements and renewals). In addition, Oracle has produced lengthy "At Risk" reports that conveniently summarize the details about the deals of customers who went to or were at risk of going to third party servicers such as SAP TN, as well as compiling the reasons for these moves from the underlying sales personnel. Defendants will use this commentary and the information from Oracle's sales custodian productions to date to attempt to undermine Oracle's damages claims for specific customers. To date, Oracle has not received comparable information from Defendants, and 12
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Defendants' limited production offers for the future will not provide Oracle with such information. Critically, there is no comparable summary report from Defendants on why Oracle customers did or didn't go to SAP TN or how important SAP TN was to their decisions to buy SAP applications, which would substitute for production from the underlying sales custodians (as Oracle's "At Risk" reports do). SAP also has not produced any sales custodians material on any of its claimed 800 SAP Safe Passage deals, included for the limited subset of those where the customer purchased service from SAP TN; nor has SAP produced such materials for any preexisting SAP TN customers who purchased services or applications from SAP after SAP acquired SAP TN. Indeed, Defendants have not even provided the list of that subset of these concededly relevant SAP/SAP TN customers. The Court recognized the unfairness of this situation at the last discovery conference. Either Defendants have to agree that they will not use the damages causation evidence Oracle has provided or they have to agree to provide comparable information, even if it means going to their sales force to get it through targeted customer-specific searches. Should Defendants choose the latter resolution of this dilemma, such production must be on top of Defendants' proposed custodian limits or those slots will be quickly and unfairly depleted. b. Defendants' Position

Oracle's complaints about "fair play" are belied by its own rhetoric about alleged damages. Oracle speculates wildly about the amount of its damages "claim" in this discovery report, even though more than a year after this case was filed, Oracle still refuses to identify with any precision the nature or amount of its alleged harm or even to provide the theory on which its damages claim is based. Oracle wants to substitute public posturing for the hard work of articulating and proving its damages claim (on which Oracle bears the burden of proof). Defendants should not be forced to undergo millions in extra expense based on hyperbole. Moreover, Oracle's proportionality argument ignores several key facts. First, this case is set for trial on February 8, 2010, with fact discovery closing on June 19, 2009. Judge Hamilton has stated that these dates will not be extended. Dkt. No. 78, at 1. Even if Defendants were to agree to Oracle's proposed number of custodians, there is no practical way for Defendants to 13
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

produce such an enormous volume of documents and still remain on ­ or even close to ­ the schedule set by Judge Hamilton. Nor, as this Court observed at the last discovery status hearing, is it possible to present such voluminous information to a jury. Defendants agree, therefore, that cost is not the only factor to be considered in setting discovery limits. Time constraints are equally important. Second, Defendants' proposals are intended to limit cumulative discovery, not to deprive Oracle of the opportunity to prove its case. Defendants have already produced around 2.3 million pages of documents from 42 custodians. Defendants' conservative estimate of the number of additional pages to be produced under its proposed limit of 115 custodians is around 4 million. That is an estimated total of around 6 million pages to be produced under Defendants' proposed limits, and that does not include the additional 6 terabytes of data already produced in native form and non-custodian based documents and information to be produced from central repositories and the like. If Defendants' alleged wrongdoing is as pervasive as Oracle claims, that surely is enough discovery to allow Oracle to present its case. Third, Defendants' burden objections are neither vague nor unsubstantiated. Defendants have provided Oracle detailed information concerning the cost of production to date and will provide more, if necessary. As Defendants' counsel represented to the Court, that cost has averaged around $100,000 per custodian. Even under Oracle's new proposed limit of 165 custodians, that amounts to $16.5 million in costs for just one portion (custodian-based documents) of one form of discovery (document production). Finally, Oracle misrepresents Defendants' positions during the meet and confer process. As discussed in detail below, Defendants have considered and agreed to many of Oracle's proposals. Aside from the number of custodians (which, for the reasons stated herein, Defendants believe should not be increased from 115) and from patently unreasonable proposals by Oracle (such as foregoing discovery on causation of damages) Defendants have been reasonable and flexible and believe that substantial progress has been made on many issues.

14

JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

(1)

Custodian Limits

Defendants firmly believe that an objective numerical limit on custodians is necessary given the extremely burdensome costs associated with the custodian productions, regardless of the amount of time left in the discovery period. As Defendants previously explained to the Court, the average cost to date per custodian for Defendants to review and produce documents is over $100,000, 10 and that figure does not include the substantial costs Defendants have incurred in collecting, preserving, and producing the electronic data that does not fall within the custodian category. Further, to date Defendants have produced on average 10 custodians a month. The number of custodians that can be produced per month necessarily depends on the average number of documents the custodian has in his or her files. Defendants' current ability to produce 10 custodians per month is based on the custodians having on average 48,000 documents in their files. That average, of course, can and will fluctuate significantly for any given custodian, and, in the end, it is the total volume of documents reviewed per month that determines how many custodians can be produced in any given month. Defendants currently have in excess of 50 lawyers continually reviewing custodian files for production. For all of the reasons previously explained to the Court, Defendants continue to maintain that a limit of 115 custodians per side is appropriate. Defendants also continue to request that additional custodians beyond that limit may only be added by leave of Court with good cause shown. Defendants agree with Oracle's request that each side be allowed to choose which custodians' material must be reviewed and in what order that review takes place, subject to logistical constraints such as the volume of data per custodian and any technical difficulties that may arise during the process. This selection process has already begun with the Parties agreeing on the names and order of the next 12 custodians whose documents Defendants will review and

Defendants have provided Oracle with a breakdown for, and related written and oral description of, the components of the $100,000 per custodian average cost. Defendants have further offered to provide Oracle with additional back up information, provided that Oracle first provides a specific list of the type of back up information it seeks. To date, Oracle has not provided its promised list so that Defendants can respond to Oracle's request for additional back up. 15
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

potentially produce. Defendants have also agreed with Oracle that each side may "reserve" custodians within the numerical limit to identify as discovery progresses and warrants. Oracle's concern that its custodian slots will be used up on the wrong targets and its claim that that Defendants have "self-selected" custodians and sequenced production to avoid producing the relevant information are unwarranted. The order of production has largely been dictated by Oracle's lists of priority custodians (most of whom, until Oracle's most recent list, were TN employees) and by the fact that this case is primarily about TomorrowNow, not SAP. It is hardly surprising, therefore, that most of the 42 custodians produced to date (18 of whom were selected by Oracle) are TN employees. When Oracle recently provided Defendants its new list of 12 priority custodians, all from SAP, Defendants promptly agreed to stop production of the custodians then in queue for production and focus their efforts on Oracle's new list of custodians. Given the extraordinarily high costs associated with custodian discovery and Defendants' willingness to work with Oracle to produce responsive documents through means other than by custodian production, a limit of 115 custodians is proportionate for this case. Further, Defendants have been working with Oracle on implementing proposed "safety nets," that it contends are necessary to provide some limited relief from the numerical limits. Finally, Defendants strongly disagree with Oracle's contention that the custodian limit should not be mutual as between the three Plaintiffs (which it attempts to argue is just one Plaintiff) and the three defendants in this case. Oracle surfaced this argument since the last hearing in an attempt to increase the number of custodians Defendants must produce, but limit the number that Oracle must produce. This "proportionality" (in reality ­ very disproportionate) argument is patently absurd. Contrary to what Oracle argues, there are three plaintiffs in this case, Oracle Corporation, Oracle International Corporation, and Oracle USA. Plaintiffs have set up their corporate structure in such a way that they believe all three entities are necessary parties to this case. Oracle has a market capitalization of $114 billion and approximately 75,000 employees worldwide. Therefore there is neither a legal difference in the number of parties per side nor a gross disparity in the relative size or global reach of the Parties. Moreover, Oracle's claims in this case relate to products developed not by Oracle but by PeopleSoft, which Oracle 16
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

acquired in January 2005, and J. D. Edwards, which PeopleSoft acquired in June 2003. Thus Defendants' discovery of Plaintiffs relates not only to the documents and information of the three Oracle-entity Plaintiffs but to two additional large software companies whose historical documents and information are now, by virtue of a stock purchase, in Oracle's custody. Finally, such a lopsided limit is contrary to the spirit of the Federal Rules of Civil Procedure, which set objective discovery limits per side not per party. Likewise, the District Judge has imposed equal, not disproportionate deposition limits and written discovery limits in this case. Accordingly, there is no basis to start applying discovery limits unevenly in this case, and whatever number the Court sets as the objective numeric limit on custodians should be applied equally to both sides of this case. (2) Targeted Searches

Defendants agree that the use of "targeted searches" is appropriate in addition to custodian searches. Defendants consider a targeted search to be a search of reasonably well-defined information that is located in common files and that are not the personal working files of a particular individual (e.g., a search for financial records maintained by a finance department or customer files maintained by a sales and marketing department). Targeted searches can also include limited searches of a custodian's files (e.g., search of an individual's files for communications with a particular company, as opposed to all search terms or a document by document review that is part of a full custodian search). In light of the enormous time and expense burdens of the full custodian searches in this case, there should be reasonable limits on targeted searches to ensure that the discovery burdens are proportionate. Defendants propose that each side should be able to propose targeted searches to the other side consistent with their discovery requests. Defendants believe that a presumptive limit of ten reasonably specific targeted searches per side is reasonable. Defendants have initially identified seven such targeted searches as follows: (1) copyright information; (2) financial information; (3) any analysis by Oracle of the intellectual property assets of PeopleSoft and/or JDE prior to, subsequent to, or in connection with its acquisition of PeopleSoft; (4) information concerning the reasons it lost customers to TomorrowNow; (5) policy 17
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

and procedure documents; (6) Oracle's communications with or about TomorrowNow; 11 (7) complete customer files regarding customers lost to TomorrowNow. Defendants have not, as Oracle claims, refused to do targeted searches on acquisition or Project Blue documents or on communications relating to lost customers. Defendants have informed Oracle that the acquisition and Project Blue documents are largely, if not completely, covered by the custodians whose productions Oracle has received or will receive. Defendants do not believe there are any such documents that are not custodian-based, but to the extent there are, Defendants have agreed to search central repositories for them. Defendants have also agreed to targeted searches relating to lost customers, as discussed further below. (3) Search Terms

The Parties have exchanged lists of search terms in order to narrow the scope of a custodian's electronic documents to be reviewed. Initially, Oracle's search term list contained 978 search terms. Defendants' first draft of its search term list contained 72 search terms. When Defendants ran Oracle's original list of search terms over nine sample custodians that were previously produced, the terms eliminated only 8.2% of the data to be reviewed and still missed 0.4% of Defendants' responsive documents. Defendants' list of 72 search terms resulted in the elimination of 20.6% of the data to be reviewed and missed 10.3% of Defendants' responsive documents. After this initial process, the Parties continued to meet and confer regarding the list of relevant search terms and their respective testing and validation of such terms. Defendants agreed to compromise by adding to their search term list and refining Oracle's list of 978 terms. Defendants have spent numerous days trying to create and evaluate a list that provides for the highest possible reduction in the data to be reviewed while capturing an acceptably high percentage of responsive documents. However, the Parties disagree on what constitutes an acceptably high percentage of responsive documents. Oracle has made it clear through the meet and confer process that it is interested in nothing less than 100% capture of responsive documents These communications between Oracle and TomorrowNow are subject to Defendants' pending objection to the Special Master's recommendation, which is scheduled for hearing at the same time as the July 1, 2008 discovery conference. 18
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)
11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

through the use of search terms. However, in the most recent conference, Oracle indicated that something less than 100% might be necessary and inevitable with Defendants' production. Defendants believe after much testing and analysis that it will be difficult, if not impossible, to locate a search term list that will capture all responsive documents while still eliminating a sufficient amount of data from review. Oracle's position is that it has adequately tested and validated a set of search terms that captures all responsive documents for the six sample custodians Oracle used. Oracle has not provided Defendants with the same level of documentation Defendants have provided to Oracle with respect to Defendants' search results. Thus, Defendants must rely in good faith on Oracle's representation that it has captured all responsive documents through its proposed search terms. Further, Oracle's statement that it has captured all responsive documents ignores the fact that Oracle has objected to several categories of documents that Defendants believe are responsive. See Part 2.b. below. 12 Defendants have thus far been unable to identify a set of search terms that captures all responsive documents for the nine sample custodians Defendants used in its testing. During the June 12 meet and confer telephone call, Oracle's counsel generally explained the methodology Oracle employed to retrieve all responsive documents with the use of search terms, but Oracle was unwilling to review the entirety of Defendants' responsive documents not captured. Since the June 12 meet and confer, Defendants have utilized Oracle's stated methodology to attempt to capture all responsive documents for the nine custodians. Defendants reviewed the responsive documents that were not captured by the original search terms and identified additional terms that could be used to attempt to capture all responsive documents. Regarding Oracle's comments in footnote 7 above, it is inevitable that some nonresponsive documents will slip through the review process in a production of this magnitude. Defendants could point to similar examples in Oracle's minimal production but will not waste the Court's time with that exercise. Interestingly, however, the document Oracle references relating to genitalia is responsive to several of Oracle's requests relating to the use of Software and Support Materials because it relates to the copying of DVDs. See, e.g., Oracle RFP No. 65 (requesting "all Documents related to that Use, including without limitation all Communications relating to that Use"). Another document cited by Oracle relates to the travel date and location of the former CEO of TomorrowNow and is relevant to his participation in potential meetings during the applicable time frame. See TN-OR00869131. 19
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)
12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Through this process, Defendants have successfully captured a substantial number of additional responsive documents. The percentage of responsive data missed has dropped from 4.2% to 1.3% for the nine custodians through use of additional search terms. While it is

Defendants' ultimate goal to increase the percentage of reduced data by eliminating words and phrases from the search term list, Defendants have been willing to add additional terms in an effort to capture more responsive documents. However, the additional search terms caused the percentage of reduced data to be reviewed to fall from 18.4% to 13%. Defendants are willing to continue refining the search term list in an effort to capture the remaining responsive documents, but it is likely that, at the end of the iterative testing and search term refinement process, some responsive data will not be captured through the use of search terms. In the meantime,

Defendants disagree with Oracle's position that it may start using search terms now without a reciprocal agreement. (4) Use of Extrapolation or Inference

During the May 6, 2008 discovery hearing, the Court suggested that the parties consider the use of sampling and inferences from that sampling as a means to reduce discovery burdens. It has been and remains Defendants' position that sampling and inferences may be appropriate for certain limited purposes in this case, but only if the process is fair, mutual and defined with sufficient precision so that Defendants will clearly understand: (1) what baseline evidence will be used for to establish the inference; (2) precisely how and in what context the inference will be applied; (3) which claims or defenses the inference will be limited to; and most importantly (4) the procedural mechanisms and related evidence that can be used in rebuttal or to otherwise respond to the inference. During a meet and confer telephone call on June 20, 2008, counsel for Oracle suggested the use of sampling and inferences. In response, counsel for Defendants stated that Defendants would need to understand the precise proposal for sampling before deciding whether to agree and further asked whether Oracle had such a proposal. In response, Oracle's counsel suggested that sampling might be useful in connection with discovery of downloads of Oracle software and support materials and the use of customer software "environments" by Defendants. Oracle did 20
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

not, however, offer any specifics or a concrete proposal to which Defendants could respond. Moreover, Defendants noted that these "downloads" and "environments" have already been produced or will be produced as part of the Data WarehouseAgreement to Oracle in electronic form, so it is unclear what discovery burdens would be eliminated by sampling them. (5) Resolution of Damages Causation Evidence

Oracle correctly understands that Defendants will challenge its proof of causation of damages. Oracle would like to make sweeping assertions that every customer that ever left Oracle for TomorrowNow represents legal damages to Oracle. Irrespective of liability, there is ample evidence that Oracle would have lost customers regardless of TomorrowNow's activities and thus Defendants did not cause the damages. Under the Copyright Act, actual damages represent the injury to the market value of the copyrighted work at the time of infringement. 4 Nimmer on Copyright § 14.02(a) at 14-13 to 1414. In appropriate circumstances, this amount is computed by determining what profits would have accrued to plaintiff but for the infringement. Nimmer §14.02(a)(1) at 14-14. Therefore, a plaintiff bears the burden of proving a causal connection between the infringement and actual damages, a requirement which is "akin to tort principles of causation and damages." Polar Bear Productions, Inc. v. Timex Corp., 384 F.3d 700, 708 (9th Cir. 2004). Thus, evidence that TomorrowNow's customers could have or would have left Oracle with or without TomorrowNow's activities presents a defense and discovery must be permitted into that area. Accordingly, both sides are entitled to discovery into the reasons customers dropped Oracle maintenance and support in favor of TomorrowNow. In addition, Oracle has argued in recent meet and confer discussions that it is also entitled to discovery of SAP's communications with customers who entered into maintenance and support agreements with TomorrowNow and simultaneously (or subsequently) replaced their Oracle enterprise software products with the competing SAP products. Apparently Oracle contends that such sales of SAP products to TomorrowNow customers was somehow caused by the alleged copyright infringement by TomorrowNow.

21

JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Oracle incorrectly implies that it may be deprived of discovery relating to damages causation. However, Defendants have already produced a large volume of TomorrowNow sales custodian data, including that of its lead sales person. In addition, Defendants are in the process of collecting, and will produce, the "comparable information" Oracle seeks, to the extent that it exists. Oracle has agreed that targeted searches is the appropriate approach to discovery this subject and it has conducted its own targeted searches for Oracle documents relating to the reasons customers left Oracle. Similarly, SAP is willing to conduct targeted searches for documents reflecting the reasons that a TomorrowNow customer replaced Oracle enterprise software with a competing SAP product. Such searches will likely include reviews of centrally maintained customer analysis information concerning reasons for the change as well as limited targeted searches of the primary account executives contacts with or about the customers' decisions. Defendants are also willing to permit discovery into SAP's communications with TomorrowNow customers who subsequently became SAP customers provided that: (1) it is reasonably narrow and tailored to its purpose; (2) Defendants get comparable discovery from Oracle of Oracle's communications with the same customers. The problem here again is primarily one of burden. For example, SAP has approximately 500 account executives in the United States alone and many more throughout the world. In addition, it is likely with respect to any given sale of enterprise software that numerous SAP employees would have had some access to documents that may reflect communications concerning customers' decisions to purchase SAP products. Accordingly, it would be impractical to search the documents of all custodians who might potentially have responsive documents. c. Joint Statement re Data Warehouse Agreement

A large volume of documents requested by Oracle, such as copies of software environments and downloaded support materials, are maintained by TomorrowNow on several servers and computers in Texas. After Judge Legge instructed Defendants to permit an Oracle engineer to inspect their systems on-site, and because of the cost associated with producing the necessary copies of all this material and in the interest of having a complete and timely 22
JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

production of the material on these servers, the Parties are negotiating the Data Warehouse Agreement. Rather than require complete production to Oracle of copies of all of these documents, the Parties' Data Warehouse Agreement allows Oracle to review the documents remotely online and identify those for which it wishes to receive a copy. Defendants will then review that subset to determine whether they will agree to produce those documents. Defendants provided instructions for the Data Warehouse on June 18, 2008 and the Parties expect the online review process to begin in mid-July. Oracle has reserved its right to seek a full forensic production of this electronic data, pending successful completion of the remote access review and production process, including the use of extrapolation or sampling methods where appropriate or necessary. Once the process is underway, Oracle will evaluate to see if this solution is adequate to the needs of the case. 2. Likely Upcoming Discovery Motions a. Oracle's Motions

Motions re Rule 30(b)(6) Depositions: Defendants' preparation of their designated corporate witnesses was inexcusably deficient on the topics of (1) SAP's acquisition of SAP TN, (2) its integration of SAP TN, and (3) the Safe Passage marketing and sales program SAP implemented using SAP TN as a means to lure Oracle service customers to become SAP TN service customers and ultimately SAP applications customers. Oracle has objected on the record regarding these deficiencies, asked for further deposition of properly prepared witnesses, and sought Defendants' agreement to reduce from Oracle's total allotted deposition hours the time wasted in these depositions. If agreement with Defendants is not reached, motion practice will be required to rectify Defendants' inadequate preparation. Defendants have recently served Rule 30(b)(6) notices on Oracle that cover topics irrelevant to this matter (e.g., all aspects of Oracle's historic and ongoing competition against SAP, Oracle's competition with other third party servicers, details on every aspect of Oracle's records, all group email accounts relating to SAP). Absent withdrawal of these topics, motion practice for a protective order or to quash the notices will be required.

23

JOINT DISC. CONF. STATEMENT Case No. 07-CV-1658 PJH (EDL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Motions re Privilege Claims: Defendants have sought to "claw back" almost 100 documents, claiming they are privileged and were inadvertently produced. Oracle believes the documents are not privileged, and that even if they were, the privilege has been waived and the production was not inadvertent. The Parties have met and conferred regarding the clawed-back documents and have a further meet and confer scheduled for tomorrow. Should the Parties be unable to reach a successful resolution, per the Court's invitation at the last discovery conference, Oracle will seek guidance from the Court at the July 1, 2008 Discovery Conference on the timing and procedure for this motion, including how to present the Court with a limited set of these documents for in camera review to demonstrate the impropriety of the claimed privilege or of the claimed inadvertence. Oracle also may need Court intervention to obtain more information than is currently contained in Defendants' privilege logs. Extension of Discovery Timing Parameters: Because discovery has revealed SAP TN engaged in improper infringement prior to the current January 2004 discovery start date, Oracle has asked for liability and targeted damages discovery from SAP TN from its inception. Defendants have agreed to produce responsive documents from all SAP TN employees who were employed prior to January 1, 2004, but have demanded the same scope of discovery from Oracle beginning in 2002, without limitation. The Parties are meeting and conferring on this. Oracle also has asked Defendants to update their damages and liability discovery to the present. Defendants have agreed to produce some of these materials, have declined to produce others, and in all instances have stated they are unable to complete such a production until close to the end of 2008. If agreement cannot be reached, motion practice will follow. De-designation of Documents Marked HC and Confidential: Oracle continues to be frustrated in its efforts to work with its own personnel by Defendants' continued irresponsible designation of deposition testimony and production of materials as Highly Confidential or Confidential. See, e.g., fn. 7 above. Oracle's counsel has sent numerous meet and confer letters on this to Defendan