music file stored on a third-party Internet server and transfer the file to the user’s locker. MP3
tunes used the standard Content-Addressable Storage system to store music files on its servers.
Based on the sequence of data bits in a particular music file, the algorithm created a hash tag as
an identification number. If different users uploaded the same song containing identical blocks
of data to MP3tunes’ servers, those blocks were assigned the same hash tag and typically saved
only once. If a user played or downloaded a song from a locker, the storage system used the
hash tags associated with the uploaded song to reconstruct the exact file the user originally
uploaded to his or her locker.^2731
MP3tunes also operated a second web site at http://www.sideload.com, which allowed users to
search for free song files. Upon entering a keyword, the site would generate a list of potential
matches by searching an index of web sites with free song files and cross-referencing the
keywords with information associated with each song file. The index was maintained on
MP3tunes’ servers. After clicking on a search return, the user was taken to a page where he or
she could play the song, follow a link to the third-party web site hosting the song, or download
the song to another computer. If the user had a locker on MP3tunes.com, Sideload.com
displayed a link that if clicked would “sideload” (download) the song from the third-party web
site and save it to his or her locker. Importantly, MP3tunes kept track of the sources of songs in
its users’ lockers. Thus, MP3tunes could identify the third-party web sites from which users
copied songs to their lockers. Sideload servers automatically generated lists of “Most Popular,”
“Featured,” and “New” songs that users could browse.^2732
Sideload.com also offered users free Sideload plug-in software. When a user of the plug-
in came across a web site with a free song file, a button appeared on that site that would copy the
song directly to the user’s MP3tunes locker without visiting Sideload.com. When a user
sideloaded a song from a third-party site, either through the Sideload plug-in or the Webload
software, that third-party site was added to Sideload.com’s index of searchable songs.
Information associated with the song, such as artist, album, title, and track was automatically
stored on a “Track Details” page, and the information became part of the searchable index. From
then on, Sideload.com returned a potential match whenever any other user searched for that song
on Sideload.com by entering keywords that matched the song file. When a downloadable song
was removed from a third-party source, the Sideload feature became inoperable and users could
no longer add the song to their lockers. However, users who sideloaded the song before it was
removed from the third-party source could continue to access the song through their lockers.^2733
The plaintiff record labels sought to hold MP3tunes liable for copies of infringing music
files stored in their lockers and MP3tunes asserted the Section 512(c) safe harbor. One of the
plaintiffs, EMI, had submitted a notice of infringement identifying 350 song titles and web
addresses that allegedly infringed EMI’s copyrights. It also provided a list of EMI artists and
demanded that MP3tunes remove all of EMI’s copyrighted works, even those not specifically
identified with web addresses. MP3tunes removed links to the specific web addresses but did
(^2731) Id. at 633-34.
(^2732) Id. at 634.
(^2733) Id. at 634-35.