UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`MICROSOFT CORPORATION,
`Petitioner
`v.
`DAEDALUS BLUE, INC.
`Patent Owner
`____________
`
`Case No.: IPR2021-00831
`U.S. Patent No. 8,671,132
`Original Issue Date: March 11, 2014
`
`Title: SYSTEM, METHOD, AND APPARATUS FOR POLICY-BASED DATA
`MANAGEMENT
`____________
`
`REPLY DECLARATION OF DR. EREZ ZADOK IN SUPPORT OF
`INTER PARTES REVIEW OF U.S. PATENT NO. 8,671,132
`____________
`
`
`
`
`Microsoft Ex. 1027, p. 1
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`TABLE OF CONTENTS
`
`B. 
`C. 
`
`I. 
`QUALIFICATIONS ....................................................................................... 2 
`II.  MATERIALS RELIED UPON IN FORMING MY OPINIONS .................. 2 
`III.  UNDERSTANDING OF GOVERNING LAW ............................................. 3 
`IV.  RESPONSES TO THE DECLARATION OF DR. JULES WHITE
`(EX2024) ........................................................................................................ 3 
`A.  Dr. White’s Narrow Definition Of Distributed Systems Is Not
`In Line With The Understandings Of A POSITA ................................ 3 
`Gelb’s Storage System Is A Distributed Storage System .................. 11 
`Combining Gelb With Tivoli Would Not Have Required A
`Ground-Up Redesign .......................................................................... 13 
`D.  A POSITA Would Have Been Aware That Computing
`Platforms Have Hardware Differences .............................................. 19 
`E.  Maximum File Sizes Are Characteristics Of Storage Pools .............. 23 
`F. 
`Callaghan Is Properly Combinable With Tivoli ................................ 24 
`G.  Devarakonda Unambiguously Discloses Servers, Including
`Application Servers, That Use Different Operating Systems ............ 27 
`V.  DECLARATION .......................................................................................... 29 
`
`
`
`
`1
`
`Microsoft Ex. 1027, p. 2
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`I, Dr. Erez Zadok, declare as follows:
`
` My name is Erez Zadok. I have been retained by Petitioner Microsoft
`
`Corporation (“Petitioner” or “Microsoft”) to assist regarding U.S. Patent No.
`
`8,671,132 (Ex. 1001, “the ’132 Patent”). Specifically, I have been asked to consider
`
`the patentability of claims 15-25 of the ’132 Patent (“the Challenged Claims”) in
`
`view of prior art and the understanding of a person of ordinary skill in the art
`
`(“POSITA”) as it relates to the ’132 Patent. I have personal knowledge of the facts
`
`and opinions set forth in this declaration and believe them to be true. If called upon
`
`to do so, I would testify competently thereto. I have been warned that willful false
`
`statements and the like are punishable by fine or imprisonment, or both.
`
` My consulting company, Zadoks Consulting, LLC,
`
`is being
`
`compensated for my time at my standard consulting rate. I am also being reimbursed
`
`for any expenses that I may incur during the course of this work. My compensation
`
`is not contingent upon the results of my study and analysis, the substance of my
`
`opinions, or the outcome of any proceeding involving the Challenged Claims. I have
`
`no financial interest in the outcome of this matter or in any litigation involving
`
`the ’132 Patent.
`
` My opinions are based on my years of education, research, and
`
`experience, as well as my investigation and study of relevant materials, including
`
`those cited herein.
`
`1
`
`Microsoft Ex. 1027, p. 3
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`
`
`I may rely upon these materials, my knowledge and experience, and/or
`
`additional materials to rebut arguments raised by the Patent Owner. Further, I may
`
`also consider additional documents and information in forming any necessary
`
`opinions, including documents that may not yet have been provided to me.
`
` My analysis of the materials produced in this proceeding is ongoing and
`
`I will continue to review any new material as it is provided. This declaration
`
`represents only those opinions I have formed to date. I reserve the right to revise,
`
`supplement, and/or amend my opinions stated herein based on new information and
`
`on my continuing analysis of the materials already provided.
`
`I.
`
`QUALIFICATIONS
`
`I set forth my qualifications in my opening declaration, EX1003,
`
`incorporated by reference herein.
`
`II. MATERIALS RELIED UPON IN FORMING MY OPINIONS
`
`In forming my opinions and reaching the conclusion given in this
`
`declaration, I relied on the documents and materials cited in this declaration and in
`
`EX1003. These materials consist of patents, patent applications, related documents,
`
`and printed publications. This material consists of the type of documents upon
`
`which experts in the field would have relied. Appendix A to this declaration is a list
`
`of materials considered in drafting this declaration.
`
`2
`
`Microsoft Ex. 1027, p. 4
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`
`III. UNDERSTANDING OF GOVERNING LAW
`
`I discuss my understandings of the governing law in EX1003, which I
`
`incorporate by reference.
`
`IV. RESPONSES TO THE DECLARATION OF DR. JULES WHITE
`(EX2024)
`
`I have reviewed the declaration of Dr. Jules White, EX2024 submitted
`
`by Patent Owner in support of the Patent Owner Response in this proceeding (Paper
`
`No. 22). I disagree with many of the opinions that Dr. White has expressed. I
`
`provide in more detail below my opinions and reasons for why Dr. White’s opinions
`
`are incorrect.
`
`A. Dr. White’s Narrow Definition Of Distributed Systems Is Not In
`Line With The Understandings Of A POSITA
` Dr. White opines that Gelb is not a “distributed system” and, therefore,
`
`would not be considered analogous art and would not be combined with Tivoli.
`
`EX2024 (Dr. White POR Decl.), ¶¶99-109, 128. Dr. White purports to “define”
`
`what a “distributed computing” or “distributed system” is, citing for support to
`
`Verissimo. EX2024 (Dr. White POR Decl.), ¶¶32-38, 74-84. Dr. White lists “at
`
`least ten” characteristics of a distributed system that allegedly distinguish it over a
`
`“single computer system.” EX2024 (Dr. White POR Decl.), ¶¶74-84.
`
` Dr. White describes distributed systems that are in the most extreme,
`
`most complicated form possible. To be sure, some of the most complex distributed
`
`systems in existence can have some of the “at least ten” characteristics that Dr. White
`
`3
`
`Microsoft Ex. 1027, p. 5
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`cites. But Dr. White is wrong in his assumptions and assertion that all distributed
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`systems are on the extreme end of a wide spectrum of possible designs.
`
` Actual distributed systems can vary widely in design and complexity,
`
`as I testified in deposition. EX2025 (Zadok Depo Tr.), at 47:5-54:14, 58:2-18. There
`
`is a wide range of possibilities on this spectrum. Moreover, nothing in the ’132
`
`Patent suggests that the Challenged Claims are limited to this most extreme end of
`
`the design spectrum; and nothing in the Challenged Claims mentions “distributed”
`
`or even “networked” systems. Dr. White mis-characterizes both the ’132 patent and
`
`the field of distributed systems in a flawed attempt to suggest that a POSITA would
`
`not combine Gelb with Tivoli.
`
` Since Dr. White and the POR mis-characterize distributed systems, I
`
`provide an expanded discussion below.
`
` Andrew S. Tanenbaum is one of the world’s foremost authorities on
`
`distributed systems, networking, and operating systems. He has written many well-
`
`known textbooks on the subject. Even Tanenbaum in 1995 acknowledged variety
`
`in the definitions of distributed systems: “Various definitions of distributed systems
`
`have been given in the literature, none of them satisfactory and none of them in
`
`agreement with any of the others.” EX1032 (Tanenbaum), p. 2. (I owned at least
`
`one physical copy of this textbook by 2003.) Tanenbaum defines a distributed
`
`system as follows: “A distributed system is a collection of independent computers
`
`4
`
`Microsoft Ex. 1027, p. 6
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`that appear to the users of the system as a single computer.”1 EX1032 (Tanenbaum),
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`p. 2. Tanenbaum explains his definition: “This definition has two aspects. The first
`
`one deals with hardware: the machines are autonomous. The second one deals with
`
`software: the users think of the system as a single computer.” EX1032 (Tanenbaum),
`
`p. 2.
`
` Tanenbaum then states that “[r]ather than going further with definitions,
`
`it is probably more helpful to give several examples of distributed systems.”
`
`EX1032 (Tanenbaum), p. 2. He follows with three examples of growing complexity,
`
`again highlighting the fact that distributed systems are not as narrowly defined as Dr.
`
`White’s definition, but rather represent a wide spectrum.
`
` Tanenbaum’s first example is “a network of workstations in a
`
`university or company department.” He explains: “Such a system might have a
`
`single file system, with all files accessible from all machines in the same way and
`
`using the same path name. [] If the system as a whole looked and acted like a classical
`
`single-processor timesharing system, it would qualify as a distributed system.”
`
`EX1032 (Tanenbaum), pp. 2-3. In this example, users can execute programs on a
`
`local or remote workstation and all users share the same centralized file system.
`
` Tanenbaum’s second example “consider[s] a factory full of robots, each
`
`containing a powerful computer for handling vision, planning, communication, and
`
`
`1 All emphasis (bold, italics, or both) is original unless otherwise noted.
`
`5
`
`Microsoft Ex. 1027, p. 7
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`other tasks. When a robot on the assembly line notices that a part it is supposed to
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`install is defective, it asks another robot in the parts department to bring it a
`
`replacement. If all the robots act like peripheral devices attached to the same central
`
`computer and the system can be programmed that way, it too counts as a distributed
`
`system.” In this example, the individual computers (robots) have greater autonomy
`
`to communicate with others.
`
` Tanenbaum’s third example considers “a large bank with hundreds of
`
`branch offices all over the world. Each office has a master computer to store local
`
`accounts and handle local transactions. In addition, each computer has the ability to
`
`talk to all other branch computers and with a central computer at headquarters. If
`
`transactions can be done without regard to where a customer or account is, and the
`
`users do not notice any difference between this system and the old centralized
`
`mainframe that it replaced, it too would be considered a distributed system.” This
`
`third example posits a fairly complex distributed systems spanning a wide-area
`
`network, where banking transactional guarantees must be strictly enforced. This
`
`example is the only type of distributed system that Dr. White contemplates, and yet
`
`even great authorities in the field of distributed systems opine that there’s a much
`
`wider spectrum of possibilities.
`
` Tanenbaum is not the only textbook author to define distributed
`
`systems across a much wider spectrum than Dr. White. “Distributed systems is a
`
`6
`
`Microsoft Ex. 1027, p. 8
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`term used to define a wide range of computer systems, from weakly-coupled systems
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`such as wide area networks to strongly-coupled systems such as local area networks
`
`to very strongly-coupled systems such as multiprocessor systems.” EX1033
`
`(Goscinski), p. 3 (italicized emphasis added). (I owned a physical copy of Goscinski
`
`since the mid-1990s while studying Distributed Systems in college.)
`
` Like Tanenbaum, Goscinski identifies a wide spectrum of possibilities;
`
`distributed systems can run over wide-area networks (geo-distributed, the most
`
`complex form), local-area networks, and even multi-processor systems (a simpler
`
`form). Dr. White focuses primarily on complex distributed systems spanning large
`
`geographic distances. EX2024 (Dr. White POR Decl.), ¶82.
`
` Even Dr. White’s own textbook reference acknowledges that there is a
`
`spectrum of possibilities for distributed systems. “Distributed systems have many
`
`different facets which are very hard to capture by a single definition.” EX2026
`
`(Verissimo), p. 22.
`
`
`
`It is my opinion that Gelb discloses a distributed system under any of
`
`the definitions set forth above, and indeed under the definitions that Dr. White and
`
`Patent Owner set forth. EX2024 (Dr. White POR Decl.), ¶74; POR, 16. Dr. White
`
`attempts unsuccessfully to suggest that Gelb is a “single computer” and hence is so
`
`different from a distributed system (e.g., Tivoli) that a POSITA would never have
`
`thought to combine the two. EX2024 (Dr. White POR Decl.), ¶¶116-128. As I
`
`7
`
`Microsoft Ex. 1027, p. 9
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`discuss in further detail below, Gelb clearly discloses multiple host processors
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`communicating with multiple storage devices over multiple I/O channels, and thus
`
`is a form of a distributed system. EX1005 (Gelb), 5:49-56, Fig. 1. Even if Gelb is
`
`deemed not to disclose a (simple) form of a distributed system, Gelb indisputably
`
`discloses a multiple processor system that is a multi-computer with closely coupled
`
`processors (as Goscinski also describes above). Dr. White’s own Verissimo
`
`reference states that “[t]here is a thin border between multi-computers and
`
`distributed systems.” EX2026 (Verissimo), p. 24. Thus, while Dr. White would like
`
`to argue that Gelb’s system with multiple host processors is too far from being a
`
`distributed system to be combined with Tivoli, Dr. White’s own reference
`
`contradicts his testimony.
`
` Verissimo goes on to acknowledge the spectrum of possibilities of what
`
`distributed systems can be, which Dr. White ignores. “Current Internet and Web-
`
`based distributed architectures lie on this side of the spectrum, whereas the trendy
`
`LAN-based cluster architectures (Pfister, 1998) lie towards the multicomputer side
`
`of the spectrum.” EX2026 (Verissimo), p. 24.
`
` Dr. White keeps asserting that Gelb is a “mainframe” and that no one
`
`would ever think to turn a mainframe into a distributed system—or that at least it
`
`would be too difficult. EX2024 (Dr. White POR Decl.), ¶¶116-128. In effect, Dr.
`
`White attacks the motivation to combine Gelb with Tivoli by trying to “put as much
`
`8
`
`Microsoft Ex. 1027, p. 10
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`distance” between Gelb and Tivoli.2 Yet, Dr. White’s own reference, Verissimo,
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`says otherwise: “Note however that recent mainframe architectures are also
`
`extremely modular despite centralized, and can almost achieve the incremental
`
`expandability of DSs [Distributed Systems].” EX2026 (Verissimo), p. 25. In other
`
`words, Verissimo demonstrates that at least by 2001, Gelb’s system would have
`
`already been far closer to being a full-fledged distributed system than Dr. White
`
`argues.
`
` Dr. White further latches on to a discussion in Verissimo about
`
`mainframes being used to build distributed systems. EX2024 (Dr. White POR Decl.),
`
`¶¶85-87. Verissimo states that “Modern mainframes are modular and expansible
`
`inside, and open to other systems and to the Internet.” EX2026 (Verissimo), p. 27.
`
`And then immediately afterwards asks “What is wrong with a distributed system of
`
`mainframe servers?” Id. Dr. White, possibly realizing that these statements in
`
`Verissimo are not helpful to his position which tries to put “as much distance”
`
`between Gelb and Tivoli as possible, then attempts to re-interpret what Verissimo
`
`meant by stating that:
`
`the word “mainframe” can be used in two different ways. First,
`‘mainframe’ may be used to describe a centralized architecture, as it
`does in Gelb. Second, “mainframe” may be used to describe the type
`of underlying computing hardware.
`
`2 I note that Dr. White never argues that Tivoli is not a distributed system.
`
`9
`
`Microsoft Ex. 1027, p. 11
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`EX2024 (Dr. White POR Decl.), ¶85. Verissimo describes modern mainframes as
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`“extremely modular,” expandable, open systems—not the inflexible ones that Dr.
`
`White would like Gelb to be.
`
` That Dr. White does not like what Verissimo says about mainframes
`
`does not suggest that mainframes at the time could not be distributed systems or at
`
`least a component thereof. In fact, despite repeatedly arguing that essentially no one
`
`would consider Gelb to be a distributed system, Dr. White then testifies and admits
`
`that at the very least, mainframes could be used to build distributed systems: “A
`
`person of ordinary skill in the art would have taken Verissimo et al.’s discussion of
`
`mainframe servers to indicate at most that a distributed system could have used a
`
`piece of mainframe hardware as a server.” EX2024 (Dr. White POR Decl.), ¶87. In
`
`other words, Dr. White now admits that Gelb’s systems are not so distant from Tivoli,
`
`as one could at least build a distributed system using Gelb’s system.
`
`
`
`Indeed, Verissimo properly characterizes distributed computing not as
`
`a single extreme point on a design spectrum (as Dr. White argues), but rather as a
`
`wide range of possibilities and an evolution that spanned decades. “The first stage
`
`of evolution of distribution techniques was the sharing of files through file transfer
`protocols (ftp) [] Transferring files back and forth soon became a nuisance, and
`
`file sharing through distributed file systems was the next and obvious step. The idea
`
`seems simple now, but it revolutionized the panorama of distributed computing: a
`
`10
`
`Microsoft Ex. 1027, p. 12
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`program should be able to open a file (fopen) resident in a remote machine, and
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`should do it pretty much in the same way as a local file open. Furthermore, users
`
`should see a directory tree that looked unique, regardless of where the files were
`
`resident [].” EX2026 (Verissimo), p. 28.
`
`
`
`In sum, distributed systems are not a narrow point on the extreme end
`
`of a spectrum, as Dr. White argues, but a wide range of design possibilities that
`
`include multiprocessor systems and even mainframe computers (at least as
`
`components). Based on this broadly accepted view of distributed systems, Gelb’s
`
`multiprocessor system clearly qualifies as a distributed system. Moreover, even
`
`under Dr. White’s overly narrow views of distributed systems and mainframes, Gelb
`
`is a distributed system (or at least is far closer to being a distributed system or being
`
`used in one than Dr. White believes). A POSITA would most certainly have been
`
`motivated to combine Gelb and Tivoli.
`
`B. Gelb’s Storage System Is A Distributed Storage System
`
`In his opinions about Gelb being a “mainframe,” Dr. White opines that
`
`Gelb’s storage system is “directly connected” to Gelb’s host processors, implying
`
`that Gelb’s storage system is not a distributed storage system. EX2024 (Dr. White
`
`POR Decl.), ¶¶27-28, 71. I disagree.
`
` As I noted above (§IV.A), Gelb clearly discloses multiple host
`
`processors communicating with multiple storage devices over multiple I/O channels,
`
`11
`
`Microsoft Ex. 1027, p. 13
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`and thus is a form of a distributed system. EX1005 (Gelb), 5:49-56 (“FIG. 1 shows
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`a data processing system having a plurality of host processors 10 sharing a peripheral
`
`data storage system 12 via a plurality of input-output connections (such as I/O
`
`channels) 11.”), Fig. 1. Gelb teaches that its plurality of host processors are
`
`“sharing” a peripheral data storage system, via a plurality of input-output
`
`connections. EX1005 (Gelb), 5:49-56 (“FIG. 1 shows a data processing system
`
`having a plurality of host processors 10 sharing a peripheral data storage system 12
`
`via a plurality of input-output connections (such as I/O channels) 11.”). This teaches
`
`a POSITA that this connection is more than a mere direct one-to-one connection
`
`from a host processor to a storage device. Indeed, a POSITA would have understood
`
`that in a “shared” system like Gelb, all hosts can access all storage devices
`
`concurrently. In Gelb, the software running on each host processor interfaces with
`
`peripheral control programs that facilitate communications between each host
`
`processor and the storage system. EX1005 (Gelb), 15:55-61 (“A large plurality of
`
`application programs 30 execute on the host processor 10. Software connections to
`
`the peripheral control programs 31 are provided in the usual manner. Data Facility
`
`Product (DFP) 32 includes programming that implements the present invention.”).
`
`These peripheral control programs include a Data Facility Product (DFP) that
`
`implements the automatic class selection algorithms of Gelb. Id. The DFP sends
`
`commands to a data mover program that moves the data to the data storage. Id.,
`
`12
`
`Microsoft Ex. 1027, p. 14
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`15:63-67 (“Data mover program 35 actually causes data movement between main
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`memory (not shown) of the [h]ost processor and peripheral data storage 12.”).
`
` This structure is quite similar to that set forth in the ’132 Patent. There,
`
`the metadata servers implement the ’132 Patent’s storage policies, and send the data
`
`to the desired storage device, over network 130. In Gelb, the DFP and data mover
`
`programs perform the same function, using Gelb’s storage policies to send data to
`
`the desired storage device over shared I/O channels 38. Compare EX1001 (’132
`
`Patent), 9:21-34 with EX1005 (Gelb), 11:51-57, 15:55-67. A POSITA would have
`
`seen little if any difference between these two architectures, and would have
`
`considered a plurality of shared I/O channels to be a “network” in a distributed
`
`storage system.
`
`C. Combining Gelb With Tivoli Would Not Have Required A
`Ground-Up Redesign
` As I described in §IV.A, Gelb discloses multiple host processors and is
`
`thus very similar to Tivoli, and a POSITA would have been motivated to combine
`
`the two. Dr. White argues that a POSITA would not combine Gelb with Tivoli
`
`because it “requires a ground-up redesign of Gelb to be a distributed system.”
`
`EX2024 (Dr. White POR Decl.), ¶119. I disagree for at least four reasons.
`
` First, by the 2003 priority date of the ’132 Patent, building distributed
`
`systems was well known and taught in schools to undergraduate Computer Science
`
`students for more than a decade: Goscinski was published in 1991; Tanenbaum was
`
`13
`
`Microsoft Ex. 1027, p. 15
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`published in 1995; and Dr. White’s own reference, Verissimo, was published in 2001.
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`Taking Gelb’s teaching of policy-based storage and applying it to Tivoli’s
`
`distributed storage system would have been well within the knowledge of a POSITA
`
`at the time.
`
` Second, Dr. White asserts that “Tivoli does not provide the information
`
`that a person of ordinary skill in the art would have needed to navigate these
`
`complexities and successfully build a cross-platform client solution that completely
`
`redesigns Gelb to be a distributed system.” EX2024 (Dr. White POR Decl.), ¶124.
`
`I disagree. A POSITA would have understood that Tivoli does not need to describe
`
`these “complexities” because (1) building any simple-to-complex distributed system
`
`would have been well within the knowledge of a POSITA at the time and (2)
`
`distributed systems do not have to be the most extreme and complex form (see
`
`§IV.A).
`
` Third, Dr. White opines that “In the early 2000s, creating solutions that
`
`were cross-platform was extremely difficult.” I completely disagree. In the 1980s
`
`and early 1990s, there were many operating systems running on many different
`
`hardware platforms: these platforms ranged from 32- to 64-bit processor systems, as
`
`well as big- and little-endian processors.
`
`
`
`In the 1980s, writing code that would compile and run identically on
`
`different operating systems and hardware platforms could be challenging. But in the
`
`14
`
`Microsoft Ex. 1027, p. 16
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`1990s, the Free Software Foundation’s GNU project began developing tools and
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`techniques to help developers build cross-platform solutions. These tools were
`
`called “Autotools” and included Autoconf, Automake, and Libtool. With these tools,
`
`one could write portable code in C, C++, Fortran, and more languages. I was among
`
`many programmers using Autotools and I even personally contributed code to the
`
`Autotools project. I used Autotools to port the Berkeley Automounter (a service to
`
`automate access to many local and remote network file resources) to dozens of
`
`different operating systems and hardware platforms, and I published a paper on these
`
`experiences in June 2002.3 GNU Autotools has built-in mechanisms to handle
`
`differences between operating systems and hardware platforms, including CPU
`
`word-sizes, endianness, and many more.4
`
`
`3 See “Overhauling Amd for the ‘00s: A Case Study of GNU Autotools”,
`
`https://www.fsl.cs.stonybrook.edu/docs/autotools/autotools.pdf.
`
`4 My own PhD dissertation was concerned with building portable code for file
`
`systems running inside an operating system. It was demonstrated on Intel and
`
`SPARC architectures running the Linux, FreeBSD, and Solaris operating systems.
`
`See “FiST: A Language for Stackable File Systems,” published in June 2000,
`
`available at https://www.fsl.cs.stonybrook.edu/docs/fist-lang/fist.pdf.
`
`15
`
`Microsoft Ex. 1027, p. 17
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
` Fourth, the Network File System (NFS) is just another example of a
`
`highly successful and well-known example of a cross-platform solution to sharing
`
`files across the network. NFSv2 was published in 1985 by Sandberg and Sun
`
`Microsystems. NFSv2 became so popular that the next generation of the protocol,
`
`NFSv3, was released in 1994; and a third (NFSv4) released by 2001. As disclosed
`
`by Callaghan, NFS ran on many clients and servers—regardless of the operating
`
`system and hardware platforms they ran on—and permitted all nodes to access files
`
`uniformly and with ease. In other words, by 2003 a POSITA would have well known
`
`how to develop network protocols that work across any platform.
`
`
`
`In sum, Dr. White is wrong: by 2003, a POSITA would certainly have
`
`known how to build distributed systems and cross-platform solutions.
`
` Dr. White lists ten reasons why in his view Gelb could not be combined
`
`with Tivoli. Many of his points are the same or slight variations of the same. In the
`
`table below I summarize and respond to these arguments.
`
`
`
`Dr. White’s Position
`have
`systems
`¶75:
`“distributed
`unreliable messaging between elements
`of the system”
`
`¶76: “distributed systems often have
`clients
`on
`different
`computing
`platforms”
`and heterogeneity of
`architectures
`
`My Counter-Position
`Reliable communications via TCP/IP
`were already well known to a POSITA.
`For example, the WWW is a distributed
`system that uses reliable messaging for
`the HTTP protocol.
`As described above, building cross-
`platforms solutions was well known to
`a POSITA, whether different operating
`systems, different hardware platforms,
`or both.
`
`16
`
`Microsoft Ex. 1027, p. 18
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`Dr. White’s Position
`underlying
`¶77:
`“the
`networks
`distributed systems have more limited
`bandwidth”
`
`underlying
`networks
`“the
`¶78:
`distributed systems may have higher
`latency”
`
`¶82: geographic distance between,
`parts, difference
`in
`latency and
`bandwidth.
`distributed
`in
`¶79:
`“scheduling
`systems, especially those with diverse
`operating systems, is more difficult”
`
`¶80: “timing of resource management
`operations is much more difficult on a
`distributed
`system
`and
`requires
`advancing
`planning
`and
`synchronization across clients”
`
`¶83: “release and cleanup of resources
`is more complex
`in distributed
`systems”
`¶81: “in a distributed system where the
`actors all have to arrive at a consensus”
`
`¶84: “locking of resources is very hard
`to do in a distributed system”
`
`My Counter-Position
`These are the same point and assume
`that a distributed system runs on wide-
`area-networks, which have
`limited
`bandwidth and have high latencies. As
`discussed above in §IV.A, distributed
`systems have a wider range of design
`choices than this narrow view. Many
`distributed systems simply do not have
`these issues.
`
`These are all the same point. Once
`again, these points assume a narrow,
`extreme view of what a distributed
`system is. Also, as shown above in
`§IV.A, well-known textbooks already
`described how to schedule operations
`and manage resources in a distributed
`system.
`
`These essentially make the same point.
`Once again, these points assume the
`most extreme and complex form of a
`distributed system where a distributed
`consensus protocol (e.g., Paxos) is
`needed. Many distributed systems did
`not require such a protocol. Either way,
`Paxos and many variants thereof were
`already well known.
`
` Finally, I observe that Dr. White opines that “The purpose of the
`
`
`
`
`
`invention in Gelb was to make mainframe application programming easier and
`
`decouple programs from the underlying peripheral storage devices.” EX2024 (Dr.
`
`17
`
`Microsoft Ex. 1027, p. 19
`Microsoft v. Daedalus Blue
`IPR2021-00831
`
`

`

`
`White POR Decl.), ¶99. That’s simply untrue and mis-characterizes Gelb. While I
`
`Reply Declaration of Dr. Erez Zadok
`ISO Inter Partes Review of U.S. Patent 8,671,132
`
`note that Dr. White cited to the ’132 patent’s “field of invention” to support his
`
`arguments, he mis-characterizes Gelb and conveniently forgets to cite to Gelb’s
`
`“field of invention,” which states: “This invention relates to computer-controlled
`
`peripheral data storage and more particularly to the global management of data
`
`storage for provided [sic] transparent isolation of the peripheral data storage
`
`characteristics from application programs using the peripheral data storage.”
`
`EX1005 (Gelb), 1:9-14 (emphasis added).
`
`
`
`Indeed, Gelb and the ’132 Patent both address the same problem. Gelb
`
`addresses the problem that prior art data storage systems required the users to select
`
`desired data storage based on the characteristics of the data storage itself. EX1005
`
`(Gelb), 1:9-14 (“This invention relates to computer-controlled peripheral data
`
`storage and more particularly to the global management of data storage for provided
`
`transparent isolation of the peripheral data storage characteristics from application
`
`programs using the peripheral data storage.”). More particularly, prior art storage
`
`access methods did not permit programmers to write code that would allow users to
`
`automatically select the appropriate s