UNITED STATES PATENT AND TRADEMARK OFFICE BEFORE THE PATENT TRIAL AND APPEAL BOARD. Oracle Corporation Petitioner,

UNITED STATES PATENT AND TRADEMARK OFFICE BEFORE THE PATENT TRIAL AND APPEAL BOARD Oracle Corporation Petitioner, v. Crossroads Systems, Inc. Patent Owner. IPR2015- U.S. Patent No. 7,934,041 PETITION FOR INTER PARTES REVIEW

TABLE OF CONTENTS EXHIBIT LIST... iv I. MANDATORY NOTICES... 1 A. Real Party-In-Interest... 1 B. Related Matters... 1 C. Lead and Back-Up Counsel... 1 D. Service Information... 1 II. PAYMENT OF FEES... 2 III. REQUIREMENTS FOR INTER PARTES REVIEW... 2 A. Grounds for Standing... 2 B. Identification of Challenge... 2 1. The Specific Art and Statutory Ground(s) on Which the Challenge Is Based... 3 2. How the Construed Claims Are Unpatentable Under the Statutory Grounds Identified in 37 C.F.R. 42.204(b)(2) and Supporting Evidence Relied Upon to Support the Challenge... 3 IV. BROADEST REASONABLE CONSTRUCTION... 4 V. GROUNDS OF UNPATENTABILITY... 6 A. Claims 1-53 are Rendered Obvious by 35 U.S.C. 103(a) by CRD-5500 User Manual in view of CRD-5500 Data Sheet and Smith... 6 1. Introduction of the CRD-5500 References... 6 2. Introduction of the Smith Reference... 9 3. Level of Skill in the Art... 10 4. The Combined System of CRD-5500 User Manual, CRD-5500 Data Sheet and Smith... 10 5. Correspondence Between Claims 1-53 and the CRD-5500 User Manual, CRD-5500 Data Sheet and Smith... 12 6. Secondary Indicia of Nonobviousness... 24 B. Claims 1-53 Are Rendered Obvious by Kikuchi taken in Combination with ii

Bergsten... 24 1. Introduction of the Kikuchi Reference... 24 2. Introduction of the Bergsten Reference... 25 3. The Combined System of Kikuchi and Bergsten... 27 4. Correspondence Between Claims 1-53 and Kikuchi and Bergsten... 30 C. Claims 1-53 Are Rendered Obvious by Bergsten taken in Combination with Hirai... 42 1. Introduction of the Hirai Reference... 42 2. The Combined System of Bergsten and Hirai... 44 3. Correspondence Between Claims 1-53 and Bergsten and Hirai... 47 VI. EXPLANATION OF NON-REDUNDANCY... 57 VII. EARLIER PROCEEDINGS AND 35 U.S.C. 325(D)... 58 VIII. CONCLUSION... 59 iii

EXHIBIT LIST 1001 U.S. Patent No. 7,934,041 ( the 041 Patent ) 1002 Select Portions of File History of the 041 Patent 1003 CRD-4400 SCSI RAID Controller User s Manual ( CRD-5500 User Manual ) 1004 CRD-4400 SCSI RAID Controller Data Sheet ( CRD-5500 Data Sheet ) 1005 Smith et al., Tachyon: A Gigabit Fibre Channel Protocol Chip, Hewlett- Packard Journal, October 1996 ( Smith ) 1006 U.S. Patent No. 6,219,771 to Kikuchi et al. ( Kikuchi ) 1007 U.S. Patent No. 6,073,209 to Bergsten ( Bergsten ) 1008 JP Patent Application Publication No. Hei 5[1993]-181609 to Hirai ( Hirai ) 1009 Infringement Contentions in Crossroads Systems, Inc. v. Oracle Corporation, W.D.Tex, Case No. 1-13-cv-00895, Crossroads Systems, Inc. v. Huawei Technologies Co. Ltd. et al., W.D.Tex, Case No. 1-13-cv-01025, and Crossroads Systems, Inc. v. NetApp, Inc., W.D.Tex, Case No. 1-14-cv-00149 1010 Declaration of Professor Chase, Professor of Computer Science at Duke University 1011 Cheating the I/O Bottleneck: Network Storage with Trapeze/Myrinet 1012 Interposed Request Routing for Scalable Network Storage 1013 Cut-Through Delivery in Trapeze: An Exercise in Low-Latency Messaging 1014 Structure and Performance of the Direct Access File System 1015 Implementing Cooperative Prefetching and Caching in a Globally- Managed Memory System iv

1016 Network I/O with Trapeze 1017 A Cost-Effective, High-Bandwidth Storage Architecture 1018 RAID-II: A High-Bandwidth Network File Server 1019 Payload Caching: High-Speed Data Forwarding for Network Intermediaries 1020 Petal: Distributed Virtual Disks 1021 File Server Scaling with Network-Attached Secure Disks 1022 Failure-Atomic File Access in an Interposed Network Storage System 1023 U.S. Patent No. 6,308,228 to Yocum et al. ( Yocum ) 1024 Select Portions of File History of Reexamination Control No. 90/007,123 (U.S. Patent No. 5,941,972) 1025 Select Portions of the File History of Reexamination Control No. 90/007,124 (U.S. Patent No. 6,421,753) 1026 Plaintiff Crossroads Systems, Inc. s Objections and Responses to Defendants First Set of Common Interrogatories in Crossroads Systems, Inc. v. Oracle Corporation, W.D.Tex, Case No. 1-13-cv- 00895, Crossroads Systems, Inc. v. Huawei Technologies Co. Ltd. et al., W.D.Tex, Case No. 1-13- cv-01025, and Crossroads Systems, Inc. v. NetApp, Inc., W.D.Tex, Case No. 1-14-cv-00149 1027 (SUBSTITUTE. Storagepath Fibre Channel Drive System, SWS/Storagepath, available at web.archive.org/web/19970114010450/http://www.storagepath.com/fi bre.htm, archived January 14, 1997 1028 Technology Brief Strategic Direction for Compaq Fibre Channel- Attached Storage, Compaq Computer Corporation, October 14, 1997 1029 Tantawy (ed., Fibre Channel (Ch. 5. of High Performance Networks, Kluwer Academic Publishers, 1994 v

1030 Deel et al., Moving Uncompressed Video Faster Than Real Time, Society of Motion Picture and Television Engineers, Inc., December 1996 1031 (SUBSTITUTE. Emulex LightPulse Fibre Channel PCI Host Adapter, Emulex Corporation, available at web.archive.org/web/19980213052222/http://www.emulex.com/fc/lig htpulse2.htm, archived February 13, 1998 1032 Select Portions of File History of Reexamination Control Nos. 90/007,125 and 90/007,317 (U.S. Patent No. 6,425,035) 1033 Local Area Networks Newsletter, Vol. 15, No. 2, Information Gatekeepers Inc., February 1997 1034 Litigation Complaint in Crossroads Systems, Inc. v. Oracle Corporation, W.D.Tex, Case No. 1-13-cv-00895Asserting Infringement of the 041 Patent 1035 Litigation Complaint in Crossroads Systems, Inc. v. Huawei Technologies Co. Ltd. et al., W.D.Tex, Case No. 1-13-cv-01025 Asserting Infringement of the 041 Patent 1036 Litigation Complaint in Crossroads Systems, Inc. v. NetApp, Inc., W.D.Tex, Case No. 1-14-cv-00149 Asserting Infringement of the 041 Patent 1037 Declaration of Monica S. Ullagaddi authenticating Ex. 1004, Ex. 1027 and Ex. 1031 vi

I. MANDATORY NOTICES Pursuant to 37 C.F.R. 42.8(a)(1), Petitioner provide the following mandatory disclosures. A. Real Party-In-Interest Pursuant to 37 C.F.R. 42.8(b)(1), Petitioner certifies that Oracle Corporation is the real party-in-interest. B. Related Matters Pursuant to 37 C.F.R. 42.8(b)(2), Petitioner states that the 041 Patent is asserted in co-pending litigation matters captioned Crossroads Systems, Inc. v. NetApp, Inc., W.D.Tex. Case No. 1-14-cv-00149, Ex. 1036, and Crossroads Systems, Inc. v. Oracle Corporation, 1-13-cv-00895, TXWD, filed October 7, 2013. The 041 patent is also the subject IPR2014-01177 (not instituted due to incorporation by reference from the expert declaration) and IPR2014-01463 (pending). All other related and copending litigation matters are set forth in Exhibit 1026. C. Lead and Back-Up Counsel Pursuant to 37 C.F.R. 42.8(b)(3), Petitioner provides the following designation of counsel: Lead counsel is Greg Gardella (Reg. No. 46,045) and back-up counsel is Scott A. McKeown (Reg. No. 42,866). D. Service Information Pursuant to 37 C.F.R. 42.8(b)(4), papers concerning this matter should be served on the following. 1

Address: Greg Gardella and Scott McKeown Oblon LLP 1940 Duke Street Alexandria, VA 22314 Email: cpdocketgardella@oblon.com, cpdocketmckeown@oblon.com Tel./Fax: (703) 413-3000 / (703) 413-2220 II. PAYMENT OF FEES The undersigned authorizes the Office to charge to Deposit Account No. 15-0030 the fee required by 37 C.F.R. 42.15(a) for this Petition for inter partes review. The undersigned further authorizes payment for any additional fees that might be due in connection with this Petition to be charged to the above referenced Deposit Account. III. REQUIREMENTS FOR INTER PARTES REVIEW As set forth below and pursuant to 37 C.F.R. 42.104, each requirement for inter partes review of the 041 Patent is satisfied. A. Grounds for Standing Petitioner certifies that is not estopped or barred from requesting inter partes review of the 041 Patent because this petition is accompanied by a motion for joinder. The one-year time bar of 35 U.S.C. 315(b) does not apply to a request for joinder. 35 U.S.C. 315(b) (final sentence) ( [t]he time limitation set forth in the preceding sentence shall not apply to a request for joinder under subsection (c) ); 37 C.F.R. 42.122(b). B. Identification of Challenge Pursuant to 37 C.F.R. 42.104(b) and (b)(1), Petitioner requests inter partes 2

review of claims 1-53 of the 041 Patent, and further requests that the Patent Trial and Appeal Board ( PTAB ) invalidate the same. 1. The Specific Art and Statutory Ground(s) on Which the Challenge Is Based Pursuant to 37 C.F.R. 42.204(b)(2), inter partes review of the 041 Patent is requested in view of the following grounds: (a) Claims 1-53 are rendered obvious under 35 U.S.C. 103(a) by the combination of the CRD-5500 SCSI RAID Controller User s Manual ( CRD-5500 User Manual, Ex. 1003), CRD-5500 SCSI RAID Controller Data Sheet ( CRD- 5500 Data Sheet, Ex. 1004), and Smith et al., Tachyon: A Gigabit Fibre Channel Protocol Chip, Hewlett-Packard Journal, October 1996 ( Smith, Ex. 1005); (b) Claims 1-53 are rendered obvious under 35 U.S.C. 103(a) by U.S. Patent No. 6,219,771 to Kikuchi et al. ( Kikuchi, Ex. 1006) in view of U.S. Patent No. 6,073,209 to Bergsten ( Bergsten, Ex. 1007); and (c) Claims 1-53 are rendered obvious under 35 U.S.C. 103(a) by Bergsten in view of JP Patent Application Publication No. Hei 5[1993]-181609 to Hirai ( Hirai, Ex. 1008). 2. How the Construed Claims Are Unpatentable Under the Statutory Grounds Identified in 37 C.F.R. 42.204(b)(2) and Supporting Evidence Relied Upon to Support the Challenge Pursuant to 37 C.F.R. 42.204(b)(4), an explanation of how claims 1-53 of the 041 Patent are unpatentable, including the identification of where each claim 3

element is found in the prior art, is provided in Section V below. Pursuant to 37 C.F.R. 42.204(b)(5), the exhibit numbers of the supporting evidence relied upon to support the challenges and the relevance of the evidence to the challenges raised, including identifying specific portions that support the challenges, are provided in Section V. IV. BROADEST REASONABLE CONSTRUCTION Petitioner bases this petition upon the broadest reasonable interpretation of the claim language. All claimed terms not specifically addressed in this section have been accorded their broadest reasonable interpretation in light of the patent specification including their plain and ordinary meaning. Petitioner s position regarding the scope of the claims under their broadest reasonable interpretation is not to be taken as stating any position regarding the appropriate scope to be given the claims in a court or other adjudicative body under the different claim interpretation standards that may apply to such proceedings. In particular, Petitioner notes that the standard for claim construction used in district courts differs from the standard applied before the U.S. Patent and Trademark Office ( USPTO ). Any claim construction offered by Petitioner in this petition is directed to the USPTO standard, and Petitioner does not acquiesce or admit to the constructions reflected herein for any purpose outside of this proceeding. Native low-level block protocol is described in the specification as being distinct from higher-level protocols that require translation to NLLBP. Ex. 1001 at 4

1:43-56; 3:43-54; 5:29-33. Examples of NLLBPs in the 041 Patent include SCSI-2 commands and SCSI-3 Fibre Channel Protocol (FCP) commands. See e.g., id. at 6:56-7:13. The specification distinguishes prior art systems that provided access through network protocols that the [network] server must translate into low level requests to the storage device. Id. at 2:12-14. During the reexamination of the grandparent patent the Patent Owner argued that a NLLBP is a set of rules or standards that enable computers to exchange information and do not involve the overhead of high level protocols and file systems typically required by network servers, citing the Markman Order of the U.S. District Court for the Western District of Texas in Crossroads v. Chaparral Network Storage, Inc., Civil Action No. A- 00-CA-217-SS and Crossroads v. Pathlight Technology, Inc., Civil Action No. A-00CA- 248-JN. 35 U.S.C. 301(a)(2); Ex. 1025 at p. 500, Patent Owner Response at p. 21. Consistent with this, the Examiner found that [t]he SCSI protocol/standard is considered a NLLBP. TCP/IP, e.g., used in Ethernet communications, however, is not considered to be a NLLBP. Id. at p. 14, Id. at NIRC a tp. 3. For the foregoing reasons, the broadest reasonable interpretation of NLLBP includes a protocol, such as SCSI command protocol, that enables the exchange of information without the overhead of high-level protocols and file systems typically required by network servers. 5

V. GROUNDS OF UNPATENTABILITY Each reference is introduced in turn and those introductions are followed by an explanation of the combined system or method and the supporting rationale. Thereafter, the correspondence between the combined system or method and each claim element is explained A. Claims 1-53 are Rendered Obvious by 35 U.S.C. 103(a) by CRD-5500 User Manual in view of CRD-5500 Data Sheet and Smith 1. Introduction of the CRD-5500 References The CRD-5500 SCSI RAID Controller User s Manual ( CRD-5500 User Manual ), Ex. 1003. and CRD-5500 SCSI RAID Controller Data Sheet ( CRD- 5500 Data Sheet", Ex. 1004. were published on November 21, 1996 and December 26, 1996, respectively, over a year before the earliest priority date of the 041 Patent (December 31, 1997). Therefore, the CRD-5500 User Manual and CRD- 5500 Data Sheet are prior art to the 041 Patent under 35 U.S.C. 102(b). The CRD-5500 User Manual and CRD-5500 Data Sheet were before the Examiner but were not discussed by the Examiner in any office action or referenced in any rejection. The CRD-5500 User Manual is presumed authentic under Fed.R.Evid. 901(b)(4) given that it was submitted by the Patent Owner as prior art and is selfauthenticating under Fed.R.Evid. 902(7) given that it bears trade inscriptions demonstrating that the document is a publication by CMD Technology, Inc. released on a date certain. Furthermore the CRD-5500 Data Sheet is authenticated by 6

the declaration of Monica S. Ullagaddi (Ex. 1037). The CRD-5500 User Manual describes a RAID controller which couples one or more host devices to virtual local storage on a RAID storage disk array. Ex. 1003 at 1-1. Devices are connected to the CRD-5500 controller through a number of I/O module slots configured to receive both host interface modules and storage device interface modules. Id. at 2-1. Figure 1-1 of the CRD-5500 User Manual illustrates how the controller's RAID set configuration utility can be used to configure virtual or logical storage regions, referred to as RAID sets, by assigning individual disk drives to logical groups. Ex. 1003 at 1-2. Each group may have a particular purpose and, as such, a particular configuration including, in some examples, striped partitions, data mirroring, or a combination thereof. Id.; see also id. at 1-5 and 1-7. The controller's Host LUN [Logical Unit Number] Mapping feature makes it possible to map RAID sets or redundancy groups (a RAID set or portion/partition thereof) differently to each host. Id. at 1-1; see also id. at 110; see also id. at 4-5. As illustrated in the Host LUN Mapping utility disclosed in the CRD-5500 User Manual, a particular host device (identified as Channel 0 ) is allotted access to one or more RAID redundancy groups (e.g., redundancy groups 0, 1, 5, and 6 through 31). The host device is provided an address for accessing each RAID redundancy group through a Host LUN (logical unit number, an addressing mechanism). See 7

e.g., id. at 4-5; 4-10; and 6-10. An administrator can allocate a particular disk as a redundancy group, such that a host LUN maps to a single physical disk or partition thereof. See, e.g., id. at 2-3, 2-4, 3-3, 3-4. Accordingly, the Host LUN Mapping utility of the CRD-5500 controller provides virtual local storage to a host device by presenting access to one or more RAID redundancy groups using LUN-based addressing. Id. at 4-5. Further, the Host LUN Mapping utility allows the CRD- 5500 controller to restrict a particular host s access to a given memory region on the RAID array by withholding addresses (i.e., Host LUNs ) for particular RAID redundancy groups to that host (e.g., redundancy groups 2 through 4 have been excluded from the list of redundancy groups for which Host LUNs have been assigned to the host illustrated). See e.g., id. at 1-1, You make the same redundancy group show up on different LUNs to different hosts, or make a redundancy group visible to one host but not to another. ; id. at 1-11, the CRD-5500 defines each RAID set or partition of a RAID set as a redundancy group. These redundancy groups may be mapped to host LUNs, either in a direct one-to-one relationship or in a manner defined by the user. The CRD-5500 Data Sheet notes that the modular design of the storage controller supports interfacing with host and/or storage devices via a high speed serial connection such as a Fibre Channel transport medium: Unlike other RAID controllers, CMD's advanced Viper RAID architecture and ASICs were designed to support tomorrow s high speed serial interfaces, such as Fibrechannel (FCAL)[]. 8

Ex. 1004 at p. 1 (emphasis added). 2. Introduction of the Smith Reference Smith et al., Tachyon: A Gigabit Fibre Channel Protocol Chip, Hewlett- Packard Journal, October 1996 ( Smith ) was published in October of 1996, over a year before the earliest priority date of the 041 Patent (December 31, 1997). Smith is therefore prior art under 35 U.S.C. 102(b). Smith describes the off-the-shelf Tachyon controller which is used in the preferred embodiment of the 041 Patent. Ex. 1001 at 6:30. The Tachyon chip is designed to serve as, among other things, a bridge between a Fibre Channel and a SCSI channel by encapsulating SCSI packets and sending them over the Fibre Channel. Ex. 1005 at 4. Indeed, [t]he second major design goal was that Tachyon should support SCSI encapsulation over Fibre Channel (known as FCP). From the beginning of the project, Tachyon designers created SCSI assists to support SCSI initiator transactions... Early in the design, Tachyon only supported SCSI initiator functionality with its SCSI hardware assists. It became evident from customer feedback, however, that Tachyon must support SCSI target functionality as well, so SCSI target functionality was added to Tachyon hardware assists. Id. The SCSI target functionality allows for mass storage support by adapting the Fibre Channel controller chip for use in a Fibre Channel target adapter card, such as the host interface modules of the CRD-5500. Id. at 3. 9

3. Level of Skill in the Art The prior art discussed herein demonstrates that a person of ordinary skill in the field, at the time the 041 Patent was effectively filed, was familiar with block storage systems (disks, RAID, and the SCSI command abstraction), storage volume management concepts, and networking technologies. Ex. 1010 14. A person skilled in the art would have at least an undergraduate degree in electrical or computer engineering or its equivalent and several years of experience with the foregoing technologies. This level of skill in the art is used for all grounds raised herein. 4. The Combined System of CRD-5500 User Manual, CRD-5500 Data Sheet and Smith It would have been obvious to one of ordinary skill in the art to combine the CRD-5500 User Manual, the CRD-5500 Data Sheet, and Smith to enhance the communication and storage options of a host computing device on a Fibre Channel transport medium, benefit from the Host LUN Mapping feature of the storage controller, and avail the host computing device of ubiquitous mass storage applications (e.g., RAID). Ex. 1010 41-45. This combination is specifically suggested in the CRD-5500 Data Sheet, which explains that CMD s advanced Viper RAID architecture and ASICs were designed to support tomorrow s high speed serial interfaces, such as Fibrechannel[] Ex. 1004 p. 1. The high bandwidth of Fibre Channels and capability of extended distances between hosts and the storage controller each provided a strong motivation to adopt the CRD- 5500 Data Sheet s 10

suggestion to enhance the CRD-5500 controller with the Fibre Channel-to-SCSI bridging capabilities of host and/or storage device modules designed with Tachyon chips. Ex. 1004 pp. 1-2. In the combined system, the Tachyon chip is incorporated into Fibre Channel enabled host device interface modules installed in I/O slots of the CRD-5500 controller. See e.g., Ex. 1010 47. Professor Chase explains that the Tachyon chip encapsulates and de-encapsulates SCSI commands on FC transport media that enables bridging to SCSI transport media. See e.g., id. 36, 38, 40, 43-45; Ex. 1005 at pp. 4, 9, and 10; see also Ex. 1004 at pp. 1-2. The CRD- 5500 controller, in the combined system, is configured to provide virtual local storage to up to four Fibre Channel host interface modules (each interfacing with a host computing device) through the Host LUN Mapping feature. See e.g., id. 37, 43-44, 46. A figure representing the combined system is shown at right. See e.g., Ex. 1003 at Fig. 1-2. In operation, the CRD-5500 controller coordinates the following process for managing a host storage command. Ex. 1010 11

44, 45. A FCP packet containing a SCSI storage access command (e.g., read or write request) is transmitted to the CRD-5500 controller by a host. Id. At the host interface module, the Tachyon chip de-encapsulates the FCP packet to access the SCSI command. Id. The host identity can be derived from either the incoming packet (e.g., FCP header or SCSI header) or the channel of the host module slot receiving the communication, if recognized. Id. The Tachyon chip pulls the host s storage address (e.g., LUN-based address) from the FCP header and passes the host identity and storage address as well as the SCSI payload to the CRD-5500 processor, where the host information is cross-referenced with the Host LUN Mapping maintained by the CRD-5500 controller to identify a redundancy group of the RAID storage disk array corresponding to the host storage address. Id. The CRD-5500 controller routes the SCSI command to the corresponding disk drive in the RAID array where the command is acted upon. Id. 5. Correspondence Between Claims 1-53 and the CRD-5500 User Manual, CRD-5500 Data Sheet and Smith The discussion below demonstrates the correspondence between the claim terms and the CRD-5500 User Manual, CRD-5500 Data Sheet, and Smith. 1. a. A storage router for providing virtual local storage on remote storage devices, comprising: The CRD-5500 Manual discloses this limitation because it teaches a CRD- 5500 RAID controller that routes data between host computers and SCSI disk 12

arrays. Ex. 1003 at 1-1, 1-4, 1-5 and 2-4. As discussed above, one or more redundancy groups can be created. Each redundancy group can be assigned a SCSI disk. Ex. 1003 at 1-5. An example of groupings of physical disk drives into redundancy groups is illustrated in Fig. 1-1 of the CRD-5500 User Manual (reproduced below). See id. at 1-1, 1-2. This provides a virtual storage region on each disk. Further, as detailed previously, the configuration provides for mapping the redundancy groups to host LUN addresses to provide the hosts with addressing capability to each redundancy group. See id. at 1-1, 4-5. As discussed above, the Host LUN Mapping utility of the CRD-5500 controller allows an administrator to assign RAID redundancy groups to host LUN addresses to provide the hosts with virtual local addressing capability to the remote storage partitions such that, from the viewpoint of the host, the remote storage device is local storage. Ex. 1003 at 1-1, 4-5; see also Ex. 1010 46-52. b) a first controller operable to interface with a first transport medium, wherein the first medium is a serial transport media; and The CRD-5500 as modified by Smith includes a first controller created through the incorporation of the Tachyon chip into a Fibre Channel host interface module designed for installation in a host I/O slot of the CRD-5500 controller, as detailed above. Ex. 1003, p. 1-1, 2-1, 2-4, 3-1. The first controller is operable to interface with a first transport medium (FC medium), and the first medium is a serial transport media. Fibre Channel is a serial transport protocol. Ex. 1001 at 13

1:54-60. As illustrated in Fig. 8 of Smith, for example, the Tachyon would interface with a SCSI transport medium via the host-based data structures, while the frame manager module of the Tachyon would interface with the FC transport medium (illustrated as the Link ). Ex. 1005 at 8, Ex. 1010 47. Professor Chase explains in his declaration that in the combined system, the FC-enabled host device interface module installed in the CRD-5500 controller is coupled to Fibre Channel, a serial transport media, on the host side. (See e.g., Ex. 1010 at 36, 47) c) a processing device coupled to the first controller, The CRD-5500 as modified by Smith includes a processing device coupled to the first controller. Ex. 1003 at 1-1, 1-3, 2-1, and Fig. 2-1. The Tachyon would interface with the MIPS R3000 RISC CPU of the CRD-5500 controller. Ex. 1003 at 1-3. Further, as explained above, a portion of the processing capability may be resident upon any or all of the interface modules installed in the nine I/O slots of the CRD-5500 Controller. See also discussion at 1010 48. d) wherein the processing device is configured to maintain a map to allocate storage space on the remote storage devices to devices connected to the first transport medium by associating representations of the devices connected to the first transport medium with representations of storage space on the remote storage devices, As described above, each redundancy group can be assigned a remote disk storage device. Ex. 1003 at 1-1, 1-2. The LUN mapping feature of the CRD-5500 can associate the redundancy groups with host devices. Ex. 1003 at 1-1, 4-5. As illustrated in the Host LUN Mapping utility (Id. at 4-5), the map associates storage 14

space on the remote disks assigned to the redundancy groups to devices connected to the first transport medium by associating representations of the devices connected to the first transport medium with representations of storage space on the remote storage devices. See also Ex. 1010 49. e) wherein each representation of a device connected to the first transport medium is associated with one or more representations of storage space on the remote storage devices; In the example of a first device connected to the first transport medium (e.g., the host at Channel 0), the host device is represented by Channel 0, and the storage space is represented by a series of RAID redundancy groups (e.g., one or more representations of storage space on the remote storage devices ) including redundancy groups 0, 1, 5, and 6 through 31. Ex. 1003 at 1-2, 4-5. Similar mappings can be configured for the remaining two to four hosts connected to the CRD-5500 controller. Id.; see also Ex. 1010 51. f) control access from the devices connected to the first transport medium to the storage space on the remote storage devices in accordance with the map; and The CRD-5500 s processing device (discussed above) uses the map to control access from the devices connected to the first transport medium to the storage space on the remote storage devices by accepting only host LUN addresses for which a mapping to a redundancy group exists and is associated with the requesting host. Ex. 1003 at 1-1, 4-5. As described by the CRD-5500 User Manual, 15

for example, the Host LUN Mapping utility allows an administrator to prevent a particular redundancy group from being visible to a particular host, effectively blocking access to that storage region by the particular host. See id. at 4-5; see also Ex. 1010 51. g) allow access from devices connected to the first transport medium to the remote storage devices using native low level block protocol. The CRD-5500 s processing device (discussed above) allow[s] access from devices connected to the first transport medium to the remote storage devices using native low level block protocol. Ex. 1003 at 4-18. Through the LUN Mapping utility of the CRD-5500 controller, a host device addresses remote RAID storage regions through LUN addressing, an addressing mechanism used to allow host devices to access and address storage as though it were local storage. See, e.g., id. at 4-10. The native low level block protocol carrying the LUN address is the FC or FCP protocol. See, e.g., Ex. 1001 at 8:22-25, claim 8. The CRD-5500, using the Host LUN Mapping, identifies a RAID redundancy group mapped to the LUN identified within the FCP request for the particular requesting host and allows access to the mapped RAID redundancy group. Id.; see also Ex. 1010 52. 2. The storage router of claim 1, wherein the map associates a representation of storage space on the remote storage devices with multiple devices connected to the first transport medium. The same RAID redundancy group can be mapped to a first LUN address for a first host and a second LUN address for a second host. Ex. 1003 at 4-5; see also 16

Ex. 1010 53. The combined system of the CRD-5500 controller with FC-enabled host interface modules functions identically in this regard. Additionally, in the combined system the storage devices may be accessed at a distance over FC links and therefore are remote as the term is used in the 041 Patent. See discussion of claim 1(d)-(g); see also Ex. 1010 53. 3. The storage router of claim 1, wherein the storage space on the remote storage devices comprises storage space on multiple remote storage devices. As described above, in the combined system, the RAID array consists of multiple SCSI drives. Ex. 1003 at 1-11. The CRD-5500 Controller, as discussed in the CRD-5500 User Manual, has multiple slots that may address storage devices. Ex. 1003 at 1-1. Each slot may address a number of individual disk drives. See id. at 6-10, Fig. 1.1, see discussion at Ex. 1010 at 54. 4. The storage router of claim 1, wherein the map associates a representation of a device connected to the first transport medium with a representation of an entire storage space of at least one remote storage device. The storage space of a RAID redundancy group may include an entire RAID set that spans the entire storage space of one or more storage devices. See id. at 1-10- 11, Figure 1-1, and 3-4 item 5; see also Ex. 1010 54. 5. The storage router of claim 1, wherein the map resides at the storage router and is maintained at the storage router. According to the CRD-5500 User Manual, the CRD-5500 controller is configured to include a mapping of Host LUNs to redundancy groups, and the 17

mapping can be updated by an administrator accessing the CRD-5500 controller through its user interface utility. See id. at 1-2 and 4-5. The mapping functionality provided by the CRD-5500 is transparent from the viewpoint of the hosts and is maintained at the CRD-5500 through its utility. See also Ex. 1010 55. 6. The storage router of claim 1, wherein the native low level block protocol is received at the storage router via the first transport medium and the processing device uses the received native low level block protocol to allow the devices connected to the first transport medium access to storage space specifically allocated to them in the map. As discussed above in relation to claim 1[g], the CRD-5500 uses the Host LUN Mapping to convert the target LUN given in a command issued by a host to a particular RAID redundancy group being accessed by the host. Ex. 1003 at 4-2 and 4-5. This redundancy group was specifically allocated to the host in the Host LUN Mapping during configuration by an administrator, in that the administrator must affirmatively enable access to the LUN and the redundancy group before a host can access the associated storage space. See id. at 4-5; see also Ex. 1010 56. 7. The storage router of claim 1, wherein the storage router is configured to receive commands according to a first low level block protocol from the device connected to the first transport medium and forward commands according to a second low level block protocol to the remote storage devices. As discussed in the CRD-5500 User Manual, the CRD-5500 issues write operations to SCSI-based RAID drives based upon write commands received from the connected hosts. Ex. 1003 at 1-4, 2-4, 4-18. These write operations would be presented in the SCSI protocol to be understood by the SCSI RAID array disk 18

devices. Id. at 4-18; Ex. 1010 57. The combined system of the CRD-5500 controller with FC-enabled host interface modules uses the FCP protocol as the NLLBP to communicate with hosts on the first transport medium. Ex. 1010 57. Additionally, in the combined system the storage devices may be accessed at a distance over FC links and therefore are remote as the term is used in the 041 Patent. Ex. 1010 57. 8. The storage router of claim 7, wherein the first low level block protocol is an FCP protocol and the second low level block protocol is a protocol other than FCP. As discussed above regarding claim 7, the combined system of the CRD-5500 controller with FC-enabled host interface modules uses the FCP protocol as the first NLLBP and the SCSI protocol as the second NLLBP. Ex. 1010 58. 9. The storage router of claim 1, wherein the map comprises one or more tables. See discussion above in connection with claim 1 regarding the Host LUN Mapping utility which assigns RAID redundancy groups to hosts, which is represented as a table in the CRD-5500 User Manual. The CRD-5500 controller additionally assigns RAID redundancy groups to RAID sets and further to physical drives, illustrated in the configuration examples in table format. Ex. 1003 at 3-5, Ex. 1010 59. 10. The storage router of claim 1, wherein the virtual local storage is provided to the devices connected to the first transport medium in a manner that is transparent to the devices and wherein the storage space allocated to the devices connected to the first transport medium appears to the devices as local storage. As discussed above, in the CRD-5500 system the hosts address the remote 19

storage using locally-generated low level LUN addresses encapsulated in a NLLBP (FCP) protocol. See also Ex. 1010 36, 38-45, 55, 60. In this manner, the location, number, and type of physical storage devices are completely transparent to the viewpoint of the host. Ex. 1010 60. 11. The storage router of claim 1, wherein the storage router provides centralized control of what the devices connected to the first transport medium see as local storage. In the CRD-5500, the LUNs and block addresses included within the transmissions from the host devices are mapped to physical disk drives in a manner that is completely transparent to the end host, and the CRD-5500 provides centralized control of the mapping via the Host LUN Mapping utility, as discussed above for claim 5. Ex. 1003 at 4-2 and 4-5; Ex. 1010 61. 12. The storage router of claim 1, wherein the representations of storage space comprise logical unit numbers that represent a subset of storage on the remote storage devices. As discussed in relation to claims 3, 4 and 9, a LUN exposed to a host may be associated with a redundancy group that represents a subset (single disk drive) of the array of a storage device, and/or with a redundancy group that is a partition of a storage device and therefore represents a subset of storage of a particular storage device of the RAID set. See also Ex. 1010 62. 13. The storage router of claim 12, wherein the storage router is operable to route requests to the same logical unit number from different devices connected to the first transport medium to different subsets of storage space on the remote storage devices. As discussed in connection with claim 2, in an example in the CRD-5500 User 20

Manual, a same RAID redundancy group can be mapped to a first LUN address for a first host (e.g., host channel 0 at LUN 5) and a second LUN address for a second host (e.g., host channel 1 at LUN 12). See id. at 4-5. As such, conversely, a particular LUN (e.g., LUN 2) may be used by the first host to access a first RAID redundancy group (e.g., redundancy group 2), while the second host uses that same LUN (LUN2) to access a second RAID redundancy group (e.g., redundancy group 4). See Ex. 1010 62, 63, and 66. 14. The storage router of claim 1, wherein the representations of devices connected to the first transport medium are unique identifiers. In the CRD-5500 system modified by Smith, host devices are identified by a channel identification, pertaining to the host module slot through which a particular host device connects. Ex. 1003 at 4-3 to 4-5. As such, each host device is associated with a unique device channel. See, e.g., id. at 4-3. Additionally, when the channel is a FC interface, the hosts connected via the FC network have unique identifiers defined by FC standards for FC host adapters. Ex. 1010 64. 15. The storage router of claim 14, wherein the unique identifiers are world wide names. World Wide Name (WWN) or World Wide Identifier (WWID) is a unique identifier used in Fibre Channel encapsulation to identify sending hardware. Ex. 1029 at 124; Ex. 1010 65. In the combined system the CRD-5500 controller identifies the sending host based upon the WWN or WWID in the FCP header. Id. 21

16. The storage router of claim 1, wherein the storage router is configured to allow modification of the map in a manner transparent to and without involvement of the devices connected to the first transport medium. The CRD-5500 maintains a Host LUN Mapping. Ex. 1003 at 4-5. Modification of the map is performed by an administrator and is transparent to the hosts. Ex. 1010 60-62, and 66. The administrator may configure the map such that a first host and a second host each use a same LUN to map to a different RAID set. 17. The storage router of claim 1, wherein the processing device is a microprocessor. As described above in relation to claim 1[c], the MIPS R3000 RISC CPU is a microprocessor device. Ex. 1003 at 1-3. 18. The storage router of claim 1, wherein the processing device is a microprocessor and associated logic to implement a stand-alone processing system. The stand-alone processing system of the combined system of the CRD-5500 would include the MIPS R3000 RISC CPU as well as the four custom ASIC components (e.g., XOR ASIC, DMA ASIC, CPU Interface ASIC, and Memory Controller ASIC). Ex. 1003 at 1-3. The ASIC components are associated logic devices. Further, a portion of the processing capability may be resident upon any or all of the interface modules installed in the nine I/O slots of the CRD-5500 Controller. Additionally, the microprocessor in the CRD-5500 system runs a software program ( associated logic ) to implement a stand-alone processing system in the same manner as the storage router of the 041 Patent. See also Ex. 1010 68. 22

19. The storage router of claim 1, wherein the first transport medium is a Fibre channel transport medium and further comprising a second transport medium connected to the remote storage devices that is a Fibre channel transport medium. Because the CRD-5500 s RAID architecture and ASICs were designed to support.fiberchannel [sic], it would be obvious to one of skill in the art to use FC-enabled slot modules in conjunction with the CRD-5500 to connect the controller with FC host devices and FC disk drives. Ex. 1027; Ex. 1010 69. 20. A storage network comprising: a set of devices connected a first transport... Independent claim 20 is similar to independent claim 1 but is a system claim that recites, in addition to the router, both the host devices and remote storage devices. The discussion set forth above in connection with claim 1 applies with equal force to claim 20. See also Ex. 1010 at 70-78. Dependent claims 21-36 Dependent claims 21-36 correspond to dependent claims 2-19. The discussion set forth above applies with equal force to claims 21-36. Ex. 1010 79-94. 37. A method for providing virtual local storage on remote storage devices comprising... Independent claim 37 recites the method performed by the apparatus recited in claim 1. The discussion set forth above applies with equal force to claim 37. See also Ex. 1010 95-100. Dependent claims 38-53 Dependent claims 38-53 correspond to dependent claims 2-19. The discussion 23

set forth above applies with equal force to claims 38-53. Ex. 1010 101-116. 6. Secondary Indicia of Nonobviousness Petitioner is aware of no objective secondary considerations that support the patentability of the challenged claims. Consistent with this, Patent Owner did not identify any such evidence in its preliminary responses filed in connection with IPR2014-01177. B. Claims 1-53 Are Rendered Obvious by Kikuchi taken in Combination with Bergsten 1. Introduction of the Kikuchi Reference U.S. Patent No. 6,219,771 to Kikuchi et al. (Ex. 1006) was filed on August 18, 1997, before the earliest priority date of the 041 Patent (December 31, 1997). Therefore, Kikuchi is prior art to the 041 Patent under 35 U.S.C. 102(e). Kikuchi was among the hundreds of references before the Examiner but was not discussed by the Examiner in any office action or referenced in any ground of rejection. Kikuchi describes a control device that receives commands from host devices via a host device interface, determines whether the host is authorized to access a specified storage device, and interprets and executes the commands from the host device. Ex. 1006 at the Abstract. The host devices are connected via a Fibre Channel transport medium to the control device and the control device is, in turn, connected to a storage unit via a SCSI transport medium. Id. at 1:31-36; see also id. at 5:37-39. Kikuchi executes access control by extracting a host address from each command 24

and determining whether the address is registered in an address registration unit. Id. at 4:35-44; see also id. at 5:3-6. Kikuchi enables multiple host devices on a Fibre Channel to access different partitions of a single SCSI storage device. Id. at 1:31-43. Kikuchi describes extracting a host address from a command sent from a host device and using the host address to obtain offset information, which indicates a disk partition corresponding to each host device. See e.g., id. at 2:52-62; 3:6-17. Kikuchi also describes extracting a disk partition address from each command. Id. at 7:53-63. An actual disk partition address is obtained by combining the offset information and the disk partition address. Id. at 7:64-8:3. Thus, in Kikuchi, a single SCSI storage device can appear as a different disk to each host device on a Fibre Channel. Id. at 8:43-45. 2. Introduction of the Bergsten Reference U.S. Patent No. 6,073,209 to Bergsten (Ex. 1007) was filed on March 31, 1997, before the earliest priority date of the 041 Patent (December 31, 1997). Therefore, Bergsten is prior art to the 041 Patent under 35 U.S.C. 102(e). Bergsten was among the hundreds of references before the Examiner but was not discussed by the Examiner in any office action or referenced in any ground of rejection. Like Kikuchi, Bergsten describes serial communication transport mediums, such as Fibre Channels and high speed serial SCSI channels, which connect a host computing device to a storage controller as well as a SCSI transport medium that connects storage subsystems to the storage controller. Ex. 1007 at 4:22-38. As was 25

well known by those of ordinary skill in the art during the relevant timeframe, the Fibre Channel packets as well as the SCSI packets are in a native low level block protocol (NLLBP). Ex. 1010 135, 140-142. The Fibre Channel-to-SCSI bridge feature of Bergsten is enabled by emulation drivers 21 (managing FCP communications between the storage controller and the host device) and physical drivers 22 (managing SCSI communications between the storage controller and storage devices). Ex. 1007 at 7:29-49; see also id. at Fig. 4. The emulations drivers 21 and physical drivers 22 each communicate with a central operating system 20 of the storage controller. Id. An exemplary arrangement is shown in Figure 1 of Bergsten illustrates a series of storage controllers daisy-chained together, where each storage controller is dedicated to service one or more host devices and a particular disk array. See, e.g., Ex. 1010 121, 122. The storage controllers are capable of intercommunicating to store back-up copies of data on each other s disk arrays and to access the back-up copies, if necessary. Id.; see also Ex. 1007 at 3:4-8; 5:23-27. Access by the host devices to the storage subsystems is facilitated by the storage controller, which operates using standard SCSI commands. Ex. 1007 at 4:19-22. Bergsten also teaches that the storage controller virtualizes the remote storage subsystems for each host computing device such that each host computing device can access the remote storage units using virtual addressing, independent of which 26

physical storage device winds up being accessed. Id. at 3:14-17; 4:47-50. The storage controller emulates a local storage array for the host computer system which it services and emulates a host computer system for the local storage array which it accesses. Host computer systems access stored data using virtual device addresses, which are mapped to real device addresses by the storage controller. Id. at 3:14-19. A local host computer accesses data by transmitting a (virtual) host address to its local storage controller. The host address is then mapped to a real address representing a location on one or more physical MSDs[.]... The mapping is completely transparent to all of the host computers.... [I]n the above described mapping process, a single host address may map to multiple physical addresses[]. Id. at 6:10-19. Because the data is replicated across a number of storage arrays serviced by a number of storage controllers, Bergsten employs a tree-style mapping to map logical addresses to physical data locations. Id. at 9:21-23; 3:14-15; 4:47-50; and Fig. 8. 3. The Combined System of Kikuchi and Bergsten A figure depicting the combined system is shown below. In the combined system of Kikuchi and Bergsten, the details related to Kikuchi s Fibre-to-SCSI controller or bridge of are enhanced by incorporating Bergsten s emulation drivers 21 and physical drivers 22. Ex. 1010 123-128. To the extent that Kikuchi fails to explicitly detail the FC-to-SCSI bridging functionality, the details of Bergsten s 27

emulation drivers 21 and physical drivers 22 fill in the gaps. Id. Additionally, the correlation chart and address conversion units of Kikuchi are modified to include the virtual mapping functionality of Bergsten s storage controller. Ex. 1010 124-125. In the resulting system, host devices send NLLBP storage commands along a Fibre Channel link to the storage controller. Id. 124. The Fibreto-SCSI storage controller allows multiple host devices to access different partitions of a single hard drive and also assigns various partitions and drives in the array to 28

various hosts in a manner which is transparent to them. Id. 127, 129, and 144. At the storage controller, the address verification unit of Kikuchi is optionally used to provide host-level access controls, denying any host not registered in the system via the address registration unit access to the storage array. Ex. 1006 at 7:13-27. Alternatively, because a host device can only access those partitions allocated to that particular host, the address verification unit may be removed as being functionally redundant. Ex. 1010 134, 161. The Fibre-to-SCSI storage controller includes the correlation chart which associates each host device with certain logical units to which the host device has access and further associates each logical address to a physical address in the SCSI storage array. Id. 125-127. As explained by Professor Chase, because a single storage controller services requests for multiple hosts in the combined system, the tree-style mapping of Bergsten is rendered unnecessary. Id. 126. Instead, Kikuchi s correlation chart may be enhanced to implement the logical addressing described in Bergsten. Id. An artisan skilled in network storage would combine the Kikuchi and Bergsten systems in this way in order to improve the Kikuchi system with the advantage of virtualized, networked storage. Id. 123-128. As explained in Bergsten, as of early 1997 it was desirable that such a storage controller not be dependent upon any particular hardware or software configuration of any host computer or mass storage device which it services. Ex. 1007 at 1:48-51. Professor Chase explains 29

that a skilled storage engineer would have been motivated to incorporate the virtual storage emulation of Bergsten into the disk apparatus of Kikuchi to increase both the number of storage devices accessible to hosts connecting to the disk apparatus and the storage address range available within the combined system. Ex. 1010 127. The combined system also benefits from increased restructuring capabilities, because an administrator could replace or update equipment and reassign host storage regions without requiring host-side involvement. Id. He further explains that the Bergsten functionality could have been readily added to the architectures as they are generally compatible with one another and the modifications could have been routinely carried out by a person of ordinary skill in the art. See id. 123-128. 4. Correspondence Between Claims 1-53 and Kikuchi and Bergsten The discussion below demonstrates the correspondence between the claim terms and Kikuchi and Bergsten. The foregoing discussion of the level of skill in the art and objective considerations of nonobviousness applies to this ground as well. 1. a) A storage router for providing virtual local storage on remote storage devices, comprising: The Kikuchi/Bergsten system involves a storage router (Kikuchi disk apparatus enhanced by Bergsten s emulation driver interfaces) capable of providing virtual local storage on remote storage devices. See Ex. 1006 at 3:61-67. The storage, for example as described by Bergsten, includes an array of remote 30