`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`International Business Machines Corporation and Oracle America, Inc.
`Petitioners
`
`go
`
`Electronics and Telecommunications Research Institute
`Patent Owner
`
`Case IPR2014-
`Patent 6,978,346
`
`DECLARATION OF ROBERT HORST, PH.D IN SUPPORT OF INTER
`PARTES REVIEW OF U.S. PATENT NO. 6,978,346
`
`Page 1 of 166
`
`IBM-Oracle 1003
`Page 1 of 166
`
`
`
`TABLE OF CONTENTS
`
`II.
`
`III.
`
`IV.
`
`A.
`
`B.
`
`V.
`
`VI.
`
`INTRODUCTION ......................................................................................... 4
`
`BACKGROUND ............................................................................................ 4
`
`LEVEL OF ORDINARY SKILL IN THE ART ........................................ 7
`
`APPLICABLE LEGAL STANDARD ......................................................... 8
`
`Claim Construction ............................................................................................................... 8
`
`Anticipation And Obviousness ............................................................................................... 9
`
`TECHNOLOGY OF THE ’346 PATENT ................................................ 11
`
`CHALLENGE #1 - CLAIMS 1-9 ARE RENDERED OBVIOUS BY
`THE MYLEX PAPER (Exhibit IBM-ORACLE-1006) IN VIEW OF
`THE TEACHINGS OF THE HATHORN PATENT (Exhibit IBM-
`ORACLE-1005) ............................................................................................ 17
`
`A. Brief Summary of the Mylex Paper ...................................................................................... 20
`
`B. Brief Summary of the Hathorn Patent ................................................................................... 25
`
`C. The Mylex Paper In View Of The Teachings Of The Hathorn Patent Compared To The ’346
`Patent, Claims 1-9 .............................................................................................................. 32
`
`VII.
`
`CHALLENGE # 2 - CLAIMS 1-9 ARE RENDERED OBVIOUS BY
`THE HATHORN PATENT (Exhibit IBM-ORACLE-1005) IN VIEW
`OF THE TEACHINGS OF THE MYLEX PAPER (Exhibit IBM-
`ORACLE-1006) ............................................................................................ 78
`
`A. The Hathorn Patent In View Of The Teachings Of The Mylex Paper Compared To The ’346
`Patent, Claims 1-9 .............................................................................................................. 81
`
`VIII.
`
`CHALLENGE # 3 - CLAIMS 1-9 ARE RENDERED OBVIOUS BY
`THE DEITZ PATENT (Exhibit IBM-ORACLE-1008) OR THE
`MYLEX PAPER (Exhibit IBM-ORACLE-1006) IN VIEW OF THE
`TEACHINGS OF THE GRIFFITH PATENT (Exhibit IBM-ORACLE-
`1007) OR THE DEKONING PATENT (Exhibit IBM-ORACLE-1010)
`¯....................................................................................................................~ 12
`
`A. Brief Summary of the Griffith and DeKoning Patents ........................................................... 113
`
`B. The Mylex Paper In View Of The Teachings Of The Griffith and/or DeKoning Patent Compared
`To The ’346 Patent, Claims 1-9 ......................................................................................... 119
`
`C. Brief Summary of the Deitz Patent ..................................................................................... 124
`
`Page 2 of 166
`
`IBM-Oracle 1003
`Page 2 of 166
`
`
`
`The Deitz Patent In View Of The Teachings Of The Griffith and/or DeKoning Patents Compared
`To The ’346 Patent, Claims 1-9 ......................................................................................... 126
`STATEMENT UNDER U.S.C. SECTION 1001 OF TITLE 18 ............ 165
`
`IX.
`
`Page 3 of 166
`
`IBM-Oracle 1003
`Page 3 of 166
`
`
`
`I.
`
`INTRODUCTION
`
`1. I have been retained on behalf of the Petitioners International
`
`Business Machines Corporation and Oracle America, Inc. to provide this
`
`Declaration concerning technical subject matter relevant to the inter partes review
`
`of U.S. Patent No. 6,978,346 (the "’346 patent"). I reserve the right to supplement
`
`this Declaration in response to additional evidence that may come to light.
`
`2. I am over 18 years of age. I have personal knowledge of the facts
`
`stated in this Declaration and could testify competently to them if asked to do so.
`
`II.
`
`BACKGROUND
`
`3. My name is Robert W. Horst. I am an independent consultant with
`
`more than 30 years expertise in the design and architecture of computer and
`
`storage systems. I hold a Bachelor of Science degree in Electrical Engineering
`
`from Bradley University, a Master of Science in Electrical Engineering from the
`
`University of Illinois at Urbana-Champaign, and a Ph.D. in Computer Science, also
`
`from the University of Illinois.
`
`4. Currently, I am an independent consultant for HT Consulting, where
`
`my work includes technology consulting and serving as an expert witness in patent
`
`and technology litigation. I am also Chief Technology Officer of Robotics for
`
`AlterG, Inc., where I am working on the design of orthotic devices to assist those
`
`with impaired mobility.
`
`Page 4 of 166
`
`IBM-Oracle 1003
`Page 4 of 166
`
`
`
`5. I have testified as an expert witness and served as a consultant in
`
`patent and intellectual property litigation and in interpartes review and re-
`
`examination proceedings.
`
`6. I have worked as a Technical Director at Network Appliance, Inc.,
`
`where I worked on processor and interconnect options for future generations of
`
`network-attached storage systems. I also served as Vice President of Research &
`
`Technology at 3ware, Inc., where I developed low-cost RAID controllers and
`
`initiated and led a proj ect on one of the industry’s first Ethernet Storage Area
`
`Network RAID storage systems. Prior to 3Ware, I worked as Technical Director at
`
`Tandem Computers/Compaq Computers where I contributed to the design and
`
`architecture of several generations of fault-tolerant systems. This work included
`
`development of CPUs, system-area networks, I/O systems, and storage systems. I
`
`have also been published in the areas of Networks, Storage, CPU Architecture,
`
`Fault Tolerance, and Bionics. In 2001, I was elected as an IEEE Fellow "for
`
`contributions to the architecture and design of fault tolerant systems and
`
`networks." I have worked with patent attorneys on numerous patent applications,
`
`and I am a named inventor on 78 issued U.S. patents.
`
`7. My qualifications and experience are set forth in more detail in my
`
`Curriculum Vitae, which is being filed as Exhibit IBM-ORACLE-1004.
`
`Page 5 of 166
`
`IBM-Oracle 1003
`Page 5 of 166
`
`
`
`8. I am being compensated for my work preparing this report. My
`
`compensation is not contingent upon the outcome of this proceeding or the
`
`opinions I develop in this matter.
`
`9. My preliminary opinions expressed herein are based on review and
`
`analysis of certain information obtained in connection with my work in this matter,
`
`together with my training, education, and experience. The opinions expressed
`
`herein are my own.
`
`10. The following identifies the information relied upon to date in
`
`connection with my work:
`
`(1)
`
`(2)
`
`(3)
`
`(4)
`
`(5)
`
`(6)
`
`(7)
`
`IBM-ORACLE-1001 : U.S. Patent No. 6,978,346 to Baek et al., foreign
`application priority date 9/19/2000 ("the ’346 patent");
`
`IBM-ORACLE- 1002: Excerpts from the Prosecution History of the ’346
`Patent;
`
`IBM-ORACLE-1005: U.S. Patent No. 5,574,950 to Hathorn et al., issued
`11 / 12/1996 ("Hathorn");
`
`IBM-ORACLE- 1006: Smith, Kevin J., "Storage Area Networks:
`Unclogging LANs and Improving Data Accessibility," Mylex Corporation
`White Paper (published 5/29/1998) ("Mylex paper");
`
`IBM-ORACLE-1007: U.S. Patent No. 6,401,170 to Griffith et al., filed
`on 8/18/1999 ("Griffith");
`
`IBM-ORACLE- 1008:
`10/28/1999 ("Deitz");
`
`IBM-ORACLE-1009:
`Archive;
`
`U.S. Patent No. 6,578,158 to Deitz et al., filed on
`
`Affidavit of Mr. Chris Butler, on behalf of Internet
`
`Page 6 of 166
`
`IBM-Oracle 1003
`Page 6 of 166
`
`
`
`(8)
`
`(9)
`
`IBM-ORACLE-1010: U.S. Patent No. 6,073,218 to DeKoning, et al., filed
`on 12/23/1996 ("DeKoning");
`
`IBM-ORACLE-1011: Clark, "Designing Storage Area Networks,’lSt
`Edition, Addison-Wesley Professional (1999);
`
`(10)
`
`IBM-ORACLE- 1012: Spainhower, "Design for Fault-Tolerance in System
`ES/9000 Model 900," IEEE (1992);
`
`(11)
`
`IBM-ORACLE- 1013: IEEE 100: Authoritative Dictionary of IEEE
`Standards Terms, 7th Edition (2000); and
`
`(12)
`
`IBM-ORACLE-1014: Siewiorek, D and Swarz R., "Reliable Computer
`Systems, Design and Evaluation," Digital Press (1992).
`
`III. LEVEL OF ORDINARY SKILL IN THE ART
`
`11. Based on my education and extensive experience relating to RAID
`
`storage systems and fault-tolerant systems, I believe I am qualified to provide
`
`opinions about the understanding and qualifications of a person of ordinary skill in
`
`the art of the technology at issue in this proceeding.
`
`12. In my opinion, a person of ordinary skill in the art of the ’346 patent,
`
`as of 2000, would have had a B.S. in Electrical Engineering or Computer Science
`
`and at least two years of experience in designing storage systems.
`
`13. My opinions below explain how a person of ordinary skill in the art
`
`would have understood the technology described in the references I have identified
`
`below around the 2000 time period.
`
`Page 7 of 166
`
`IBM-Oracle 1003
`Page 7 of 166
`
`
`
`IV. APPLICABLE LEGAL STANDARD
`
`A. Claim Construction
`
`14. I understand that in an interpartes review proceeding, the claims of a
`
`patent are to be given their broadest reasonable meaning as they would be
`
`understood by one of ordinary skill in the art, consistent with the specification of
`
`the patent.
`
`15. It is my understanding that the Patent Trial and Appeal Board
`
`previously construed certain ’346 patent claim terms in an interpartes review filed
`
`by Dell Inc., Hewlett-Packard Company, and NetApp. For the purposes of my
`
`opinions set forth herein, I have used the Patent Trial and Appeal Board’s
`
`constructions identified below. I reserve the right to offer an opinion as to the
`
`proper construction of other claim terms in this proceeding. At this time, I have no
`
`opinion as to whether these constructions would be the proper constructions for
`
`any district court litigation involving the ’346 patent.
`
`Claim Term
`"RAID controlling unit" and "RAID
`controller"
`"RAID"
`"exchange/exchanges information"
`
`"connection unit"
`
`Construction1
`"A component that controls operation of
`the RAID"
`"Redundant array of inexpensive disks"
`"To transmit and receive information
`reciprocally"
`"a hub or switch"
`
`See IPR2013-00635, Paper 19 at pp. 8-11.
`
`Page 8 of 166
`
`IBM-Oracle 1003
`Page 8 of 166
`
`
`
`16. Further, I understand that the Patent Owner stated in a prior IPR
`
`proceeding (IPR2013-00635, Paper 14 at p. 19) that a "network interface controller
`
`is the part of a RAID controller that allows the RAID controller to communicate
`
`with the ’connection units.’" Therefore, for purposes of this proceeding, I
`
`incorporate the construction of the claim terms "network interface controller,"
`
`"network controlling unit," and "network interface controlling unit," as "the part of
`
`a RAID controller that allows the RAID controller to communicate with the
`
`’connection units.’"
`
`B.
`
`17.
`
`Anticipation And Obviousness
`
`I have been informed that a patent claim is invalid as anticipated
`
`under 35 U.S.C. § 102 if each and every element of a claim, as properly construed,
`
`is found either explicitly or inherently in a single prior art reference. Under the
`
`principles of inherency, if the prior art necessarily functions in accordance with or
`
`includes the claimed elements, it anticipates.
`
`18. I have been informed that a claim is invalid under 35 U.S.C. § 102(a)
`
`if the claimed invention was known or used by others in the U.S., or was patented
`
`or published anywhere, before the applicant’s invention. I further have been
`
`informed that a claim is invalid under 35 U.S.C. § 102(b) if the invention was
`
`patented or published anywhere, or was in public use, on sale, or offered for sale in
`
`this country, more than one year prior to the filing date of the patent application
`
`Page 9 of 166
`
`IBM-Oracle 1003
`Page 9 of 166
`
`
`
`(critical date). I further have been informed that a claim is invalid under 35 U.S.C.
`
`§ 102(e) if an invention described by that claim was disclosed in a U.S. patent
`
`granted on an application for a patent by another that was filed in the U.S. before
`
`the date of invention for such a claim.
`
`19. I have been informed that a patent claim is invalid as "obvious" under
`
`35 U.S.C. § 103 in light of one or more prior art references if it would have been
`
`obvious to one of ordinary skill in the art, taking into account (1) the scope and
`
`content of the prior art, (2) the differences between the prior art and the claims, (3)
`
`the level of ordinary skill in the art, and (4) any so called "secondary
`
`considerations" of non-obviousness, which include: (i) "long felt need" for the
`
`claimed invention, (ii) commercial success attributable to the claimed invention,
`
`(iii) unexpected results of the claimed invention, and (iv) "copying" of the claimed
`
`invention by others. For purposes of my analysis above, and because I know of no
`
`indication from the patent owner or others to the contrary, I have applied a date of
`
`September 19, 2000, as the date of invention in my obviousness analyses, although
`
`in many cases the same analysis would hold true even at an earlier time than
`
`September 19, 2000.
`
`20. I have been informed that a claim can be obvious in light of a single
`
`prior art reference or multiple prior art references. To be obvious in light of a
`
`single prior art reference or multiple prior art references, there must be a reason to
`
`Page 10 of 166
`
`IBM-Oracle 1003
`Page 10 of 166
`
`
`
`modify the single prior art reference, or combine two or more references, in order
`
`to achieve the claimed invention. This reason may come from a teaching,
`
`suggestion, or motivation to combine, or may come from the reference or
`
`references themselves, the knowledge or "common sense" of one skilled in the art,
`
`or from the nature of the problem to be solved, and may be explicit or implicit
`
`from the prior art as a whole. I have been informed that the combination of
`
`familiar elements according to known methods is likely to be obvious when it does
`
`no more than yield predictable results. I also understand it is improper to rely on
`
`hindsight in making the obviousness determination. KSR Int ’l Co. v. Teleflex Inc.,
`
`550 U.S. 398, 421 (2007).
`
`V.
`
`TECHNOLOGY OF THE ’346 PATENT
`
`21. The ’346 patent relates to interconnections between host computers
`
`and storage systems. The storage systems referenced in the patent and claims are
`
`those known by the acronym RAID, which stands for Redundant Array of
`
`Inexpensive (or sometimes, Independent) Disks. The term RAID was first used in
`
`a 1987 paper by David Patterson and Randy Katz to describe storage systems built
`
`from multiple low-cost disk drives and configured to improve the reliability and/or
`
`performance of the storage system. Over time, multiple different levels of RAID
`
`were developed, including disk striping (RAID 0), disk mirroring (RAID 1), and
`
`various forms of parity protection across groups of drives (RAID 2 through RAID
`
`Page 11 of 166
`
`IBM-Oracle 1003
`Page 11 of 166
`
`
`
`6). The ’346 patent does not specifically address a particular type of RAID system
`
`and instead focuses on connections between the RAID and the host computers.
`
`22. By the 2000 time frame, several different types of networks were
`
`being used to connect RAID systems to hosts. Networks providing block-level
`
`access to storage were called SANs (for storage-area networks) and these networks
`
`included FCAL (fibre channel arbitrated loop), switched fibre channel and IBM’s
`
`ESCON (Enterprise Systems Connection). Other networks used for block-level or
`
`file level access to storage included ATM (Asynchronous Transfer Mode),
`
`ServerNet, InfiniBand and Ethernet. The ’346 patent gives examples of some
`
`networks for connecting storage, but does not make claims specific to any
`
`particular network. The ’346 patent groups them together, calling them "industrial
`
`standard communications networks," and also describes them using the general
`
`term "network" (’346 patent at 3:25-29).
`
`23. The terms "fault tolerance" and "fault tolerant system" describe
`
`systems that continue to function when part of the system encounters a fault. The
`
`concept of fault tolerant systems dates back to fault tolerant telephone switching
`
`systems introduced by AT&T in the 1960s and commercial fault-tolerant systems
`
`Page 12 of 166
`
`IBM-Oracle 1003
`Page 12 of 166
`
`
`
`first introduced by Tandem Computers in the 1970s.2 The focus of the ’346 patent
`
`is on the configuration of redundant network connections between hosts and RAID
`
`storage to assure that fault tolerance and performance are maintained when a RAID
`
`controller experiences a fault.
`
`24. I reproduced Figure 4 of the ’346 patent below:
`
`2 For a summary of commercial fault tolerant systems, see Siewiorek, D and Swarz
`
`R., Reliable Computer Systems, 1992. (Exhibit IBM-ORACLE-1014). Page 568
`
`shows redundant host to storage connections in the AT&T 3B20D system (1981).
`
`Page 589 shows redundant host to storage connections in the Tandem NonStop
`
`(1976). Page 619 shows redundant host to storage connections in the Tandem
`
`Integrity $2 (1991).
`
`Page 13 of 166
`
`IBM-Oracle 1003
`Page 13 of 166
`
`
`
`______ 1390
`HOST COMFurEFIl
`
`FIG. 4
`._ _____________ _
`__ ._ ._ ._
`_
`_
`1401
`1402
`1403
`"1.40“
`rHOST COMFUTER‘EINOST CONFUIER‘: [HOST CCMFOTER‘TIHOST COMPUTER}
`4| 1
`412
`413
`414
`‘
`If'.‘.-'II
`' 'l '
`J‘l
`1
`Yl‘. -;'
`INTERFACE
`INTERFACE
`OOHTHOIJ£FI
`IE!
`
`II
`{:
`
`4:55
`______ 1.. _
`rHOST COMPUTEFII
`‘
`Fl .1. V
`INTERFACE
`CONTROLLER
`
`'F-“r F-
`
`W whiff-El
`
`IE]
`
`'
`
`I
`
`r—___.—....—.____I IE!
`
`mar-R
`
`Iii!
`
`.53!
`
`IE:-
`EI-
`NETWORK INTERFACE NETwORK INTERFACE .
`CNTROLLER
`C
`CONTROLLER
`4?!)
`
`EI-
`NETWORK INTERFACE
`CONTROLLER
`
`25.
`
`The ’346 patent describes Figure 4 as “a block diagram showing one
`
`embodiment of a host interface system as an internal installment system between a
`
`RAID and host computers in accordance with the present invention.” (’346 patent
`
`at 2:46-49.) Figure 4 identifies a system where three “host computer[s]” (labeled
`
`400-402) are connected to a first “hub or switch” (labeled 440), and further
`
`connected to a RAID (labeled 490) through a network interface controlling unit
`
`port (labeled 470) on a first RAID controller (labeled 460) and a network interface
`
`controlling unit port (labeled 480) on a second RAID controller (labeled 461).
`
`Figure 4 also illustrates that three other “host computers” (labeled 403—405) are
`
`connected to a second “hub or switch” (labeled 441), and further connected to the
`
`RAID through a network interface controlling unit port (labeled 471) on the first
`
`Page 14 of 166
`
`IBM-Oracle 1003
`IBM-Oracle 1003
`Page 14 of 166
`Page 14 of 166
`
`
`
`RAID controller and a network interface controlling unit port (labeled 481) on the
`
`second RAID controller.
`
`26. As shown in Figure 4 above, network interface controlling unit port
`
`470 is networked through "hub or switch" 440 to network interface controlling unit
`
`port 480, and network interface controlling unit port 471 is networked through
`
`"hub or switch" 441 to network interface controlling unit port 481. The ’346
`
`patent discloses that "information" can be transmitted on these networks, but does
`
`not describe what type of information is transmitted.
`
`27. The ’346 patent does not disclose any specific modifications to the
`
`network interface controlling unit ports, the RAID controllers, or the hubs or
`
`switches that need to be made in order to transmit information on the networks
`
`between network interface controlling units. At most, the ’346 patent discloses
`
`that the mere addition of communication lines allowing both "hubs or switches" to
`
`connect both RAID controllers (e.g., in Figure 4, see lines labeled 450) is sufficient
`
`to allow communication between the network interface controlling units on two
`
`RAID controllers. The ’346 patent describes these communication lines broadly as
`
`follows:
`
`a communication line, representatively shown as 450 in
`the drawing, for connecting the network interface
`controller to the hub is a copper line or an optical fibre,
`which is matched to a corresponding standard.
`(’346
`patent at 3:39-42.)
`
`Page 15 of 166
`
`IBM-Oracle 1003
`Page 15 of 166
`
`
`
`28. I understand that the claims at issue in this proceeding are claims 1-9,
`
`with claims 1 and 9 being independent.
`
`29. Claim 1 of the ’346 patent recites the following:
`
`[1 a] An apparatus for a redundant interconnection between multiple hosts
`and a RAID, comprising:
`
`[1 b] a first RAID controlling units and a second RAID controlling unit for
`processing a requirement of numerous host computers, the first RAID
`controlling unit including a first network controlling unit and a second
`network controlling unit, and the second RAID controlling unit including a
`third network controlling unit and a fourth network controlling unit;
`
`[1 c] a plurality of connection units for connecting the first RAID
`controlling units and the second RAID controlling unit to the numerous
`host computers, wherein the first RAID controlling unit and the second
`RAID controlling unit directly exchange information with the connecting
`units, and the first network controlling unit exchanges information with the
`fourth network controlling unit, and the second network controlling unit
`exchanges information with the third network controlling unit.
`
`30. Claim 9 of the ’346 patent recites the following:
`
`[9a] An apparatus for a redundant interconnection between multiple host
`computers and a RAID, the apparatus comprising:
`
`[9b] a plurality of connecting units for connecting the host computers and
`the RAID;
`
`[9c] a first and a second RAID controllers, included in the RAID, each of
`which having a first network interface controller and a second network
`interface controller for processing requests from the plurality of the host
`computers connected through the plurality of the connection units,
`
`[9d] wherein the first network interface controller in the first RAID
`controller supplies data to the host computers connected through the
`plurality of connection units and processes information transmitted from
`the second network interface controller in the second RAID controller,
`
`Page 16 of 166
`
`IBM-Oracle 1003
`Page 16 of 166
`
`
`
`[9e] wherein the first network interface controller in the second RAID
`controller supplies data to the host computers connected through the
`plurality of connection units and processes information transmitted from
`the second network interface controller in the first RAID controller,
`
`[9f] wherein the second network interface controller in the first RAID
`controller is used for fault tolerance by performing functions of the first
`network interface controller in the second RAID controller when the
`second RAID controller is faulty, and
`
`[9g] wherein the second network interface controller in the second RAID
`controller is used for fault tolerance by performing functions of the first
`network interface controller in the first RAID controller when the first
`RAID controller is faulty, and
`
`[9h] wherein the first network controlling unit in the first RAID controlling
`unit exchanges information with the second network controlling unit in the
`second RAID controlling unit, and the second network controlling unit in
`the first RAID controlling unit exchanges information with the first
`network controlling unit in the second RAID controlling unit.
`
`VI. CHALLENGE #1 - CLAIMS 1-9 ARE RENDERED OBVIOUS BY
`THE MYLEX PAPER (Exhibit IBM-ORACLE-1006) IN VIEW OF THE
`TEACHINGS OF THE HATHORN PATENT (Exhibit IBM-ORACLE-1005)
`
`31. It is my opinion that the Mylex paper in view of the teachings of the
`
`Hathorn patent renders obvious claims 1-9 of the ’346 patent.3 I provide a brief
`
`description of the Mylex paper and the Hathorn patent below and then a more
`
`3 In setting forth my opinions regarding obviousness within this declaration, I am
`
`not setting forth the opinion that any reference does not anticipate any of the ’346
`
`patent’s claims.
`
`Page 17 of 166
`
`IBM-Oracle 1003
`Page 17 of 166
`
`
`
`detailed discussion identifying the disclosures in the Mylex paper and teachings of
`
`the Hathorn patent that support my opinion.
`
`32. As described below, a person of ordinary skill would understand that
`
`the Mylex paper discloses every element of the ’346 patent’s claims 1-9, with the
`
`exception of a direct exchange of information between network interface
`
`controlling units. Instead, the Mylex paper discloses a direct "heartbeat"
`
`communication path between controllers for exchanging information. The Hathorn
`
`patent, on the other hand, teaches that communication paths are expensive, and that
`
`this expense can be reduced by modifying network interface controlling unit ports
`
`to use the existing switch network for communications between RAID controllers
`
`(instead of using a direct "heartbeat" path).
`
`33. A person of ordinary skill in the art would have been motivated to
`
`apply these Hathorn patent teachings to the system disclosed in the Mylex paper in
`
`order to render every claim in the ’346 patent obvious. For example, both the
`
`Mylex paper and the Hathorn patent are in the same field of endeavor. Both
`
`references disclose redundant RAID systems that connect multiple hosts to
`
`switches or hubs, which in turn connect to RAID controllers with two or more
`
`ports. Both references disclose redundancy in terms of sending communications
`
`between two or more RAID controllers and/or network interface controlling unit
`
`ports. Additionally, both references are concerned with RAID 1 (disk
`
`Page 18 of 166
`
`IBM-Oracle 1003
`Page 18 of 166
`
`
`
`mirroring/shadowing). (See Mylex paper at 12 ("SAN-attached RAID arrays
`
`should support disk mirroring"); see Hathorn patent at 1:9-12 ("The present
`
`invention relates generally to remote data shadowing... ").) One of ordinary skill in
`
`the art would have been motivated to study multiple instances of systems for disk
`
`mirroring when designing a new RAID system. Further, both references disclose
`
`redundant RAID systems and disclose using off-the-shelf components for
`
`constructing the RAID system, and, as such, their combination is merely the use of
`
`known techniques to achieve predictable results. (See, e.g., Mylex paper at 15
`
`(marketing "Mylex controllers"); Hathorn patent at 6:25-34 (describing an IBM
`
`Enterprise Systems/9000 (ES/9000)processor running DFSMS/MVS operating
`
`software, IBM 3990 Model 6 storage controllers, and an IBM ESCON Director
`
`dynamic switch).)
`
`34. Finally, one of ordinary skill in the art would be motivated to combine
`
`the teachings of the Hathorn patent with the Mylex controllers because there was a
`
`close relationship between IBM, the assignee of the Hathorn patent, and Mylex
`
`Corporation. In September of 1999, IBM completed the acquisition of Mylex.
`
`Storage system designers at both companies in that timeframe would have been
`
`strongly motivated to combine and leverage storage technology from the other
`
`company. In fact, later IBM products were partly based on the acquired Mylex
`
`Page 19 of 166
`
`IBM-Oracle 1003
`Page 19 of 166
`
`
`
`technology, demonstrating that the motivation to combine the features actually
`
`resulted in new products.
`
`A.
`
`35.
`
`Brief Summarv of the Mvlex Paper
`
`The whitepaper titled "Storage Area Networks: Unclogging LANs and
`
`Improving Data Accessibility" by Kevin J. Smith of Mylex Corporation ("the
`
`Mylex paper") generally discloses Mylex’s Fibre Channel RAID controllers and
`
`the use of storage area networks to configure reliable and high-performance pools
`
`of storage. The Mylex paper was published on May 29, 1998, and made available
`
`on the Mylex public web site (www.Mylex.com). I understand that Petitioners are
`
`submitting a declaration by a business records custodian for archive, org
`
`identifying that the Mylex paper was publically available on the Mylex web site
`
`(www.Mylex.com) at least as early as February 4, 1999. (Exhibit IBM-
`
`ORACLE-1009).
`
`36. The Mylex paper includes illustrations of several configurations of the
`
`"seamless product line of external Raid Controllers" and explains that "Mylex
`
`array controllers are available in simplex configurations for network servers and
`
`duplex (dual) configurations for SAN’s and clusters. In duplex mode, advanced
`
`features are implemented to accelerate performance, protect data and guarantee
`
`data accessibility." (Mylex Paper at 14). While some of the figures focus on
`
`particular features of one configuration, one of ordinary skill in the art would
`
`Page 20 of 166
`
`IBM-Oracle 1003
`Page 20 of 166
`
`
`
`understand that the features could be combined in a single system. The Mylex
`
`paper discloses Fibre Channel arbitrated loop hub topologies and switched
`
`topologies, and combination hub and switch topologies that include redundant
`
`connections between hosts and RAID arrays to allow host-independent failover.
`
`For example, the Mylex paper states:
`
`Mylex controllers have dual SAN ports which doubles
`the bandwidth to controllers and allows redundant
`paths from other SAN devices to the controllers to
`increase the resiliency of the SAN topology. As
`described later in this paper, dual host ports are particular
`critical for controller failover in Fibre Channel
`topologies. The SAN ports can be connected directly to
`UNIX and NT servers or indirectly through hubs and
`switches. (Mylex Paper at 16, emphasis added.)
`
`37. For example, the Mylex paper’s Figures 6 and 20-21 are reproduced
`
`below, showing (i) a combination switch and hub cascade topology (Figure 6), (ii)
`
`hub topologies (Figure 20), and (iii) switch topologies (Figure 21):
`
`Fabric
`Loop .....
`
`Switch
`
`Figure 6.
`
`SAN With Switched and Shared (Loop) Interconnects
`
`Page 21 of 166
`
`IBM-Oracle 1003
`Page 21 of 166
`
`
`
`FC HBA
`
`FC HBA
`
`FC HBA
`
`FC HBA
`
`Hub
`
`Port 1
`
`Port 1
`
`Port 1
`
`Port 1
`
`FL Controller 0
`
`FL Controller 1
`
`FL Controller 2
`
`FL Controller 3
`
`Figure 20. Mvlex External Array Controllers in a Loop SAN Topology
`
`FC HBA
`
`FC HBA
`
`FC HBA
`
`FC HBA
`
`Switch
`
`Port 1
`
`Port 1
`
`Port 1
`
`Port 1
`
`FL Controller 0
`
`FL Controller 1
`
`FL Controller 2
`
`FL Controller 3
`
`Figure 21. Mvlex External Array Controllers in a Switched SAN Topology
`
`38. As shown in in the figures above, the Mylex paper discloses that
`
`multiple hosts (each associated with a fibre channel host-bus adapter "FC HBA" in
`
`Figures 20 and 21) may be connected to at least two RAID controllers (labeled,
`
`e.g., FL Controllers 0 and 1 in Figures 20 and 21) through a plurality of connecting
`
`Page 22 of 166
`
`IBM-Oracle 1003
`Page 22 of 166
`
`
`
`units (labeled as hubs and switches). While only one hub or switch is shown in
`
`Figures 20 and 21, one of ordinary skill in the art would understand the SANs
`
`implemented with hubs and switches are often implemented and extended with
`
`multiple hubs or switches. The RAID controllers include at least two network
`
`interface controlling units (labeled Port 0 and Port 1 in Figures 20 and 21).
`
`39. The Mylex paper also discloses a feature of the Mylex controller that
`
`provides high availability by heartbeat monitoring and transparent controller
`
`failover/failback. For example, the Mylex paper’s Figures 17 and 18 are
`
`reproduced below:
`
`HBA
`
`Controller 0 ~
`
`Hemtbea ts
`
`Port 2 Reserved
`
`~ Controller 1
`
`Figure 17.
`
`Normal Operating Mode of Dual DAC SF and FL Fibre Controllers
`
`Page 23 of 166
`
`IBM-Oracle 1003
`Page 23 of 166
`
`
`
`Figure 18.
`
`DAC SF and FL Fibre Array Controller Failover
`
`40.
`
`As shown in Figures 17 and 18, the Mylex paper teaches that RAID
`
`controllers can provide fault tolerance. Two Mylex RAID controllers exchange
`
`information by sending "heartbeat" signals through a path linking the two
`
`controllers. When one RAID controller has a fault, that fault is detected by the
`
`reserved network controller port on the non-faulty RAID controller, which assumes
`
`the network ID of the faulty network controller port in order to process the host I/O
`
`requests being sent to the faulty network controller port. As such, in the event of a
`
`fault occurrence, there is no loss in bandwidth to the RAID storage.
`
`41. The Mylex paper’s Figure 19 reproduced