`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`____________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`____________________
`
`EXPERT DECLARATION OF DR. GODMAR BACK IN SUPPORT OF
`THE PATENT OWNER’S RESPONSE
`
`APPLE, INC.,
`Petitioner
`
`v.
`
`REALTIME DATA LLC,
`Patent Owner
`
`____________________
`
`
`
` Case IPR2016-01737
`Patent 8,880,862
`
`____________________
`
`
`
`
`
`
`
`
`
`
`
`
`
`Realtime 2008
`Page 1 of 69
`
`
`
`
`
`TABLE OF CONTENTS
`
`
`
`I.
`
`INTRODUCTION .......................................................................................... 1
`
`A. Summary of Issues ........................................................................................... 1
`
`B. Summary of Opinions ...................................................................................... 4
`
`II.
`
`PROFESSIONAL BACKGROUND .............................................................. 5
`
`III. TECHNOLOGY OVERVIEW ....................................................................... 9
`
`A. ‘862 Patent ....................................................................................................... 9
`
`B. Sukegawa Reference ...................................................................................... 11
`
`C. Dye Reference ................................................................................................ 13
`
`D. Settsu Reference ............................................................................................ 14
`
`E. Burrows Reference ........................................................................................ 14
`
`F. Zwiegincew Reference .................................................................................. 15
`
`IV. LEGAL STANDARDS TO BE APPLIED .................................................. 18
`
`V.
`
`PERSON OF ORDINARY SKILL IN THE ART ....................................... 21
`
`VI. CLAIM CONSTRUCTION ......................................................................... 23
`
`A. Term “Boot Data List” ................................................................................... 23
`
`B. Term “Non-Accessed Boot Data” .................................................................. 29
`
`C. Apple’s Proposed Construction of “Boot Data” ............................................ 32
`
`VII. VALIDITY OF THE ‘862 PATENT ............................................................ 33
`
`A. Sukegawa Does Not Teach or Suggest the Claimed “Boot Data List.” ........ 33
`
`B. Zwiegincew’s Teachings Do Not Provide a Basis to Modify Sukegawa to
`Render Obvious the Claimed “Boot Data List.” ............................................ 39
`
`
`
`ii
`
`Realtime 2008
`Page 2 of 69
`
`
`
`
`
`C. Sukegawa Does Not Teach or Suggest “Disassociating Non-Accessed Boot
`Data from the Boot Data List.” ...................................................................... 44
`
`D. Sukegawa Does Not Teach or Suggest “Loading” Boot Data “That is
`Associated with a Boot Data List.” ................................................................ 47
`
`E. Settsu’s Teachings Do Not Provide a Basis to Modify Sukegawa to Render
`Obvious “Loading” Boot Data “That is Associated with a Boot Data
`List.” .............................................................................................................. 51
`
`F. Zwiegincew’s Teachings Do Not Provide a Basis to Modify Sukegawa to
`Render Obvious “Loading” Boot Data “That is Associated with a Boot Data
`List.” .............................................................................................................. 52
`
`G. Sukegawa in View of Dye Does Not Render Obvious the Challenged
`Claims. ........................................................................................................... 53
`
`H. Sukegawa Does Not Teach or Suggest “Boot Data Compris[ing] a Program
`Code Associated with…an Application Program.” ....................................... 59
`
`I. Dye Does Not Disclose Storing Compressed Boot Data in a Hard Disk
`Drive. ............................................................................................................. 60
`
`J. Dye Does Not Disclose “a Plurality of Encoders.” ....................................... 62
`
`
`
`iii
`
`
`
`
`
`Realtime 2008
`Page 3 of 69
`
`
`
`
`I, Godmar Back, declare as follows:
`
`I.
`
`INTRODUCTION
`
`
`
`1. My name is Dr. Godmar Back. I have been retained by Realtime Data
`
`LLC to offer my opinions concerning the validity of U.S. Patent No. 8,880,862
`
`(“the ‘862 Patent”).
`
`2.
`
`Specifically, I have been asked to analyze arguments made by Apple,
`
`Inc. and its expert, Dr. Charles J. Neuhauser, in the petition for inter partes review
`
`(“IPR”) proceeding of the ‘862 Patent, Case No. IPR2016-01737. I understand that
`
`on March 14, 2017, the Patent Trial and Appeal Board (“the Board”) entered a
`
`decision instituting (“the Institution Decision”) this IPR proceeding.
`
`A.
`
`3.
`
`Summary of Issues
`
`I understand that Apple’s Petition (and Dr. Neuhauser’s Declaration)
`
`allege the following five grounds of unpatentability:
`
`a. Ground 1: claims 1-4, 6-7, 13, 23-34, 47-58, 83-96, 99-100, 105-
`111, 113, and 116 of the ‘862 Patent are obvious over the
`combination of U.S. Patent No. 5,860,083 (“Sukegawa”) in view
`of U.S. Patent No. 6,145,069 (“Dye”);
`b. Ground 2: claims 1-4, 6-7, 13, 23-34, 47-58, 83-96, 99-100, 105-
`111, 113, and 116 of the ‘862 Patent are obvious over the
`combination of Sukegawa in view of Dye and U.S. Patent No.
`6,374,353 (“Settsu”);
`c. Ground 3: claims 1-4, 6-7, 13, 23-34, 47-58, 83-96, 99-100, 105-
`111, 113, and 116 of the ‘862 Patent are obvious over the
`
`1
`
`
`
`Realtime 2008
`Page 4 of 69
`
`
`
`
`
`
`
`combination of Sukegawa in view of Dye and Burrows et al., “On-
`line Data Compression in a Log-structured File System”
`(“Burrows”);
`d. Ground 4: claims 1-4, 6-7, 13, 23-34, 47-58, 83-96, 99-100, 105-
`111, 113, and 116 of the ‘862 Patent are obvious over the
`combination of Sukegawa in view of Dye, Settsu, and Burrows;
`and
`
`e. Ground 5: claims 1-4, 6-7, 13, 23-34, 47-58, 83-96, 99-100, 105-
`111, 113, and 116 of the ‘862 Patent are obvious over the
`combination of Sukegawa in view of Dye and U.S. Patent No.
`6,317,818 (“Zwiegincew”).
`
`4.
`
`I understand that in its Institution Decision, the Board instituted IPR
`
`on Ground 5 for claims 1–4, 6–7, 13, 23–34, 47–58, 83–96, 99–100, 105–111, 113,
`
`and 116 (“Challenged Claims”). On page 5 of the Institution Decision, the Board
`
`identifies Sukegawa, Dye, and Zwiegincew as the asserted prior art for Ground 5.
`
`Similarly, Apple’s Petition identifies those three references as the basis for Ground
`
`5.1 However, on page of 24 of the Institution Decision, the Board identifies Settsu
`
`as another reference at issue in Ground 5. I address the combination of Sukegawa,
`
`Dye, and Zwiegincew in my declaration. I also considered the combination of
`
`Ground 5 including the teachings of Settsu, and reached the same conclusions. As
`
`
`1 Petition at 2-3.
`
`
`
`
`2
`
`Realtime 2008
`Page 5 of 69
`
`
`
`
`such, my opinion regarding Ground 5 would not change if Settsu were one of the
`
`
`
`prior art references at issue in Ground 5.
`
`5.
`
`I understand that in its Institution Decision, the Board instituted IPR
`
`on Ground 1 (Sukegawa in view of Dye) for the Challenged Claims. I also
`
`understand that the Board instituted IPR on Ground 2 (Sukegawa in view of Dye
`
`and Settsu) for the Challenged Claims. Further, I understand that the Board
`
`instituted IPR on Ground 3 (Sukegawa in view of Dye and Burrows) for the
`
`Challenged Claims. Lastly, I understand that the Board instituted IPR on Ground 4
`
`(Sukegawa in view of Dye, Settsu, and Burrows) for the Challenged Claims.
`
`6.
`
`In forming my opinions, I have reviewed the ‘862 Patent, its file
`
`history, priority application 60/180,114 listed on the cover of the ‘862 Patent, Dr.
`
`Neuhauser’s declaration (“the Neuhauser Declaration”), Apple’s Petition for Inter
`
`Partes Review, the references upon which Apple’s Petition and Dr. Neuhauser
`
`rely, Realtime’s Preliminary Response, the Institution Decision, and materials
`
`referenced herein.
`
`7. My opinions are based on my experience and knowledge of the
`
`relevant art, the documents identified above, as well as the documents discussed in
`
`this declaration.
`
`
`
`
`3
`
`Realtime 2008
`Page 6 of 69
`
`
`
`
`
`B.
`
`8.
`
`Summary of Opinions
`
`As explained in detail below, it is my opinion that claims 1-4, 6-7, 13,
`
`
`
`23-34, 47-58, 83-96, 99-100, 105-111, 113, and 116 of the ‘862 Patent are not
`
`obvious over:
`
`a. Sukegawa in view of Dye;
`
`b. Sukegawa in view of Dye and Settsu;
`
`c. Sukegawa in view of Dye and Burrows;
`
`d. Sukegawa in view of Dye, Settsu, and Burrows; or
`
`e. Sukegawa in view of Dye and Zwiegincew.2
`
`9.
`
`Specifically, Sukegawa does not disclose “a boot data list,” as
`
`required by each Challenged Claim. Similarly, Sukegawa, alone or in view of
`
`Zwiegincew, does not render obvious (a) “disassociating non-accessed boot data
`
`from the boot data list” or (b) “updating the boot data list” in response to the
`
`“accessing step,” as required by claims 92, 100, 106, 111, 113, and 116. Sukegawa
`
`also does not disclose “load[ing]” or “accessing” boot data “that is associated with
`
`a boot data list,” as required by each Challenged Claim. Moreover, Sukegawa in
`
`view of the relevant prior art does not render obvious “load[ing]” or “accessing”
`
`
`2 As explained above, I also considered the combination of Sukegawa in view of
`
`Dye, Zwiegincew, and Settsu and reach the same conclusion that the Challenged
`
`Claims are not obvious over those references.
`
`
`
`
`4
`
`Realtime 2008
`Page 7 of 69
`
`
`
`
`boot data “that is associated with a boot data list.” Also, for each Challenged
`
`
`
`Claim, Dye teaches away from storing compressed boot data in Sukegawa’s hard
`
`drive for loading into flash memory 1. Even further, Sukegawa does not disclose
`
`“boot data compris[ing] a program code associated with…an application program,”
`
`as required by claims 29, 53, and 89. Lastly, Dye does not disclose “a plurality of
`
`encoders,” as required by claims 34, 58, and 94.
`
`10. As such, it is my opinion that the challenged claims are not obvious as
`
`alleged by Apple and Dr. Neuhauser.
`
`II.
`
`PROFESSIONAL BACKGROUND
`
`11.
`
`I have been working in the field of computer science for over 25
`
`years. My areas of expertise include computer systems, operating systems, and
`
`kernels. My experience includes, as a few examples, research, publications,
`
`lectures, and workshops in the field of computer systems, operating systems, and
`
`kernels. My Curriculum Vitae is attached hereto (Exhibit 2009).
`
`12.
`
`I obtained my undergraduate degree in Mathematics and Computer
`
`Science from Humboldt University of Berlin in 1992, and I studied Computer
`
`Science at the Technical University of Berlin from 1992-1994.
`
`13. From September 1994 to May 1995, I was a Teaching Assistant in the
`
`Department of Computer Science at University of Utah, where I co-taught senior-
`
`
`
`
`5
`
`Realtime 2008
`Page 8 of 69
`
`
`
`
`level undergraduate courses and entry-level graduate courses in operating systems,
`
`
`
`networking, and compilers.
`
`14. From June 1995 to November 2001, I was a Research Assistant in the
`
`Computer Systems Laboratory at University of Utah, where I conducted research
`
`on component-based operating systems (OSKit) and microkernel systems (Fluke).
`
`My research was published at the Second Symposium on Operating Systems
`
`Design and Implementation (OSDI) in 1996 and at the 16th ACM Symposium on
`
`Operating Systems Principles (SOSP) in 1997. Also during this time period, I
`
`conducted my dissertation research on runtime systems that support multiple
`
`applications. My research was published at the Seventh Workshop on Hot Topics
`
`in Operating Systems (HotOS) in 1999, at the Fourth Symposium on Operating
`
`Systems Design and Implementation (OSDI) in 2000, and at the USENIX 2000
`
`Annual Technical Conference in 2000. I also received travel scholarship awards
`
`from Usenix, ACM, and the IEEE for various conferences such as these.
`
`15.
`
`In May 2002, I received my Ph.D. in Computer Science from the
`
`University of Utah. I wrote my dissertation on the topic, “Isolation, Resource
`
`Management and Sharing in the KaffeOS Java Runtime System,” which went on to
`
`win the 2003 ACM SIGPLAN Doctoral Dissertation Award.
`
`16. Between November 2001 and June 2004, I was a Postdoctoral Scholar
`
`in the Computer Systems Laboratory at Stanford University. During my time at
`
`
`
`
`6
`
`Realtime 2008
`Page 9 of 69
`
`
`
`
`Stanford, I researched static analysis tools. As part of my research, I developed the
`
`
`
`MJ system for checking properties and implementing bug-finding analyses in Java
`
`code. I also worked on the design and implementation of DataScript, an input
`
`description language that supports code generation. I published my work on this
`
`language at the ACM Conference on Generative Programming and Component
`
`Engineering Proceedings (GPCE) in 2002. I also taught courses on “Introduction to
`
`Compilers” during my time at Stanford.
`
`17.
`
`In August 2004, I was appointed as Assistant Professor in the
`
`Department of Computer Science at Virginia Tech. In June 2010, I was promoted
`
`to Associate Professor, the position I currently hold. Between 2004 and 2015, I
`
`taught both graduate and undergraduate courses in “Operating Systems.” I have
`
`also taught undergraduate courses such as “Computer Systems,” “Introduction to
`
`Software Design,” “Systems and Networking Capstone,” and “Cloud Software
`
`Engineering,” and graduate courses such as “Advanced Topics in Program
`
`Analysis,” “Network Architectures and Protocols,” and “Execution Environments
`
`for Cloud Applications.” My current research interests include: operating and
`
`runtime systems, virtualization, software engineering, software visualization, web
`
`technology, cloud-based systems, high-performance computing, domain-specific
`
`languages, and library technology.
`
`
`
`
`7
`
`Realtime 2008
`Page 10 of 69
`
`
`
`
`
`
`
`18. Throughout my career, I have been an external reviewer for several
`
`professional publications and organizations, including the Journal of Parallel and
`
`Distributed Computing; the Journal of STEM Education; National Science
`
`Foundation; IEEE Transactions on Parallel and Distributed Systems; the Journal of
`
`Simulation Modelling Practice and Theory; IEEE Computer; ASEE Southeast
`
`conference; Proceedings of the IEEE; ACM Transactions on Programming
`
`Languages; the Journal of the ACM; Software Practice and Experience;
`
`Transactions on Information Systems (TOIS); USENIX; the Journal of Systems
`
`and Software; ICCD; SOSP; OSDI; PACT; ECOOP; EUROPAR; and the
`
`Informatik Forum Journal.
`
`19.
`
`I have also held
`
`leadership positions at several professional
`
`conferences and workshops. For instance, I have been the Program Co-Chair for
`
`the Operating Systems track at ICCD; a Program Committee Member for the
`
`International Conference on Parallel Processing (ICPP); a Program Committee
`
`Member of the International Workshop on Programming Support Innovations for
`
`Emerging Distributed Applications; and a Program Committee Member for the
`
`SPLASH/Wavefront conference.
`
`20.
`
`I have published 21 conference papers, nine journal articles, chapters
`
`in 2 books, and 11 workshop papers. Many of my works relate to computer
`
`systems, operating systems, and kernels.
`
`
`
`
`8
`
`Realtime 2008
`Page 11 of 69
`
`
`
`
`
`
`
`21. My compensation is not dependent on the outcome of this case, and I
`
`have no financial interest in the outcome.
`
`III. TECHNOLOGY OVERVIEW
`
`A.
`
`‘862 Patent
`
`22. The ‘862 Patent is generally directed to systems and methods for
`
`providing accelerated loading of operating systems and application programs in a
`
`computer system.3
`
`23. One method of increasing computer performance at the time of
`
`invention was the use of onboard memory and onboard caches. These onboard
`
`memories and caches are faster than the common-place magnetic HDD’s and thus
`
`allow host systems to quickly access necessary data.4 Thus, data is temporarily
`
`stored in a cache or other high-speed memory, and host systems do not have to
`
`wait for relatively slow hard drives to retrieve the needed data.
`
`24. Computers at the time of the invention suffered from slow boot times,
`
`even with high-speed onboard memories and caches.5 One reason for this is that
`
`upon reset, conventional boot device controllers of the time would wait for a
`
`3 Ex. 1001, ‘862 Patent at 1:20-26.
`
`4 E.g., Ex. 1001, ‘862 Patent at 1:29-31, 20:36-49; Ex. 1005, Sukegawa at 1:14-16,
`
`42-49.
`
`5 E.g., Ex. 1001, ‘862 Patent at 20:45-49; Ex. 1005, Sukegawa at 1:46-49.
`
`
`
`
`9
`
`Realtime 2008
`Page 12 of 69
`
`
`
`
`command before loading data for processing.6 Since boot device controllers are
`
`
`
`typically reset along with the host system, the time spent by the boot device
`
`controller waiting for commands is unproductive.7 Also, once the CPU is
`
`initialized and issues commands requesting data to the boot device controller, time
`
`is wasted while the CPU waits for the boot device controller to carry out the CPU’s
`
`command and provide the requested information. All this wasted time results in
`
`delay experienced by the user. Traditional high-speed memories of the time were
`
`also volatile memories, the contents of which are erased upon power reset.8
`
`Therefore, storing desired information, such as boot information, ahead of time
`
`was not possible.
`
`25. To address these problems, the ‘862 Patent discloses and claims
`
`methods and systems for loading compressed boot data associated with a boot data
`
`list, and updating the boot data list. Specifically, the claims of the ‘862 are directed
`
`to, among other things, maintaining a list of boot data, loading boot data in
`
`compressed form (based on the list) from a boot device into a cache memory,
`
`6 Ex. 1001, ‘862 Patent at 20:38-49.
`
`7 Ex. 1001, ‘862 Patent at 20:38-49.
`
`8 See Ex. 1005, Sukegawa at 1:21-26. See also Ex. 1003, Neuhauser Dec. at ¶ 44
`
`(recognizing that flash memory based designs were expensive on a per bit basis at
`
`the time of invention).
`
`
`
`
`10
`
`Realtime 2008
`Page 13 of 69
`
`
`
`
`accessing the loaded boot data and decompressing the boot data at a rate that
`
`
`
`decreases boot time of the operating system relative to loading the operating
`
`system with uncompressed boot data.9
`
`26. Another aspect of the inventions of the ‘862 Patent is updating the list
`
`of boot data during the boot process by adding to the list any boot data requested
`
`by the computer which was not previously stored in the list, as well as removing
`
`from the list any boot data previously stored in the list but not requested by the
`
`CPU.10 The ‘862 Patent’s system also includes a processor configured to load
`
`compressed boot data associated with a boot data list into memory, to access the
`
`loaded boot data, to decompress the access portion of boot data, and to update the
`
`boot data list.11 These systems and methods result in a faster bootup of computer
`
`systems.
`
`B.
`
`Sukegawa Reference
`
`27. Sukegawa describes a “permanent storage” solution in which
`
`“information…necessary for starting up…the OS and AP are permanently stored
`
`in the flash memory.”12 Sukegawa’s disclosed solution contrasts with the ‘862
`
`9 E.g., Ex. 1001, ‘862 Patent at 3:35-52, Claims 1, 5, 6, 8, 11, 13 and 14.
`
`10 E.g., Ex. 1001, ‘862 Patent at 3:53-59, Claims 1, 5, 6, 8, 11, 13, and 14.
`
`11 Ex. 1001, ‘862 Patent at 4:4-20.
`
`12 Ex. 1005, Sukegawa at 2:11-16. (emphasis added).
`
`
`
`
`11
`
`Realtime 2008
`Page 14 of 69
`
`
`
`
`Patent’s approach to reducing boot times that, during the same power-on cycle,
`
`
`
`loads compressed boot data associated with a boot data list, accesses the loaded
`
`portion, and decompresses the accessed portion at a rate that decreases boot time of
`
`the operating system.
`
`28. Sukegawa discloses that there was a problem with cache systems: the
`
`caches utilized a portion of the high-speed DRAM main memory (which is a
`
`volatile storage medium), and such memory was cleared when the power to the
`
`system is switched off. As a result, “the cache system does not function when the
`
`power is switched on.”13 To overcome this drawback, Sukegawa proposes using
`
`non-volatile memory to permanently store data needed for system startup instead
`
`of a traditional volatile cache.14
`
`29. Apple and Dr. Neuhauser contend that Sukegawa discloses the claim
`
`elements “a boot data list” and “loading boot data…that is associated with a boot
`
`data list.”15 I disagree, as explained below in Sections VII.A. and VII.E.
`
`
`13 E.g., Ex. 1005, Sukegawa at 1:50-61.
`
`14 Ex. 1005, Sukegawa at 1:53-61(Because “[t]he flash memory … is a non-volatile
`
`storage medium and has a higher access speed than the HDD,” “the cache function
`
`is effectively performed at the time of turning on power.”).
`
`15 E.g., Petition at 11-13.
`
`
`
`
`12
`
`Realtime 2008
`Page 15 of 69
`
`
`
`
`
`C. Dye Reference
`
`
`
`30. Dye describes a flash memory controller having a compression and/or
`
`decompression engine to support, for example, Execute-In-Place architectures,
`
`which results in improved memory density and bandwidth.16 Dye’s flash memory
`
`system comprises a flash memory array 100 and a Compression Enhanced Flash
`
`Memory Controller (“CEFMC”) 200.17 Dye’s memory controller (CEFMC 200)
`
`controls the transmission of small data segments (i.e., row and column data
`
`addressed in DRAM) to and from memory. Embedded within CEFMC 200 are
`
`compression and decompression engines 260, 280.18
`
`31. Dye
`
`does
`
`not
`
`teach
`
`or
`
`suggest
`
`using
`
`the
`
`disclosed
`
`compression/decompression system with traditional platter drives, and is instead
`
`limited to flash media. Also, Dye does not teach or suggest using the disclosed
`
`system in a data storage controller, as in Sukegawa, for accessing data sectors used
`
`to store and access operating system files. Therefore, a POSITA would not have
`
`looked to Dye when considering whether to apply compression to a data storage
`
`controller for a traditional hard drive such as Sukegawa’s HDD 2.
`
`
`16 Ex. 1008, Dye at Abs., Figs. 7-9, 2:32-39, 2:42-53.
`
`17 Ex. 1008, Dye at 8:29-31.
`
`18 Ex. 1008, Dye at Abs., 8:48-52.
`
`
`
`
`13
`
`Realtime 2008
`Page 16 of 69
`
`
`
`
`
`D.
`
`Settsu Reference
`
`
`
`32. Settsu is directed to “[a] method of booting up an information
`
`processing apparatus.”19 In addition, Settsu discloses a process for booting up a
`
`system that comprises a boot device divided into a mini-operating system (“OS”)
`
`module and an OS main body wherein modules of the OS main body may be
`
`stored as compressed files.20
`
`33.
`
`In Settsu’s system, certain application programs that need only a
`
`subset of the OS’s functionality can be started after loading only a subset of
`
`functional modules, thus speeding up boot times for this special case.21
`
`E.
`
`Burrows Reference
`
`34. Burrows is directed to a particular type of file system utilizing data
`
`compression and reports the results of tests of that system.22 Specifically, Burrows
`
`discloses a log-structured file system aimed at improving performance by
`
`eliminating non-sequential disk writes, which also uses compression software
`
`routines so that data occupies less space on disk.23
`
`19 Ex. 1006, Settsu at Abs.
`
`20 Ex. 1006, Settsu at Abs., 1:51-65, 3:6-12.
`
`21 E.g., Ex. 1006, Settsu at Abs., 1:51-65, 7:66-9:3.
`
`22 Ex. 1007, Burrows at v.
`
`23 Ex. 1007, Burrows at 8, 10.
`
`
`
`
`14
`
`Realtime 2008
`Page 17 of 69
`
`
`
`
`
`
`
`35. Burrow’s disclosures regarding its log-structured file system illustrate
`
`certain obstacles and challenges faced by a POSITA at the time of the filing of the
`
`‘862 Patent when trying to integrate compression into filesystem designs for
`
`traditional, platter-based hard drives.
`
`F.
`
`Zwiegincew Reference
`
`36. Zwiegincew is directed at the management of program code and data
`
`pages of application program modules during hard page fault intensive scenarios
`
`with the aim of improving performance. To understand hard page faults, it is
`
`helpful to understand virtual memory and paging in the context of modern
`
`computer systems.
`
`37. Virtual memory is a memory management technique that uses both
`
`hardware and software. When using virtual memory, program code utilizes virtual
`
`addresses that are mapped to the physical locations of the data in RAM. The blocks
`
`of data that are mapped in this way are known as pages.
`
`38. When a user or the system starts a new process, modern operating
`
`systems do not load the process’s program code into RAM all at once. Especially
`
`when a program is large, not all parts of the program may be needed, and loading
`
`them upfront would waste time and memory. Instead, these systems use a method
`
`called “on-demand paging,” where parts of a program are not loaded until the
`
`process running the program actually tries to execute them. If and when this
`
`
`
`
`15
`
`Realtime 2008
`Page 18 of 69
`
`
`
`
`happens, the OS recognizes which part of the program is requested, loads it from
`
`
`
`disk into memory, and resumes the process. This memory management process is
`
`performed by the OS’s virtual memory management module, commonly referred to
`
`as “virtual memory manager.” The virtual memory manager keeps track of which
`
`virtual addresses have been loaded to RAM and which ones have not. For virtual
`
`addresses that have been loaded to RAM, the virtual memory manager instructs the
`
`CPU’s memory management unit (MMU) where to find the physical address of the
`
`page in RAM.
`
`39. A hard page fault occurs when a process references a page in its
`
`virtual address space that has not been loaded to RAM. In this situation, the
`
`process is interrupted while the page is retrieved from the hard disk and loaded to
`
`RAM. The virtual memory manager updates its tables to indicate that the requested
`
`page is now available in RAM and identifies the location of that page in RAM. The
`
`process can then resume and utilize the page. Because handling a hard page fault
`
`requires accessing the hard disk which is much slower than RAM, these hard page
`
`faults slow down the process. If the OS knew which pages the program was likely
`
`to access, it could prefetch those pages into memory. Then, if the program
`
`accesses them, they are available when needed, thereby avoiding a “hard page
`
`fault.”
`
`
`
`
`16
`
`Realtime 2008
`Page 19 of 69
`
`
`
`
`
`
`
`40. To reduce the occurrence of hard page faults, Zwiegincew discloses
`
`that a “scenario file” may pre-fetch pages of application programs prior to the
`
`occurrence of a hard page fault sequence.24 In other words, Zwiegincew proposes
`
`to prevent hard page faults from occurring through the use of these “scenario
`
`files.”25
`
`41. Zwiegincew’s “scenario file” is a file that identifies characteristics,
`
`markers, or other indicators that a hard page fault is likely to occur—a so-called
`
`“page fault scenario.”26 The scenario file can also include a copy or identification
`
`of the page file that is needed to avoid the impending hard page fault.27 The
`
`system is monitored based on the information in the scenario file and when a hard
`
`page fault scenario is detected and thus a hard page fault is likely to occur, the
`
`system can load the page identified by the scenario file.28 Thus, the scenario file
`
`anticipates and prevents hard page faults, thereby increasing system speed.29
`
`
`24 Ex. 1010, Zwiegincew at 4:6-19.
`
`25 Ex. 1010, Zwiegincew at 4:6-19.
`
`26 Ex. 1010, Zwiegincew at Abs., Fig. 3.
`
`27 Ex. 1010, Zwiegincew at 6:64-67, 7:7-10.
`
`28 Ex. 1010, Zwiegincew at 6:29-39.
`
`29 Ex. 1010, Zwiegincew at 6:29-43.
`
`
`
`
`17
`
`Realtime 2008
`Page 20 of 69
`
`
`
`
`
`
`
`42. Zwiegincew further discloses the idea of automatically refining the
`
`scenario file so it can more accurately identify hard page fault scenarios.30
`
`Towards that end, Zwiegincew also discloses a mode in which hard page faults are
`
`recorded in a log, thus allowing a subsequent pattern-based algorithm to analyze
`
`this log to refine the page fault markers and indicators in the scenario file to better
`
`predict the occurrence of page faults.31
`
`43. Apple contends that Zwiegincew in combination with Sukegawa
`
`renders obvious “a boot data list,” and “loading” boot data “associated with a boot
`
`data list.” I disagree, as explained below in Sections VII.B. and VIII.F.
`
`IV. LEGAL STANDARDS TO BE APPLIED
`
`44.
`
`I understand that in an inter partes review, Apple carries the burden of
`
`proving invalidity on a claim-by-claim basis. Each claim must be analyzed
`
`independently. I also understand that Apple must prove invalidity by a
`
`preponderance of the evidence. To meet this burden, Apple must show that it is
`
`more likely than not that the claim is invalid.
`
`45.
`
`I understand that, under version of 35 U.S.C. § 103 of the patent
`
`statutes that applies to the ‘862 Patent, a patent claim may be invalid as obvious in
`
`
`30 Ex. 1010, Zwiegincew at 7:24-49.
`
`31 Ex. 1010, Zwiegincew at 6:30-37, 7:25-39, claim 2.
`
`
`
`
`18
`
`Realtime 2008
`Page 21 of 69
`
`
`
`
`view of a combination of prior art references. Obviousness is determined from the
`
`
`
`perspective of a hypothetical person of ordinary skill in the art (“POSITA”).
`
`46.
`
`I understand that obviousness may be based upon a combination of
`
`references. I further understand that the combination of familiar elements
`
`according to known methods is likely to be obvious when it does no more than
`
`yield predictable results. However, I also understand that a patent claim composed
`
`of several elements is not proved obvious merely by demonstrating that each of its
`
`elements was, independently, known in the prior art.
`
`47.
`
`I understand that when a patented invention is a combination of
`
`known elements, a court (or the Board) must determine whether there was an
`
`apparent reason to combine the known elements in the fashion claimed by the
`
`patent at issue by considering the teachings of prior art references, the background
`
`knowledge possessed by a person having ordinary skill in the art, and the effects of
`
`demands known to people working in the field or present in the marketplace.
`
`48.
`
` Further, I understand that a patent claim composed of several
`
`limitations is not proved obvious merely by demonstrating that each of its
`
`limitations was independently known in the prior art. I also understand that
`
`identifying a reason those elements would be combined can be important because
`
`inventions in many instances rely upon building blocks long since uncovered, and
`
`claimed discoveries almost of necessity will be combinations of what, in some
`
`
`
`
`19
`
`Realtime 2008
`Page 22 of 69
`
`
`
`
`sense, is already known. Accordingly, it is improper to use hindsight in an
`
`
`
`obviousness analysis and to use the patent’s claims as a “road map.”
`
`49.
`
`I also understand that a POSITA would not combine references in
`
`order to achieve a benefit or feature that is already present in one of the references
`
`being combined. Furthermore, I understand that an invention may not be obvious if
`
`the prior art reference(s) teaches away from the proposed modification or
`
`combination of prior art references. In addition, an invention may not be obvious if
`
`the proposed modification or combination would render the prior art inoperable for
`
`its intended purpose. As well, an invention may not be obvious if the proposed
`
`modification destroys the principle of operation of the prior art reference subject to
`
`modification.
`
`50.
`
`I understand that an obviousness analysis requires consideration of:
`
`(1) the scope and content of the prior art; (2) differences between the prior art and
`
`the claims at issue; (3) the level of ordinary skill in the art; and (4) objective
`
`indicia of non-obviousness. Examples of objecti