Oracle America, Inc. v. Google Inc.

Filing 493

MOTION in Limine NO. 2 filed by Google Inc.. Responses due by 10/21/2011. Replies due by 10/28/2011. (Attachments: #1 Oracle Opposition)(Kamber, Matthias) (Filed on 10/7/2011)

1
2
3
4
5
6
7
8
9
10
11
12
13
14

KEKER & VAN NEST LLP
ROBERT A. VAN NEST - #84065
[e-mail]
CHRISTA M. ANDERSON - #184325
[e-mail]
633 Battery Street
San Francisco, CA 94111-1809
Telephone:
415.391.5400
Facsimile:
415.397.7188

KING & SPALDING LLP
DONALD F. ZIMMER, JR. - #112279
[e-mail]
CHERYL A. SABNIS - #224323
[e-mail]
101 Second St., Suite 2300
San Francisco, CA 94105
Tel: [phone]
Fax: [phone]

KING & SPALDING LLP
SCOTT T. WEINGAERTNER (Pro Hac Vice)
[e-mail]
ROBERT F. PERRY
[e-mail]
BRUCE W. BABER (Pro Hac Vice)
1185 Avenue of the Americas
New York, NY 10036
Tel: [phone]
Fax: [phone]

IAN C. BALLON - #141819
[e-mail]
HEATHER MEEKER - #172148
[e-mail]
GREENBERG TRAURIG, LLP
1900 University Avenue
East Palo Alto, CA 94303
Tel: [phone]
Fax: [phone]

Attorneys for Defendant
GOOGLE INC.

15

UNITED STATES DISTRICT COURT

16

NORTHERN DISTRICT OF CALIFORNIA

17

SAN FRANCISCO DIVISION

18
19

Case No. 3:10-cv-03561-WHA

ORACLE AMERICA, INC.,

20

Plaintiff,

21

v.

22

GOOGLE INC.,

23
24

Defendant.

GOOGLE’S MOTION IN LIMINE NO. 2
TO EXCLUDE ALLEGED
PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED
FUNCTIONALITY
Judge:

Hon. William Alsup

Date Comp. Filed:

October 27, 2010

Trial Date:

October 31, 2011

25
26
27
28

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

MOTION AND RELIEF REQUESTED

2

Under Federal Rules of Evidence 401, 402, 403 and 702, defendant Google Inc.

3

(“Google”) hereby moves the Court for an order excluding in limine all evidence and argument

4

concerning the alleged performance benefits of Android with the accused functionality.

5

DISCUSSION

6

Oracle submitted four short reports by its own employees, Robert Vandette, Noel Poore,

7

Erez Landau and Seeon Birger (“the performance reports”), that collectively purport to quantify

8

the performance of Android with and without the alleged inventions claimed in five of Oracle’s

9

asserted patents. Based on these reports, Oracle experts Dr. Mitchell (on patent infringement),

10

Dr. Goldberg (on patent validity) and Dr. Cockburn (on damages) conclude that alleged

11

infringement of the asserted patents results in performance gains on Android devices, and by

12

extension, is the basis for consumer demand and Android’s success.
First, the Oracle employees who submitted the performance reports conceded that they

13
14

knew little about the Oracle patents, the asserted claims or the accused functionality. Moreover,

15

the employees lack any semblance of objectivity and failed to comply with the basic tenets of

16

Daubert and Rule 702. Second, the opinions are irrelevant to any issue of consequence in the

17

trial’s initial liability phase—and thus are inadmissible under Rule 402 except in the damages

18

phase, at most. Third, the opinions should be excluded under Rule 403 even for the damages

19

phase of trial because the limited probative value of the performance reports is substantially

20

outweighed by the danger of misleading the jury into thinking that actual Android devices

21

running actual Android applications would experience a similar performance impact.

22

1.

The Performance Reports Fail to Comply with Daubert and Rule 702

23

The performance reports attempt to demonstrate the benefits of implementing five of the

24

asserted patents in the Android operating system. (The ‘447 and ‘476 patents are not addressed

25

in the reports). To accomplish this, Oracle ran various “performance benchmark” tests against

26

unmodified Android code to generate various performance statistics (e.g., how fast applications

27

are executed or how much storage space they require in memory). Oracle then modified the

28

Android code to try disabling the accused functionality, and re-ran the same tests in order to
1

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

compare the two sets of results and point to any drop in performance as indicative of the benefits

2

provided by the patented techniques. The idea may seem simple in theory, but it is a complicated

3

task in practice, and Oracle’s employees took countless shortcuts, disregarded approved

4

methodologies, and ran particular tests that were assured (and evidently designed) to reach the

5

conclusions sought by Oracle. Most critically, they made subpar modifications to the Android

6

code that were crude, hasty and in many instances crippled substantial non-accused

7

functionalities as well. The performance reports are so fraught with errors, assumptions and an

8

overall lack of detail that, had the reports been submitted for peer review to any reputable

9

technical journal or academic institution, they would be immediately rejected. See Ellis v.

10

Costco Wholesale Corp., No. 07-15838, 2011 U.S. App. LEXIS 19060 at *25-26 (9th Cir. Sept.

11

16, 2011) (“Under Daubert, the trial court must act as a ‘gatekeeper’ to exclude junk science that

12

does not meet Federal Rule of Evidence 702’s reliability standards by making a preliminary

13

determination that the expert’s testimony is reliable.” (citing Kumho Tire Co. v. Carmichael, 526

14

U.S. 137, 145 (1999).)

15

The essential prerequisite to any expert testimony is that it rest “on a reliable foundation.”

16

Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 584-587 (1993). As codified in Rule 702:

17

If scientific, technical, or other specialized knowledge will assist
the trier of fact to understand the evidence or to determine a fact in
issue, a witness qualified as an expert by knowledge, skill,
experience, training, or education, may testify thereto in the form
of an opinion or otherwise, if (1) the testimony is based upon
sufficient facts or data, (2) the testimony is the product of reliable
principles and methods, and (3) the witness has applied the
principles and methods reliably to the facts of the case.

18
19
20
21
22

A. It is Unclear What Was Done, Why, and By Whom

23

Dr. Mitchell has no firsthand knowledge about any code modifications, test conditions, or

24

the actual tests. He participated in some phone calls, after which he believed the employees used

25

“modifications that [he] suggested” and that “the actual experiments that were carried out

26

reflected [his] suggestions and requests on how to do it.” Mitchell Dep. Tr. (attached as Ex. 2 to

27

the Declaration of Daniel Purcell in Support of Google Inc.’s Motions in Limine (“Purcell

28

Decl.”) filed herewith) at 81:16 – 82:6; see also Purcell Decl. Ex. 3 (Mitchell Patent Report) at ¶
2

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

196. Yet, the Oracle engineers attribute everything they did to Dr. Mitchell. Robert Vandette

2

submitted a report on the ‘104 and ‘205 patents and admitted that he “relied on Dr. Mitchell to

3

interpret [the patent].” Purcell Decl. Ex. 4 (Vandette Dep. Tr.) at 74:16-17. When asked if he

4

had the “knowledge necessary to determine what functionality should be disabled,” he responded

5

that he was “not a patent attorney” and did not “know exactly which claim in the [‘104] patent

6

corresponds to [the] functionality” he disabled. Id. at 76:23-24; 77:24-78:1. Likewise, Noel

7

Poore submitted a report regarding the ‘520 and ‘702 patents and testified that he “looked

8

through the patent[s]” but had no idea which claims were asserted, and was “not necessarily

9

familiar enough with the detail of th[e] patent[s]” to know how they related to the accused

10

functionality. Purcell Decl. Ex. 5 (Poore Dep. Tr.) at 33-35; 74-75. Erez Landau revealed that

11

he reviewed only the Abstract of the ‘720 patent and “part of the claims.” Purcell Decl. Ex. 6

12

(Landau Dep. Tr.) at 21-22.

13

In addition, Vandette, Poore and Landau all indicate in their August 9, 2011 reports that

14

they downloaded the Android code, modified it, compiled it and tested it. Purcell Decl. Ex. 7

15

(Vandette report) at ¶ 27-39; Purcell Decl. Ex. 8 (Poore report) at ¶ 12-14; Purcell Decl. Ex. 9

16

(Landau report) at ¶ 25-31. Now, in a report Oracle submitted on September 12, 2011 (after the

17

close of expert discovery), a fourth employee named Seeon Birger claims that he was the one

18

who downloaded the Android code, modified it, compiled it, and provided it to the other Oracle

19

employees. See Purcell Decl. Ex. 10 (Birger report) at ¶ 11-21. Landau admitted as much

20

during his deposition (see Purcell Decl. Ex. 6 (Landau Dep. Tr.) at 58-63), while Vandette and

21

Poore appear to have concealed the fact that someone else performed the work disclosed in their

22

reports. See, e.g., Purcell Decl. Ex. 5, (Poore Dep. Tr.) at 25:9-11 (“Q. So you did all the work

23

in this report by yourself, correct? A. That’s correct”). In short, Oracle failed to provide the

24

foundational report underlying the other performance reports until weeks later, without

25

explanation, and prejudicing Google’s discovery into the benchmarking analyses. Even more

26

appalling, Oracle refuses to present Mr. Birger for deposition. See Purcell Decl. Ex. 11

27

(September 21, 2011 email from M. Francis to M. Peters.) For these reasons, all of the

28

benchmarking analyses are legitimately in question and should be excluded for Oracle’s failure
3

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

to comply with its expert-discovery obligations.
B. Oracle Made Improper Modifications to the Android Code

2

Oracle’s modifications of Android code improperly disabled or corrupted functionality in

3
4

Android that is not accused of infringement. For example, Vandette disabled the JIT compiler

5

for his ‘104 patent tests, even though the JIT compiler was never accused in connection with that

6

patent. Purcell Decl. Ex. 4 (Vandette Dep. Tr.) at 84:18-25. With respect to the ‘205 patent,

7

which is asserted against a single feature of the JIT compiler, Vandette again disabled the whole

8

JIT compiler, even though the ‘205 patent itself recognizes the existence of JIT compilers in the

9

prior art. For tests relating to the ‘702 patent, Poore modified Android code so existing

10

duplicates would remain in the Android “.dex” files—but his report openly admits that the

11

modified code created new duplicates that did not previously exist. Purcell Decl. Ex. 8 (Poore

12

Report) at ¶ 23. Landau admitted that when disabling the copy-on-write functionality allegedly

13

claimed by the ‘720 patent, his test program loaded 1,264 files into memory, but readily agreed

14

that “if copy-on-write is disabled, it would just be a waste of memory to load these [files] …”

15

and it would make no sense to have Android run that way. Purcell Decl. Ex. 6 (Landau Dep. Tr.)

16

at 86:1-87:24. As a whole, the performance reports fail to account for changes beyond the

17

accused functionality responsible for differences in performance between the original and

18

modified code. See Claar v. Burlington N.R.R., 29 F.3d 499 (9th Cir. 1994) (testimony excluded

19

where the expert failed to consider other obvious causes for the plaintiff’s condition). Oracle’s

20

experts offer no recognized methodology guiding their modifications of the Android source

21

code. Under Daubert, the Court needs to exclude expert testimony unless the “methodology

22

underlying the testimony is scientifically valid.” 509 U.S. at 592-93.
C. Oracle’s Reports Draw Conclusions Not Supported by Its Tests

23

Even assuming that Oracle’s performance reports quantify the difference in running

24
25

Android code with and without the accused functionality, Oracle uses the test results to make

26

unfounded conclusions that those results are indicative of the patents’ impact on the real-world

27

performance of Android, as follows:

28



Vandette and Poore did not test the performance of any actual Android applications;
4

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1



Vandette and Poore did not test the performance of any actual Android devices;

2



Landau tested 17 Android applications at one time on the Nexus One device released nearly

3

two years ago, and—as mentioned above—he modified the code to operate in a manner that

4

would never be implemented by Google and made no practical sense. See Purcell Decl. Ex.

5

12 (Landau Dep. Ex. 494) (Android’s website specifically states that the Nexus One should

6

not be used for testing);

7



simply crippled the code in a manner that would never actually be implemented;

8
9

None of the employees attempted to implement reasonable non-infringing substitutes—they



None of the tests measure overall system performance on an Android device and they fail to

10

recognize that the Dalvik VM is used only 1/3 of the time by Android—the other 2/3 of time

11

Android is executing native machine code. See Purcell Decl. Ex. 13 (Vandette Dep. Ex. 462,

12

May 2010 Google IO presentation) at slide 5;

13



source code and not any particular code release that has been used by actual devices;

14
15

Oracle’s employees downloaded and tested a development version of Android’s “Froyo”



There is nothing in the employee reports that indicates whether they followed the specific

16

instructions on Android’s website for downloading, compiling and installing Android on

17

devices. (For example, Landau testified that Mr. Birger used version 6 of the JDK software

18

to compile Android “Froyo” code, whereas the Android website specifically requires that

19

version 5 of the JDK be used. See Purcell Decl. Ex. 6 (Landau Dep. Tr.) at 61:17-21; Purcell

20

Decl. Ex. 14 (Landau Dep. Ex. 491)
Taking these factors into account, “there is simply too great an analytical gap between the

21
22

data and the opinion proffered.” General Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997). All of

23

Oracle’s expert conclusions based on these reports amount to just “subjective belief or

24

unsupported speculation” and should be excluded. Daubert, 509 U.S. at 588-90 (“[I]f an opinion

25

is fundamentally unsupported, then it offers no expert assistance to the jury.”).

26
27

2.

The Performance Reports Should Be Excluded From the Liability Phase of the Case
under Rule 402

28

Evidence regarding the purported performance impact of the asserted patents is
5

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

potentially probative of only damages theories, and all damages issues have been bifurcated.

2

Thus, the purported performance impact of the asserted patents has no bearing on alleged

3

infringement or invalidity and has no relevance to the liability phase of the trial.

4

As stated in Google’s motion regarding “commercial success,” claims about the

5

performance with and without the asserted patents pose at least a “modest likelihood” of unfair

6

prejudice and a more than a “small risk” of misleading the jury because it tends to create an

7

inflated sense of the patent’s importance to the success of Android. See United States v.

8

Gonzalez-Flores, 418 F.3d 1093, 1098 (9th Cir. 2005) (applying Rule 403 and holding that

9

“[w]here the evidence is of very slight (if any) probative value, it’s an abuse of discretion to

10

admit it if there’s even a modest likelihood of unfair prejudice or a small risk of misleading the

11

jury”) (citation omitted); see also United States v. Espinoza-Baza, 647 F.3d 1182, 1190 (9th Cir.

12

2011) (“evidence presenting even a ‘modest likelihood of unfair prejudice’ is ‘high enough to

13

outweigh the . . . probative value’ of marginally relevant evidence” under Rule 403) (citation

14

omitted).

15
16

3.

The Performance Reports Should Be Excluded From the Damages Phase of the
Case under Rule 403

17

As explained above, Oracle’s performance reports fail to analyze the performance impact

18

of the accused infringing features on real-world Android devices and applications. Furthermore,

19

the results are unreliable because they do not account for realistic non-infringing alternatives or

20

the detrimental impact of Oracle’s sloppy code modifications to non-infringing Android

21

functionality. In view of these limitations, Oracle’s reports offer little probative value, which

22

value is substantially outweighed by its capacity to be used in prejudicial ways. For example,

23

Oracle may contend that consumers’ Android products would suffer a noticeable performance

24

impact without the accused functionality. But this premise is unsupported by the performance

25

reports, which focus on micro-benchmarks of the virtual machine, not overall system

26

performance. Even if Oracle presents testimony within the scope of the reports, it threatens to

27

mislead the jury to conclude that actual Android performance will mimic Oracle’s test results,

28

given all the flaws discussed above. Either way, the Court should exclude evidence regarding
6

581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA

1

the performance reports. See Rule 403 (“evidence may be excluded if its probative value is

2

substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading

3

the jury…”); Daubert, 509 U.S. at 595 (“Expert evidence can be both powerful and quite

4

misleading because of the difficulty in evaluating it.” (citation omitted)).

5

As a final note, in order to safeguard the effectiveness of its ruling and thus avoid the

6

danger of prejudice, the Court also should bar counsel for both parties from making any remarks

7

in opening argument about the alleged performance benefits of the asserted patents in Android.
Respectfully submitted,

8
9

Dated: September 24, 2011

KEKER & VAN NEST LLP

10
11
12

By: s/ Robert A. Van Nest
ROBERT A. VAN NEST
Attorneys for Defendant
GOOGLE INC.

13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
7
581324.02

GOOGLE’S MOTION IN LIMINE NO. 2 TO EXCLUDE ALLEGED PERFORMANCE BENEFITS OF
ANDROID WITH THE ACCUSED FUNCTIONALITY – CASE NO. 3:10-cv-03561-WHA