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
`
`
`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 1 of 26
`
`Michael A. Sherman (SBN 94783)
`masherman@stubbsalderton.com
`Jeffrey F. Gersh (SBN 87124)
`jgersh@stubbsalderton.com
`Sandeep Seth (SBN 195914)
`sseth@stubbsalderton.com
`Wesley W. Monroe (SBN 149211)
`wmonroe@stubbsalderton.com
`Stanley H. Thompson, Jr. (SBN 198825)
`sthompson@stubbsalderton.com
`Viviana Boero Hedrick (SBN 239359)
`vhedrick@stubbsalderton.com
`STUBBS, ALDERTON & MARKILES, LLP
`15260 Ventura Blvd., 20th Floor
`Sherman Oaks, CA 91403
`Telephone:
`(818) 444-4500
`Facsimile:
`(818) 444-4520
`
`Attorneys for Plaintiffs
`[Additional Attorneys listed
`below]
`
`
`UNITED STATES DISTRICT COURT
`
`NORTHERN DISTRICT OF CALIFORNIA
`
`SAN JOSE DIVISION
`
`
`CASE NO. 5:18-md-02834-BLF
`
`
`
`FIRST AMENDED COMPLAINT
`
`DEMAND FOR JURY TRIAL
`
`
`Case No. 5:18-cv-00183-BLF
`
`
`
`IN RE: PERSONAL WEB TECHNOLOGIES,
`LLC ET AL., PATENT LITIGATION
`
`
`
`
`
`
`______________________________________
`PERSONALWEB TECHNOLOGIES, LLC, a
`Texas limited liability company, and
`LEVEL 3 COMMUNICATIONS, LLC,
`a Delaware limited liability company,
`
`
`
`v.
`
`SQUARE, INC., a Delaware corporation,
`
`
`
`
`Defendant.
`
`Plaintiffs,
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`
`
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`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
`
`
`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 2 of 26
`
`Plaintiff PersonalWeb Technologies, LLC (“Plaintiff” or “PersonalWeb”) files this First
`
`Amended Complaint (“Complaint”) for patent infringement against Defendant Square, Inc.
`
`(“Defendant”). Plaintiff PersonalWeb Technologies, LLC alleges:
`
`PRELIMINARY STATEMENT
`
`1.
`
`PersonalWeb and Level 3 Communications, LLC (“Level 3”) are parties to an
`
`agreement between Kinetech, Inc. and Digital Island, Inc. dated September 1, 2000 (the “Agreement”).
`
`Pursuant to the Agreement, PersonalWeb and Level 3 each own a fifty percent (50%) undivided
`
`interest in and to the patents at issue in this action: U.S. Patent Nos. 6,928,442, 7,802,310, 7,945,544,
`
`and 8,099,420 (“Patents-in-Suit”). Level 3 has joined in this Complaint pursuant to its contractual
`
`obligations under the Agreement, at the request of PersonalWeb.
`
`2.
`
`Pursuant to the Agreement, Level 3 has, among other rights, certain defined rights to
`
`use, practice, license, sublicense and enforce and/or litigate the Patents-in-Suit in connection with a
`
`particular field of use (“Level 3 Exclusive Field”). Pursuant to the Agreement PersonalWeb has,
`
`among other rights, certain defined rights to use, practice, license, sublicense, enforce and/or litigate
`
`the Patents-in-Suit in fields other than the Level 3 Exclusive Field (the “PersonalWeb Patent Field”).
`
`3.
`
`All
`
`infringement allegations, statements describing PersonalWeb, statements
`
`describing any Defendant (or any Defendant’s products) and any statements made regarding
`
`jurisdiction and venue are made by PersonalWeb alone, and not by Level 3. PersonalWeb alleges that
`
`the infringements at issue in this case all occur within, and are limited to, the PersonalWeb Patent
`
`Field. Accordingly, PersonalWeb has not provided notice to Level 3—under Section 6.4.1 of the
`
`Agreement or otherwise—that PersonalWeb desires to bring suit in the Level 3 Exclusive Field in its
`
`own name on its own behalf or that PersonalWeb knows or suspects that Defendant is infringing or
`
`has infringed any of Level 3’s rights in the patents.
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`1
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 3 of 26
`
`THE PARTIES
`
`4.
`
`Plaintiff PersonalWeb Technologies, LLC is a limited liability company duly organized
`
`and existing under the laws of Texas with its principal place of business at 112 E. Line Street, Suite
`
`204, Tyler, TX 75702.
`
`5.
`
`Plaintiff Level 3 Communications, LLC is a limited liability company organized under
`
`the laws of Delaware with its principal place of business at 100 CenturyLink Drive, Monroe,
`
`Louisiana, 71203.
`
`6.
`
`PersonalWeb’s infringement claims asserted in this case are asserted by PersonalWeb
`
`and all fall outside the Level 3 Exclusive Field. Level 3 is currently not asserting patent infringement
`
`in this case in the Level 3 Exclusive Field against any Defendant.
`
`7.
`
`Defendant Square, Inc. is, upon information and belief, a Delaware corporation having
`
`a principal place of business and regular and established place of business at 1455 Market Street, Suite
`
`600, San Francisco CA 94103.
`
`JURISDICTION AND VENUE
`
`8.
`
`The court has subject matter jurisdiction pursuant to 28 U.S.C. §§ 1331 and 1338(a)
`
`because this action arises under the patent laws of the United States, 35 U.S.C. §§ 1 et seq.
`
`9.
`
`Venue is proper in this federal district pursuant to 28 U.S.C. §§ 1391(b)–(c) and
`
`1400(b) because, on information and belief, Defendant has a regular and established place of business
`
`in this District and has committed acts of infringement in this District.
`
`10.
`
`This court has personal jurisdiction over Defendant because, in addition to the
`
`allegations in above paragraphs, on information and belief, Defendant is domiciled in this District.
`
`Further, Defendant purposefully directed activities at residents of California, the claims herein arise
`
`out of and relate to those activities, and assertion of personal jurisdiction over Defendant would be
`
`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
`
`fair.
`
`26
`
`27
`
`28
`
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`2
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 4 of 26
`
`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
`
`
`PERSONALWEB BACKGROUND
`
`11.
`
`The Patents-in-Suit cover fundamental aspects of cloud computing, including the
`
`identification of files or data and the efficient retrieval thereof in a manner which reduces bandwidth
`
`transmission and storage requirements.
`
`12.
`
`The ability to reliably identify and access specific data is essential to any computer
`
`system or network. On a single computer or within a small network, the task is relatively easy: simply
`
`name the file, identify it by that name and its stored location on the computer or within the network,
`
`and access it by name and location. Early operating systems facilitated this approach with standardized
`
`naming conventions, storage device identifiers, and folder structures.
`
`13.
`
`Ronald Lachman and David Farber, the inventors of the Patents-in-Suit, recognized
`
`that the conventional approach for naming, locating, and accessing data in computer networks could
`
`not keep pace with ever-expanding, global data processing networks. New distributed storage systems
`
`use files that are stored across different devices in dispersed geographic locations. These different
`
`locations could use dissimilar conventions for identifying storage devices and data partitions.
`
`Likewise, different users could give identical names to different files or parts of files—or unknowingly
`
`give different names to identical files. No solution existed to ensure that identical file names referred
`
`to the same data, and conversely, that different file names referred to different data. As a result,
`
`expanding networks could not only become clogged with duplicate data, they also made locating and
`
`controlling access to stored data more difficult.
`
`14.
`
`Lachman and Farber developed a solution: replacing conventional naming and storing
`
`conventions with system-wide “substantially unique,” content-based identifiers. Their approach
`
`assigned substantially unique identifiers to “data items” of any type: “the contents of a file, a portion
`
`of a file, a page in memory, an object in an object-oriented program, a digital message, a digital
`
`scanned image, a part of a video or audio signal, or any other entity which can be represented by a
`
`sequence of bits.” Applied system-wide, this invention would permit any data item to be stored,
`
`located, managed, synchronized, and accessed using its content-based identifier.
`
`15.
`
`To create a substantially unique, content-based identifier, Lachman and Farber turned
`
`to cryptography. Cryptographic hash functions, including MD4, MD5, and SHA, had been used in
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`3
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`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
`
`
`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 5 of 26
`
`computer systems to verify the integrity of retrieved data—a so-called “checksum.” Lachman and
`
`Farber recognized that these same hash functions could be devoted to a vital new purpose: if a
`
`cryptographic hash function was applied to a sequence of bits (a “data item”), it would produce a
`
`substantially unique result value, one that: (1) virtually guarantees a different result value if the data
`
`item is changed; (2) is computationally difficult to reproduce with a different sequence of bits; and
`
`(3) cannot be used to recreate the original sequence of bits.
`
`16.
`
`These cryptographic hash functions would thus assign any sequence of bits, based on
`
`content alone, with a substantially unique identifier. Lachman and Farber estimated that the odds of
`
`these hash functions producing the same identifier for two different sequences of bits (i.e., the
`
`“probability of collision”) would be about 1 in 2 to the 29th power. Lachman and Farber dubbed their
`
`content-based identifier a “True Name.”
`
`17.
`
`Using a True Name, Lachman and Farber conceived various data structures and
`
`methods for managing data (each data item correlated with a single True Name) within a network—
`
`no matter the complexity of the data or the network. These data structures provide a key-map
`
`organization, allowing for a rapid identification of any particular data item anywhere in a network by
`
`comparing a True Name for the data item against other True Names for data items already in the
`
`network. In operation, managing data using True Names allows a user to determine the location of
`
`any data in a network, determine whether access is authorized, and to selectively provide access to
`
`specific content not possible using the conventional naming arts.
`
`18.
`
`On April 11, 1995, Lachman and Farber filed their patent application, describing these
`
`and other ways in which content-based “True Names” elevated data-processing systems over
`
`conventional file-naming systems. The first True Name patent issued on November 2, 1999. The last
`
`of the Patents-in-Suit has expired, and the allegations herein are directed to the time period before
`
`expiration of the last of the Patents-in-Suit.
`
`19.
`
`PersonalWeb has successfully enforced its intellectual property rights against third
`
`party infringers, and its enforcement of the Patents-In Suit is ongoing. This enforcement has resulted
`
`in PersonalWeb obtaining settlements and granting non-exclusive licenses regarding the Patents-in-
`
`Suit.
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`4
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 6 of 26
`
`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
`
`
`GENERAL BACKGROUND
`
`20.
`
`A webpage is a type of document that is typically retrieved over the World Wide Web,
`
`made viewable and formatted (rendered) by a web browser, and displayed electronically. A “webpage”
`
`often refers to what is visible in a browser, but sometimes also refers to a computer file (“webpage
`
`base file”), usually written in Hypertext Markup Language (“HTML”) or a comparable markup
`
`language. Such HTML webpage base files typically include text, formatting, and references
`
`(hyperlinks) to other web content, such as style sheets, scripts, and images that make up part of the
`
`webpage. Web content referenced in an HTML or similar file are also called “asset files” herein. The
`
`web browser coordinates the retrieval of the various asset files of a webpage and renders the webpage
`
`for display from the webpage base file and the asset files referenced in the webpage base file or
`
`referenced in other asset files.
`
`21.
`
`On the World Wide Web, hyperlinks generally include Uniform Resource Identifiers
`
`(“URIs”), which each typically include an address of a server (“host”) from which the asset file is to
`
`be retrieved (e.g., “www.website.com”), a “path” to the location of that asset file on the host server
`
`(e.g., “/directory/”), and a filename (e.g., “filename.ext”).
`
`22.
`
`On the Internet, a web browser typically retrieves a webpage base file from a remote
`
`web server and retrieves referenced asset files from the same or different servers. The web browser
`
`retrieves a webpage base file or an asset file by making a GET “request” to a web server using the
`
`Hypertext Transfer Protocol (“HTTP”), an industry standard. The web server may respond to such an
`
`HTTP request with a HTTP “response” that includes the requested web content and may include other
`
`information or instructions.
`
`23.
`
`A static webpage is delivered exactly as stored, as web content in the web server’s file
`
`system or memory. In contrast, a dynamic webpage is generated by a web server application, usually
`
`driven by server-side software, upon receipt of a request from a browser (user). For example, a picture
`
`of a building might be delivered as static content (a picture) whereas the latest traffic conditions may
`
`be delivered dynamically based on real time traffic information.
`
`24.
`
`The speed of a browser retrieving webpage base files and incorporated asset files can
`
`be increased by the browser storing previously retrieved webpage base files and asset files in a browser
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`5
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 7 of 26
`
`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
`
`
`“cache” on the computer running the browser. If a browser’s user later requests a previously retrieved
`
`webpage base file or requests a webpage that includes an asset file previously used by the browser in
`
`rendering the same or a different webpage (for example, by reloading a webpage or visiting the same
`
`webpage again), the browser may use the cached webpage base file or asset file rather than having to
`
`download the same file repeatedly over the Internet again.
`
`25.
`
`Two computers communicating over the Internet usually are not directly connected to
`
`each other but rather interact via chains of network appliances and other computers (e.g., “switches”
`
`and “intermediate” servers). Many intermediate servers have caches similar to and complementing
`
`the browser cache that store webpage base files and assets that pass through that intermediate server.
`
`If a browser or server requests a file from the intermediate server that is present in that intermediate
`
`server’s cache, the intermediate server can use the content in its cache to respond to the request rather
`
`than send the request upstream towards the web server from which the file initially originated (also
`
`called the “origin server”).
`
`26.
`
`Responses to HTTP requests may include header elements (control elements) and a
`
`body (the “object” that was requested). Under HTTP, web servers can include a “cache-control”
`
`header with a response that includes a webpage or asset file. A “cache-control” header includes one
`
`or more directives that instruct browsers and intermediate server caches (“intermediate caches”) as to
`
`whether and for how long the file (object) included in the response may be cached or under what
`
`circumstances and under what conditions the cached content may be used. HTTP also provides for
`
`including other headers in responses that provide similar types of instructions to browsers and
`
`intermediate caches. Collectively, these other headers and directives in a “cache-control” header are
`
`referred to herein as “cache-control headers.”
`
`27.
`
`Given that webpage content changes, sometimes rather quickly and regularly, a
`
`problem that website owners face is effectively instructing a browser that is re-rendering a previously
`
`cached webpage that one or more of its cached files for that webpage are no longer the correct and
`
`authorized content (the content of those files has changed) and similarly reauthorizing the use of those
`
`cached files whose content has not changed.
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`6
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`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
`
`
`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 8 of 26
`
`28.
`
`On one hand, website owners want to encourage the browsers that render their web
`
`pages to use cached files thereby reducing the number of requests for these files that are being made
`
`to their webpage servers. Therefore, they frequently will set cache-control headers that authorize the
`
`browser to cache their webpage base files and asset files so the files are on hand when the browser
`
`needs to render that webpage again. On the other hand, website owners want the browsers to use the
`
`latest authorized files so that their users do not see the wrong content when viewing their webpage.
`
`DEFENDANT’S BACKGROUND
`
`29.
`
`On information and belief, Defendant has operated a website located at squareup.com,
`
`and has done so since before expiration of the last to expire of the Patents-in-Suit, which has operated
`
`to provide authorized webpage content to its users in the manner herein described.1
`
`30.
`
`On information and belief, Defendant’s web servers utilized a system of notifications
`
`and authorizations to control the distribution of content, e.g., what webpage content may be served
`
`from web servers and intermediate caches and what cached webpage content a browser is re-authorized
`
`to use to render Defendant’s webpage(s).
`
`31.
`
`On information and belief, Defendant’s system and its associated method of providing
`
`webpage content used “conditional” HTTP GET requests with If-None-Match headers and associated
`
`content-based ETag values for various webpage base files and asset files required to render various
`
`webpages of the Defendant.
`
`32.
`
`On information and belief, Defendant’s system and its associated method of providing
`
`webpage content also inserted fingerprints generated based on the content of asset files into the
`
`filenames of asset files required to render various webpages of the Defendant.
`
`33.
`
`On information and belief, Defendant’s system and associated method used these
`
`ETags and fingerprints to instruct both the intermediate cache servers and the endpoint caches at
`
`browsers to verify whether they were still authorized to reuse the previously cached webpage base
`
`
`
`1 While the complaint is sometimes written in the present or present perfect tense, all specific
`allegations are directed to the system’s operations and the method’s performance in the relevant time
`
`period.
`
`
`FIRST AMENDED COMPLAINT
`
`
`7
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 9 of 26
`
`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
`
`
`files of Defendant and to instruct them to obtain newly authorized content in rendering Defendant’s
`
`webpage when that content had changed. In other words, whether the previously cached content was
`
`still considered valid for use by the Defendant website operator.
`
`34.
`
`On information and belief, Defendant thereby reduced the bandwidth and computation
`
`required by its origin servers and any intermediate cache servers to field user requests to render
`
`Defendant’s webpages as those servers only need to serve files whose content has changed. On
`
`information and belief, this has allowed for the efficient update of cached information only when such
`
`content has changed, thereby reducing transaction overhead and bandwidth and allowing the
`
`authorized content to be served from the nearest cache.
`
`35. More particularly, on information and belief, each of Defendant’s webpages included
`
`a webpage base file (e.g., a main or initial HTML file) and one or more asset files referenced in the
`
`webpage base file (or referenced in other asset files that contained references to other asset files). On
`
`information and belief, the references in the webpage base file to the asset files needed to render the
`
`webpage were typically Uniform Resource Identifiers (“URIs”), which each typically included a
`
`filename, the address of a host server from which the asset file could be retrieved, and a “path” to the
`
`location of that asset file on that server.
`
`36.
`
`On information and belief, Defendant’s website used a web application framework to
`
`develop and compile various webpages of the Defendant, including asset files that were used in
`
`rendering the webpages, and to generate fingerprints of the contents of asset files. On information and
`
`belief, the fingerprints of individual asset files that were part of the webpage’s content were included
`
`in the respective filenames of the individual asset files. On information and belief, the modified
`
`filenames were then used as part of the URI used to access the individual asset files over the Internet.
`
`On information and belief, when an asset file’s content was changed, a new fingerprint was generated
`
`and included in the filename, its URI thus being changed accordingly.
`
`37.
`
`On information and belief, the asset file fingerprint was generated with a hash function
`
`and used to identify content changes. Furthermore, on information and belief, asset file URIs (with
`
`respective fingerprints) were included in webpage base files or other asset files contained references
`
`to other asset files. On information and belief, static webpage base files, if any, were recompiled when
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`8
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 10 of 26
`
`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
`
`
`any URI of a referenced asset file was changed (due to the fingerprint of the referenced asset file
`
`changing). Thus, a content change in an asset file for a given webpage would result in a change to its
`
`fingerprint, its URI, and a subsequent change to the content of any static webpage base files
`
`referencing that changed asset file for that webpage.
`
`38.
`
`On information and belief, a dynamic webpage base file generated for a webpage of
`
`Defendant webpages in response to one request from a user could be the same as it was when it was
`
`generated in response to a prior request from that or another user. However, on information and belief,
`
`this would not be the case if any of the asset files referenced in the webpage base file had changed
`
`between the time of the two requests and the URIs of the changed asset files included fingerprints as
`
`described above.
`
`39.
`
`On information and belief, when an asset file’s content was changed, a new fingerprint
`
`was generated and included in the filename, and its URI was thus changed accordingly, resulting in a
`
`content change to any webpage base file or other asset file that referenced that URI. This, in turn,
`
`caused a new and different ETag being generated for such webpage base file or other asset file that
`
`referenced that URI.
`
`40.
`
`On information and belief, for at least one of the asset files (“CBI ETag asset files”),
`
`the asset file comprised a sequence of bits and an associated ETag value was generated by Defendant
`
`by applying a hash function to the sequence of bits; wherein any two CBI ETag asset files comprising
`
`identical sequences of bits had identical associated ETag values. Thus, on information and belief,
`
`when a CBI ETag asset file’s content was changed a new associated ETag value was generated by
`
`Defendant. On information and belief, Defendant caused the origin server for each CBI ETag asset
`
`file to serve such CBI ETag asset file with its associated Etag value in response to HTTP GET requests
`
`for the CBI ETag asset file.
`
`41.
`
`On information and belief, Defendant contracted with Amazon to use Amazon’s S3
`
`system to store and serve at least some of Defendant’s CBI ETag files (“S3 asset files”) on its behalf.
`
`On information and belief, once Defendant’s S3 asset files were compiled and are complete, Defendant
`
`uploaded them to an Amazon S3 server as objects. On information and belief, such objects comprised
`
`a sequence of bits and, upon upload, an associated ETag value was generated by the S3 system on
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`9
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 11 of 26
`
`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
`
`
`behalf of Defendant by applying a hash function to the sequence of bits, wherein any two S3 asset
`
`files comprising identical sequences of bits had identical associated ETag values. On information and
`
`belief, in this way, Defendant generated the associated ETag values for its CBI ETag asset files that
`
`were S3 asset files. On information and belief, the S3 server for each S3 asset file served the S3 asset
`
`file with the its associated ETag value to HTTP GET requests for the S3 asset file.
`
`42.
`
`On information and belief, when Defendant created a webpage base file for a webpage,
`
`whether dynamic or static, that webpage base file included a sequence of bits and an associated ETag
`
`value was generated by Defendant by applying a hash function to the sequence of bits; wherein any
`
`two webpage base files comprising identical sequences of bits had identical associated ETag values.
`
`Thus, on information and belief, when a webpage base file’s content was changed and a new associated
`
`ETag value was generated by Defendant, it thereafter instructed the respective service by intermediate
`
`cache servers or use by endpoint caches such as browser caches to no longer use the previous cached
`
`webpage base file’s content. Conversely, when the webpage base file content had not changed and
`
`thus its ETag was unchanged, the cached asset files with fingerprints in their URIs referenced in the
`
`webpage base file had not changed and were still valid to use.
`
`43.
`
`On information and belief, when an intermediate cache server or a browser requested
`
`a webpage from the Defendant for the first time, it sent an HTTP GET request with the webpage’s
`
`URI and Defendant’s origin server or an upstream cache server responded by sending an HTTP 200
`
`(OK) response message containing the webpage base file, along with its respective associated ETag.
`
`On information and belief, a browser then sent individual HTTP GET requests, each with an asset
`
`file’s URI that was referenced in the webpage base file, and the asset files’ origin servers or
`
`intermediate cache servers responded by sending individual HTTP 200 responses containing the
`
`requested asset files, along with, if available, their respective associated ETags. On information and
`
`belief, upon receipt of the HTTP 200 responses, the intermediate cache server or browser cached the
`
`webpage base file and asset files with their associated URI and associated ETag values and the browser
`
`used them in rendering the requested web page of the Defendant. On information and belief, the origin
`
`servers, intermediate cache servers, and browser caches were caused to maintain databases/tables
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`10
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 12 of 26
`
`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
`
`
`which mapped the URIs of webpage base files and asset files to their respective responses and, if
`
`applicable, associated cache-control headers and ETags.
`
`44.
`
`On information and belief, by responding to an HTTP GET request for a given webpage
`
`by transmitting content of a webpage base file or asset file with an associated ETag, Defendant
`
`instructed the browser cache and all intermediate cache servers, to use an HTTP conditional GET
`
`request the next time that webpage base file or asset file is requested. More specifically, on information
`
`and belief, the browser or intermediate cache is instructed to include the ETag in the HTTP conditional
`
`GET request with an “If-None-Match” header to re-verify that they are still authorized to serve or use
`
`that content or determine that they are no longer authorized to use that content and therefore must use
`
`new content.
`
`45.
`
`On information and belief, Defendant did this, for example, by causing cache-control
`
`headers to be included in HTTP responses containing its webpage base file or asset files. On
`
`information and belief, Defendant benefits from using the ETags to control the distribution of its
`
`webpage content by communicating to a downstream cache and to a browser which of Defendant’s
`
`cached webpage base files it is reauthorized to serve/use and what newly authorized files it must first
`
`obtain in serving/rendering Defendant’s webpages.
`
`46. More particularly, on information and belief, when a browser again requested the
`
`Defendant’s webpage, the browser either used a cached copy, if allowed by the cache-control headers,
`
`or retrieved a new copy of the webpage base file for Defendant’s webpage. Similarly, on information
`
`and belief, for asset files referenced in the new or cached webpage base file, the browser either used a
`
`cached copy, if allowed by the cache-control headers, or retrieved a new copy of the asset files for
`
`Defendant’s webpage.
`
`47.
`
`On information and belief, for a webpage base file or an asset file stored in the
`
`browser’s cache with an ETag, and based on the cache-control headers received in the original
`
`response, the browser sent a conditional GET request with an If-None-Match header using the
`
`associated ETag value and the URI for the webpage base file or asset file so as to be notified whether
`
`the browser still had Defendant’s authority to render the webpage with its locally cached webpage
`
`
`
`
`FIRST AMENDED COMPLAINT
`
`
`11
`
`
`
`CASE NO. 5:18-MD-02834-BLF
`CASE NO. 5:18-CV-00183-BLF
`
`

`

`
`
`Case 5:18-cv-00183-BLF Document 32 Filed 10/03/18 Page 13 of 26
`
`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
`
`
`base file or asset file. In other words, whether the cached content was still valid for use in rendering
`
`Defendant’s webpage.
`
`48.
`
`On information and belief, under most circumstances, a responding intermediate cache
`
`server having content cached for the URI in the conditional GET request and having an ETag for that
`
`URI responded to the request by determining whether it had the same associated ETag value for that
`
`URI. If it had no ETag value for that URI, on information and belief, the request was passed up to an
`
`upstream intermediate cache server capable of responding or, if none, to the URI’s origin server, which
`
`responded to the request. On information and belief, if the intermediate cache server did not have
`
`content cached for the URI in the conditional GET request, the request was similarly passed up to an
`
`upstream intermediate cache server capable of responding or, if none, to the URI’s origin server.
`
`49.
`