Disney Enterprises, Inc. et al v. Hotfile Corp. et al

Filing 301

MOTION for Summary Judgment >WARNER BROS. ENTERTAINMENT INC.'S MOTION FOR SUMMARY JUDGMENT AND MEMORANDUM OF LAW IN SUPPORT OF MOTION (PUBLIC REDACTED VERSION)< by Warner Bros. Entertainment Inc.. Responses due by 3/15/2012 (Attachments: # 1 Affidavit Declaration of Scott A. Zebrak in Support of Warner's Motion for Summary Judgment (public redacted version), # 2 Exhibit A to Declaration of S. Zebrak, # 3 Exhibit B to Declaration of S. Zebrak, # 4 Affidavit Declaration of Dr. Ian Foster in Support of Warner's Motion for Summary Judgment (public redacted version), # 5 Exhibit A to Declaration of I. Foster, # 6 Affidavit Declaration of david Kaplan in Support of Warner's Motion for Summary Judgment (public redacted version), # 7 Affidavit Declaration of Kerry Hopkins in Support of Warner's Motion for Summary Judgment (public redacted version), # 8 Exhibit A to Declaration of K. Hopkins, # 9 Affidavit Declaration of Jennifer Yeh in Support of Warner's Motion for Summary Judgment (public redacted version), # 10 Exhibit A to Declaration of J. Yeh, # 11 Exhibit B to Declaration of J. Yeh, # 12 Exhibit C to Declaration of J. Yeh, # 13 Exhibit D to Declaration of J. Yeh, # 14 Exhibit E to Declaration of J. Yeh, # 15 Exhibit F to Declaration of J. Yeh, # 16 Exhibit G to Declaration of J. Yeh, # 17 Exhibit H to Declaration of J. Yeh, # 18 Exhibit I to Declaration of J. Yeh, # 19 Exhibit J to Declaration of J. Yeh, # 20 Exhibit K to Declaration of J. Yeh, # 21 Exhibit L to Declaration of J. yeh, # 22 Exhibit M to Declaraton of J. Yeh, # 23 Exhibit N to Declaration of J. Yeh, # 24 Exhibit O to Declaration of J. Yeh, # 25 Exhibit P to declaration of J. Yeh, # 26 Exhibit Q to Declaration of J. Yeh, # 27 Text of Proposed Order)(Stetson, Karen)

UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF FLORIDA
CASE NO. 11-20427-WILLIAMS/TURNOFF

DISNEY ENTERPRISES, INC.,
TWENTIETH CENTURY FOX FILM CORPORATION,
UNIVERSAL CITY STUDIOS PRODUCTIONS LLLP,
COLUMBIA PICTURES INDUSTRIES, INC., and
WARNER BROS. ENTERTAINMENT INC.,
Plaintiffs,
v.
HOTFILE CORP., ANTON TITOV, and
DOES 1-10.
Defendants.
/
HOTFILE CORP.,
Counterclaimant,
v.
WARNER BROS. ENTERTAINMENT INC.,
Counterdefendant.
/

DECLARATION OF DR. IAN FOSTER IN SUPPORT OF WARNER’S MOTION FOR
SUMMARY JUDGMENT

PUBLIC REDACTED VERSION
1.

My name is Ian Foster and I currently hold the position of Director of the

Computation Institute at Argonne National Laboratory and the University of Chicago. I also
hold the positions of Arthur Holly Compton Distinguished Service Professor of Computer
Science at the University of Chicago, and of Argonne Distinguished Fellow at the Argonne

National Laboratory. My professional and academic background include extensive experience in
designing, operating, maintaining, and studying large, distributed computer systems, including
the management of large databases required for such systems to operate. A copy of my
curriculum vitae, including lists of my previous publications, is attached hereto as Exhibit A. I
have not given testimony in any legal proceeding the past four years.
2.

I have been asked by Warner, for purposes of Hotfile’s counterclaim, to extract

and describe certain data from data sets maintained and produced by www.hotfile.com
(“Hotfile”). The observations and conclusions set forth below are based on my own observation
and use of Hotfile’s produced data, as informed by my specialized knowledge, education, and
expertise as applied to the facts and circumstances in this case. In addition, I have for purposes
of this declaration consulted and reviewed the following data sets and documents produced by
Hotfile (including later-supplemented versions):
a. Actiondat.csv;
b. Affpay.csv;
c. Uploads.csv;
d. Userdat.csv;
e. Dmcanotices.csv;
f. Users_cowner_upload.csv;
g. Strikes.csv;
h. File.csv;
i. Uploadsurl.csv;
j. HF0285583;
k. Takedown notices sent to Hotfile produced by Hotfile in this litigation;
l. Transcripts and selected exhibits from of the November 16, 2011 and December
5-8 depositions of Anton Titov;
m. Source code related to the removal and blocking of files produced by Hotfile.
3.

I am being compensated for my study and testimony in this case at a rate of $500

per hour. If called as a witness at trial, I would testify as to the contents of this declaration.
Methodology
4.

Hotfile provided various data sets in this case, which I understand were extracted

from its database(s), in comma separated value (or “CSV”) format, a common format used to

2

represent tabular data. Analyzing and querying data sets in this format is a straightforward
proposition: they were loaded to a database server (using MySQL, a common database
management system); I then constructed and indexed tables (which associate subsets of the data
based on defined criteria) using the MySQL software, and then ran queries using common
database commands as discussed below.
Number of Takedown Notices Received by Hotfile
5.

Included in Hotfile’s data production were various data sets and other sources of

information reflecting takedown notices Hotfile has received from copyright owners. I have
been asked by Warner’s counsel to provide the number of such notices received. Hotfile has
provided data reflecting notices referring to a cumulative total of approximately ten million files
(and more than a million after the filing of this lawsuit on February 8, 2011). The methodology
for calculating this number is as follows.
6.

First, Hotfile provided a data set “dmcanotices.csv” (as well as subsequent

supplements) containing entries corresponding to takedown notices that Hotfile has received by
email. These can be summed using common database commands. Second, Hotfile produced
another data set – “users_cowner_upload.csv” – containing records of takedown notices sent by
means of Hotfile’s web-based interface or special rightsholder account (“SRA”), which can
likewise be summed using common database commands. In theory, these two sources ought to
be sufficient to capture takedown notices received by Hotfile. In analyzing the dmcanotices.csv
data set, however, I observed that the data set only contains takedown notices from July 2, 2009
through September 15, 2009, and then again from September 19, 2010 through the present. In
other words, a year of data (running from September 2009 through September 2010) is missing
from the dmcanotices.csv data set entirely. I understand that Hotfile has stated that this data was
lost and cannot be recovered.
7.

Although this data is missing from the dmcanotices.csv data set, it can be

substantially reconstructed from the actual email takedown notices that Hotfile received and
produced, many of which were provided in text form. Using a script (i.e., a small computer
program), I extracted the “uploadid” (i.e., the unique numerical identifier that Hotfile assigns to
each file uploaded to its service) of the files contained in the accessible takedown notices for the

3

period for which Hotfile’s dmcanotices.csv data set was missing entries entirely, then summed
those and added them to the total count from the other two sources.1
Number of Users Suspended by Hotfile for Repeat Infringement Subsequent to Litigation
8.

I have been asked by Warner’s counsel to provide a count, from Hotfile’s data, of

the total number of users Hotfile suspended based on alleged copyright infringement subsequent
to the filing of Plaintiffs’ lawsuit against Hotfile.
9.

Hotfile maintains records of its suspension of users based on alleged copyright

infringement. One of Hotfile’s produced data sets, “actiondat.csv” (and supplements to that data
set) contains records of actions Hotfile took regarding users. The actiondat data set associates
the unique identification number that Hotfile assigns to each user (“userid”) with the “type” of
action (which includes “suspenduser”), the date of the action (“dtaction”) and other information
about the suspension, including the reason for the suspension (“params”). Therefore, this data
set readily identifies users whom Hotfile terminated based on alleged copyright infringement
subsequent to the filing of the Complaint on February 8, 2011. Such suspensions – which began
on February 18, 2011 – exceed 22,000 as of the latest supplement to Hotfile’s actiondat data set.
Data Regarding Counterclaim Files and Their Uploading Users
10.

I understand that Plaintiffs’ expert Scott Zebrak is submitting a declaration

containing information concerning each of the 890 files that Hotfile is claiming that Warner
wrongfully took down. At Warner’s counsel’s request, I have extracted and provided certain
data from Hotfile’s produced data sets, for use in connection with Mr. Zebrak’s report, that relate
to each of the files identified by Hotfile in the Counterclaim. This includes, for each file:
a. The unique identifier (“userid”) that Hotfile assigns to the uploading user, as well
as the name of the user’s account. Both values are straightforward associations of
data contained in Hotfile’s userdat and uploads data sets (which contain various
data concerning users and concerning uploaded files, respectively).
b. Whether or not a file identical to the file identified in the Counterclaim (i.e., a file
with the same “hash” value)2 had been the subject of a takedown notice sent by a
1

I note here that my methodology was conservative in that it was limited to takedown notices
accessible in text form. Hotfile produced a much smaller number of notices in other formats that
I did not consider due to the additional difficulty and higher potential for introducing inadvertent
errors into the count. Therefore, the actual number of takedown notices received by Hotfile is
likely somewhat higher.
4

copyright owner other than Warner. This can be determined by identifying the
hash values of the files in the Counterclaim (which are represented as “md5”and
“sha1” in the uploads and file data sets), then comparing those hash values to the
hash values of files on which Hotfile received a takedown notice (excluding
Warner’s own takedowns through the SRA), and identifying any matches.3
c. Search results reflecting the names of files with similar names that had been
removed from Hotfile for reasons related to claimed infringement.
i. This can be determined by creating a table of the files removed from
Hotfile due to infringement claims. This includes files on which Hotfile
received a takedown notice (the process for identifying which I described
in Paragraphs 5-7 above) as well as files that Hotfile itself deleted (which
includes files removed in response to a specific takedown notice, but also
includes, e.g., files Hotfile removed due to takedown notices that removed
multiple files at once, files Hotfile removed or blocked due to a hash
match with a different file removed due to claimed infringement, etc.).
ii. This latter category can be shown by the file’s “status” in Hotfile’s
uploads data set; Hotfile (per the testimony of Anton Titov and associated
deposition exhibits) uses various different codes to refer to different file
statuses, with status “5” and “2” reflecting Hotfile’s own suspension and
deletion of files (as opposed to files deleted for other specific reasons, e.g.,
deleted by the user or deleted due to inactivity).
iii. Once this table was constructed, I excluded files to avoid counting files as
that had been removed for reasons other than claimed infringement.
1.

First, I excluded Warner’s own takedowns through the SRA.

2. Second, as a conservative measure, unless there was a takedown
notice for the specific file, I also excluded files where the user had
been suspended, for any reason. This restriction is very
conservative; files removed for infringement were many cases
uploaded by users who were later suspended. Nevertheless, I
understand from Mr. Titov’s testimony, as well as from my
analysis of the data, that Hotfile appears to have for certain time
periods and in certain cases used statuses “5”and “2” to designate
files suspended or deleted because the user was suspended, even in
the absence of alleged infringement with respect to the specific file
itself. Therefore, to err on the side of underinclusion rather than
overinclusion, I did not include such files in the table.
3. Third, I excluded files where Hotfile had chosen not to “block” the
file from future upload (i.e., where the “blocked” value for the file
2

A “hash” does not purport to recognize when two different files contain the same content (since
the representation of that content in a digital file can differ). However, it does usefully identify
when two digital files are exact copies of one another.

3

MD5 and SHA1 are both common algorithms for identifying a file’s hash used by Hotfile.
5

in Hotfile’s uploads data set was set to a value of “0”). Again, this
is a conservative measure, as Hotfile may have declined to block
uploads of files removed for reasons related to infringement in
some cases. Nevertheless, to err on the side of underinclusion
rather than overinclusion, I did not include files where Hotfile had
itself chosen not to block future uploads of the file.
iv. Once this table was constructed, the identification of file name matches is
a simple matter of using MySQL’s search functionality to identify
candidate matches. Here, I used search terms for the 890 counterclaim
files and ran them against the files in the table constructed and narrowed
as described above. I then provided up to 25 candidate matches for each
file to Plaintiffs for analysis.
d. Search results reflecting the names of copyright owners, with names similar to the
copyright owners that Plaintiffs had identified as being the owners of the files in
question, that had sent takedown notices to Hotfile. I conducted this search by
using the search functionality of MySQL to search both the bodies of the
takedown notices reflected in the dmcanotices data set and the usernames and
email addresses of copyright owners in the users_cowner_upload data set. I then
provided up to 25 candidate matches for each file to Plaintiffs for analysis.
11.

In addition, also from Hotfile’s data sets, I have extracted at counsel’s request the

following data concerning the uploading user of each file:
a. How many copyright “strikes” Hotfile had itself assigned to the user based on its
own count of takedown notices sent against the user’s files (a straightforward
association of the userid with Hotfile’s “strikes” data set).
b. The number of discrete days on which Hotfile received a takedown notice for a
file uploaded by the user (which I have termed “Notice Days”). This can be
calculated by associating takedown notices with the user, then in turn summing
the distinct number of days on which such notices are dated.
c. Whether or not the user was suspended for copyright infringement, and if so, the
date of the first such suspension. As described in paragraph 9 above, Hotfile
maintains records of suspension events, dates, and reasons in its actiondat data
set. It also maintains records (in the userdat data set) of whether or not a user is
suspended. In some cases, Hotfile’s userdat data reflects the suspension of a user
but the actiondat data set did not contain an entry reflecting the reason. In those
cases, I have indicated “Reason Unknown” (“RU”).
d. The number of days, if any, that elapsed between the user’s suspension and the
date of the Warner notice at issue (calculated by means of a simple script
reflecting the difference between the date of the notice and the date of the
suspension).
e. The number of discrete days after February 18, 2011 on which the takedown
notices over which Hotfile is suing in its counterclaim were sent against a file

6

uploaded by the user. (This is a simple sorting and/or manual calculation within
Microsoft Excel).
f. Whether or not the user is a “Premium” user, and, if so, until when. This data is
reflected in the userdat data set.
g. Whether Hotfile has paid the user as an “Affiliate” and, if so, the sum of such
affiliate payments. This data is reflected in Hotfile’s userdat and affpay data sets.
h. Whether Hotfile has a copy of the file. Hotfile maintains data in the uploads data
set showing the size of each file; whether or not Hotfile has such data can act as a
proxy for whether Hotfile has a copy of the file itself.

I declare under penalty of perjury that the foregoing is true and correct.
Executed in the State of Illinois this 10th day of February, 2012.

___________________________________
Ian Foster, Ph.D.

7