`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`___________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`___________________
`
`
`HEWLETT PACKARD ENTERPRISE COMPANY,
`Petitioner
`
`v.
`
`INTELLECTUAL VENTURES II LLC,
`Patent Owner
`_________________________
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`_________________________
`
`
`
`
`PATENT OWNER’S RESPONSE
`
`
`
`
`
`Mail Stop “PATENT BOARD”
`Patent Trial and Appeal Board
`U.S. Patent and Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-145
`
`
`
`
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`TABLE OF CONTENTS
`
`I.
`II.
`
`Introduction ..................................................................................................... 1
`The ’818 Patent ............................................................................................... 2
`A. Overview of the Specification ............................................................... 2
`1. Multi-Tier Hierarchical Quality of Service (QoS)
`Control ....................................................................................... 2
`Classification of a SAN Storage Command .............................. 6
`2.
`The Virtual I/O Server Configuration ...................................... 11
`3.
`B. Overview of the Claims ....................................................................... 15
`III. Level of Ordinary Skill in the Art ................................................................ 18
`IV. Claim Construction ....................................................................................... 18
`V.
`SUMMARY OF THE APPLIED REFERENCES ....................................... 19
`A.
`Srinivasa .............................................................................................. 19
`B.
`Edsall ................................................................................................... 25
`C. Wu ....................................................................................................... 29
`VI. Ground 1 Fails Because Srinivasa is Not Section 102(e) Prior Art. ............ 31
`A.
`Legal Standards ................................................................................... 33
`B.
`The Dynamic Drinkware burden-shifting framework does not
`apply here; Patent Owner does not have a burden of
`production. ........................................................................................... 35
`Petitioner fails to prove that Srinivasa is “by another.” ...................... 37
`The Board should follow Applied Materials; Petitioner’s cited
`cases are readily distinguishable. ........................................................ 39
`VII. Ground 1: Srinivasa Does Not Disclose “Classifying the Storage Command
`Relative to the Hierarchical Token Bucket Resource Allocation to Determine a
`Current Amount of Tokens Available” ................................................................... 43
`VIII. Ground 2: The Combination of Edsall and Wu Does Not Render Obvious The
`Challenged Claims. ................................................................................................. 51
`Petitioner Fails to Consistent Map Edsall’s Disclosure to the
`A.
`Claim Elements ................................................................................... 52
`Edsall Does Not Disclose the Claimed “Virtual Storage Node
`Identifier.” ........................................................................................... 57
`
`C.
`D.
`
`B.
`
`
`
`- i -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`1.
`
`2.
`
`Edsall’s virtual address does not identify a “virtual
`storage network interface layer” located on the
`application server ..................................................................... 58
`Edsall’s virtualization switch does not present the
`“virtual address” to a SAN ....................................................... 61
`Edsall does not disclose a “virtual storage network interface
`layer” of an application server. ........................................................... 62
`Edsall and Wu Do Not Render Obvious “Classifying the
`Storage Command Relative to the Hierarchical Token Bucket
`Resource Allocation to Determine a Current Amount of
`Tokens Available.” .............................................................................. 64
`A POSITA Would Not Have Been Motivated to Combine
`Edsall and Wu. .................................................................................... 66
`Edsall and Wu do not render obvious the other independent
`claims or the dependent claims. .......................................................... 68
`IX. Conclusion .................................................................................................... 70
`
`
`C.
`
`D.
`
`E.
`
`F.
`
`
`
`
`
`- ii -
`
`
`
`
`
`Exhibit No.
`2001
`2002
`2003
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`PATENT OWNER’S UPDATED EXHIBIT LIST
`
`Description
`Declaration of Jacob Sharony, Ph.D.
`
`Curriculum vitae of Jacob Sharony, Ph.D.
`
`Deposition Transcript of Aaron Striegel, Ph.D., August 19, 2022
`
`
`
`- iii -
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`I.
`
`INTRODUCTION
`Petitioner challenges claims 1-3, 6-7, 17-19, 30, 32-34, 37-38, and 40 of
`
`U.S. Patent No. RE44,818 (“the ’818 patent”) on two separate grounds. Both
`
`grounds, however, fail demonstrate that the challenged claims are anticipated or
`
`obvious.
`
`The first ground relies on anticipation in view of U.S. Patent No. 7,782,869
`
`(“Srinivasa”). As further explained below, Srinivasa does not qualify as prior art
`
`because it is not “by another” as required by pre-AIA §102(e). Petitioner does not
`
`dispute that Srinivasa’s sole inventor is the same inventor listed on the ’818 patent.
`
`Based on the relevant case law, Petitioner has therefore failed to meet its burden
`
`that Srinivasa qualifies as prior art.
`
`But even beyond that issue, Srinivasa does not anticipate the claims because
`
`it does not perform the required claim limitation of “classifying the storage
`
`command” as recited in the independent claims. The content of the ’818 patent
`
`distinguishes itself from Srinivasa and highlights this active classification as
`
`important to its quality of service process. Srinivasa does not disclose this
`
`“classifying.”
`
`Petitioner’s second ground based on Edsall and Wu suffers similar flaws.
`
`Regarding Edsall, Edsall does not disclose the same networking configuration
`
`process recited in the “maintaining” and “presenting” steps recited in the claims.
`
`
`
`- 1 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`Specifically, the claims require an apparatus such as a virtual I/O server to present
`
`a virtual storage node identifier corresponding to an application server’s “virtual
`
`storage network interface layer” to a “storage area network” (SAN). Edsall’s
`
`system does not present such an identifier but rather converts a virtual address to a
`
`physical address and instead presents that physical address to a SAN. This is
`
`because Edsall does not teach any application server having a virtual storage
`
`network interface layer.
`
`Additionally, like Srinivasa, Wu also does not perform a classification of the
`
`storage command as recited in the claims. Wu does not classify any specific
`
`command but rather groups all commands from a particular client together.
`
`Further, a person of ordinary skill in the art (“POSITA”) would not have been
`
`motivated to combine Edsall and Wu in view of their differing contextual
`
`environments.
`
`In view of these deficiencies, Petitioner has not met its burden to show that
`
`the challenged claims of the ’818 patent are unpatentable.
`
`II. THE ’818 PATENT
`A. Overview of the Specification
`1. Multi-Tier Hierarchical Quality of Service (QoS) Control
`The ’818 patent provides a novel and non-obvious “multi-tier” quality of
`
`service (QoS) that “allows virtual I/O servers to be scalable and provide
`
`appropriate QoS granularity.” EX1001, Abstract; EX2001, ¶¶31-38. To provide
`
`- 2 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`this fine level of control, the ’818 patent recites a “two-tier hierarchical QoS
`
`management process” that is executed at a virtual I/O server. EX1001, 2:10-18,
`
`8:20-44. The “first hierarchical QoS process” is performed on “aggregated virtual
`
`I/O subsystem traffic.” Id. For example, this first process determines whether
`
`received traffic is destined for a storage area network (SAN) or a local area
`
`network (LAN). EX1001, FIG. 1, 2:66-3:36. The ’818 patent explains that these
`
`two types of traffic result in different conditions and bandwidth requirements:
`
`In applications such as system backup files transfer to a
`
`SAN device, the application’s demand for bandwidth is
`
`usually high, has relaxed latency requirements, and
`
`occurs infrequently. In applications such as Internet
`
`Protocol telephony application, accesses to LAN I/O
`
`subsystems use little bandwidth, but require very low
`
`latency.
`
`EX1001, 3:30-18 (emphasis added).
`
`After performing this first hierarchical QoS process and performing an
`
`initial classification of a received command, the ’818 patent proceeds to the second
`
`classification process that provides finer grain QoS control. EX1001, 1:66-2:9,
`
`13:5-30. Specifically, “a second hierarchical QoS process is performed on each
`
`group for further classification.” EX1001, 2:10-2:18. “This hierarchical technique
`
`
`
`- 3 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`divides the QoS process into sub-processes, providing the flexibility to scale and
`
`fine tune the granularity of QoS as necessary without affecting other sub-
`
`processes.” EX1001, 8:20-39; see also id., 11:25-36. Regarding a storage
`
`command directed to a SAN, the ’818 patent explains:
`
`The second hierarchical QoS process provides much
`
`finer grain QoS control to the virtual I/O server. For
`
`communication group destined for SAN I/O subsystems,
`
`SAN QoS manager 424 can allocated QoS on different
`
`SAN commands such as read. These SAN read
`
`commands dominate the bandwidth usage of the SAN
`
`I/O subsystems as they involve the transfer of larger data
`
`size, while other commands utilize negligible
`
`bandwidth. SAN QoS manager 424 can emphasize finer
`
`QoS control over read commands, and can effectively
`
`ignore other SAN commands.
`
`EX1001, 13:21-30 (emphasis added).
`
`In this manner, the ’818 patent describes a second classification
`
`corresponding to the particular type of SAN command received. EX2001, ¶¶36-38.
`
`With this classification, the ’818 patent executes a corresponding action based on
`
`
`
`- 4 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`the bandwidth requirements of that particular type of SAN command. Id. For
`
`example, the ’818 patent continues in explaining that:
`
`The bulk of the bandwidth usage for such SAN I/O
`
`subsystems is related to read and write operations, with
`
`other operations such as setup and management
`
`constituting a very small percentage of bandwidth
`
`usage. In the present invention, SAN QoS manager 424
`
`is used to further classify read commands, allowing
`
`other less bandwidth intensive virtual I/O
`
`communications to proceed directly to the SAN I/O
`
`subsystems.
`
`EX1001, 14:12-21 (emphasis added).
`
`As seen from these passages, the ’818 patent provides fine-grain control
`
`over SAN commands. EX1001, 1:66-2:9, 11:25-36. For example, the ’818 patent
`
`describes discerning between read, write, setup, and management commands,
`
`which are all storage commands directed to a SAN. EX1001, 14:12-21. As
`
`Petitioner’s expert has admitted, there are many types of storage commands—
`
`being less resource intensive than read or write commands—that would be
`
`prioritized and immediately passed for execution. EX2003, 55:7-58:11, 67:3-9,
`
`96:13-98:21.
`
`
`
`- 5 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`Classifying the storage command allows for fine-grain control—such as
`
`allowing less bandwidth intensive storage commands to proceed directly to the
`
`SAN—while performing additional scheduling and traffic shaping processes on
`
`more bandwidth intensive storage commands. EX1001, 10:21-29, 14:12-21;
`
`EX2001, ¶¶33-38. For example, the ’818 patent uses a hierarchal token bucket
`
`(HTB) to manage traffic as well. EX1001, 9:61-10:53. This fine-grain control
`
`results in a better QoS that is not anticipated or otherwise rendered obvious by any
`
`of the art cited in the petition. EX2001, ¶¶33-38.
`
`2.
`Classification of a SAN Storage Command
`To illustrate the multi-tier classification of SAN Storage Command, the ’818
`
`patent provides several figures and example use cases. EX2001, ¶¶39-46. For
`
`example, “FIG. 4 is a component diagram showing the two-tier hierarchical
`
`components of the virtual I/O server QoS. This two-tier QoS management process
`
`is hierarchical and scalable, separating the virtual I/O traffic QoS management
`
`process into smaller sub-processes.” EX1001, 8:40-44.
`
`
`
`- 6 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`EX1001, FIG. 4 (annotated).
`
`
`
`As seen in Figure 4, “fabric receive process 416” performs the first tier of
`
`classification to determine whether received traffic or a received command is
`
`directed to a SAN or a LAN. EX1001, 9:50-65.
`
`The first hierarchical QoS process is performed by
`
`fabric receive QoS manager 414 along with fabric
`
`receive process 416 and fabric receive buffer 412. After
`
`
`
`- 7 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`virtual I/O communications are classified and separated
`
`into either SAN or LAN I/O subsystems groups, SAN
`
`I/O subsystems group virtual I/O communications are
`
`forwarded to SAN receive process 426 and LAN I/O
`
`subsystems group virtual I/O communications are
`
`forwarded to LAN receive process 430.
`
`EX1001, 9:11-18 (emphasis added).
`
`Subsequent to classifying the virtual communication as SAN or LAN traffic,
`
`“SAN QoS Manager 424” performs the second level of classification: “The second
`
`hierarchical QoS process is performed by SAN QoS manager 424 along with SAN
`
`receive buffer 428 on SAN I/O subsystems group.” EX1001, 9:18-21 (emphasis
`
`added). “The second hierarchical QoS process imposes further QoS classification
`
`on each I/O subsystems destination groups.” Id., 13:6-11. Depending on the results
`
`of this second hierarchical QoS process, SAN QoS manager 424 forwards storage
`
`commands to SAN Receive Process 426, which then provides the storage
`
`command to a SAN. Id., 13:6-11.
`
`The ’818 patent also depicts its application of a second hierarchical QoS
`
`process to a SAN storage command in Figures 5 and 7. Id., FIG. 5, FIG. 7, 11:38-
`
`12:28, 13:32-14:29.
`
`
`
`- 8 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`
`
`EX1001, FIG. 5 (annotated).
`
`EX1001, FIG. 7 (annotated).
`
`
`
`As seen in Figure 5, at step 516 “fabric receive QoS manager 414 separates
`
`the virtual I/O communication into either SAN or LAN I/O subsystems groups
`
`based on the I/O subsystems destination type.” EX1001, 11:67-12:3. At step 518,
`
`“virtual I/O communications that are SAN write commands are further evaluated.”
`
`Id., 12:4-5. If the system determines that the storage command is in fact a write
`
`command, the process proceeds to 522 where the system determines whether there
`
`are sufficient tokens to process the data transfer size of the write command. Id.,
`
`12:5-7. Otherwise, storage commands that are not write commands are directly
`
`forwarded to SAN receive process 426 for execution at the SAN. Id., 12:24-26.
`
`
`
`- 9 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`
`
`Similarly, Figure 7 depicts a process based on whether the storage command
`
`is classified as a read command. Id., 13:47-51, 13:66-2:3. Specifically, at “step
`
`706, the virtual I/O communication is analyzed to determine if it is a SAN read
`
`command.” Id., 13:47-49. “If it is a SAN read command, in step 712, the
`
`associated data size transfer of the SAN read command is determined.” Id., 13:49-
`
`51. In contrast, if the SAN storage command is not classified as a read command,
`
`no data size determination is needed. Id., 13:66-14:3.The system proceeds with
`
`executing the command when sufficient tokens are available. Id., 14:7-11. As seen
`
`from Figures 5 and 7, when the storage command is not a read or write command,
`
`the additional step of checking a data transfer size is not needed. EX2001, ¶¶42-46.
`
`In this manner, Figures 5 and 7 illustrate the second hierarchical QoS
`
`process and the ’818 patent’s fine-grain control. EX1001, 13:21-30. By classifying
`
`and determining the type of storage command that is received, the ’818 patent can
`
`exercise fine-grain control over bandwidth-intensive operations while being able to
`
`“effectively ignore other SAN commands” and “allowing other less bandwidth
`
`intensive virtual I/O communications to proceed directly to the SAN.” EX1001,
`
`13:21-30 (emphasis added). This classification achieves improved QoS by
`
`prioritizing less bandwidth-intensive SAN commands and allowing those to be
`
`executed directly. EX1001, 14:12-21; see also EX2003, 55:7-58:11, 67:3-9;
`
`EX2001, ¶46. As further discussed below, the references asserted in the Petition do
`
`
`
`- 10 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`not disclose this type of classification to achieve fine-grain QoS control. EX2001,
`
`¶46.
`
`3.
`The Virtual I/O Server Configuration
`The ’818 patent describes executing the multi-tier classification at “virtual
`
`I/O server 106” as depicted in Figure 1. EX1001, FIG. 1, 2:66-3:36; EX2001,
`
`¶¶47-54.
`
`
`
`- 11 -
`
`
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`EX1001, FIG. 1 (annotated).
`
`Virtual I/O server 106 communicates with application servers 102 via I/O
`
`switch fabric 104. EX1001, 2:66-3:36, 7:58-67. “Virtual I/O server 106 provides
`
`the storage and external networking needs of application servers 102 connected to
`
`I/O switch fabric 104, allowing transparent, shared access to SAN I/O subsystems
`
`114.” Id., 3:15-18. The ’818 patent explains that “virtual I/O server 106 enables
`
`application servers 102 to read a remote physical storage device target within SAN
`
`I/O subsystems 114 as if it is physically attached.” Id., 15:49-52.
`
`When an application server 102 seeks to issue a storage command to SAN
`
`114 (e.g., read, write, setup, management), the application server 102 uses a locally
`
`executed virtual interface referred to as virtual host bus adapter (HBA) 208a. Id.,
`
`4:63-5:15, 5:56-6:2, 15:49-59. Virtual HBA 208a is depicted in Figure 2. Id.
`
`“[V]irtual HBA layer 208a is operative to establish a connection with virtual block
`
`interface of virtual I/O server 106 to forward commands or other messages.” Id.,
`
`5:10-15, 5:56-60.
`
`
`
`
`
`
`
`- 12 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`EX1001, FIG. 2 (annotated).
`
`
`
`“At the application server 102 where a SAN read command is initiated by a
`
`given application, the Virtual HBA 208a intercepts the SAN read command and
`
`the SAN read command is encapsulated with an identifying header in
`
`encapsulation module 206. The encapsulated SAN read command passes through
`
`the I/O fabric PHY interface 202 to a virtual I/O server 106 over the I/O switch
`
`fabric 104 for further processing.” EX1001, 15:52-59. In this manner, the
`
`
`
`- 13 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`application server 102 uses virtual HBA 208a to transmit commands to virtual I/O
`
`server 106. EX2001, ¶52.
`
`As part of this command transmission, application server 102 also transmits
`
`a unique identification corresponding to virtual HBA 208a. EX1001, 7:28-67,
`
`15:52-59; see also EX2003, 41:20-42:11, 46:20-48:5, 49:18-51:17, 121:12-21
`
`(Petitioner’s expert explaining that a virtual I/O server 106 would need this
`
`identification to return data to the particular application server 102 or virtual HBA
`
`208a).
`
`“The virtual HBA layer 208a is assigned one or more virtual World Wide
`
`Names (WWNs).” EX1001, 7:30-36. The same virtual WWN that is assigned to
`
`the virtual HBA layer 208 on the application server 102 is also exposed or
`
`presented to SAN 114. Id., 7:32-36. Specifically, “virtual I/O server 106 exposes
`
`these virtual WWN on SAN I/O subsystems 114 using N-Port Identifier
`
`Virtualization (NPIV) functionality.” Id., 7:32-36. By virtue of being a “World
`
`Wide Name,” application server 102, the virtual I/O server 106, and the SAN 114
`
`are all aware and utilize the same virtual WWN to reference command and
`
`communications from virtual HBA 208a. Id.; EX2001, ¶¶53-54. The virtual I/O
`
`server 106 need not perform any conversion and instead presents the virtual WWN
`
`for an application server 102 directly to a SAN 114 to perform operations.
`
`EX1001, 7:32-36; EX2001, ¶¶53-54. The virtual I/O server 106 subsequently uses
`
`
`
`- 14 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`this virtual WWN when returning response messages or data back to the virtual
`
`HBA 208a providing the command. Id., 7:57-58.
`
`B. Overview of the Claims
`The claims reflect the aspects of the ’818 patent described above. EX2001,
`
`¶¶55-60. For example, claim elements [1.1] and [1.2] recite a “virtual storage node
`
`identifier.” EX1001, 17:45-50.
`
`• [1.1] maintaining a connection, over a network fabric, to a virtual storage
`
`network interface layer of an application server, wherein the virtual
`
`storage network interface layer is associated with a virtual storage
`
`node identifier;
`
`• [1.2] presenting, at a physical storage network interface, the virtual
`
`storage node identifier to a storage area network;
`
`Claim elements [1.1] and [1.2] therefore recite a “virtual storage node
`
`identifier” that identifies a “virtual storage network interface layer” residing on
`
`application server 102. EX1001, 17:45-50. This same “virtual storage node
`
`identifier” is subsequently is presented to a storage area network (e.g., SAN 114).
`
`Id.; EX2003, 40:1-41:3. The antecedent basis of this claim term is clear and
`
`indicates that they are the same identifier. EX1001, 17:45-50; EX2001, ¶¶56-57.
`
`In the context of the specification, an example of the “virtual storage node
`
`identifier” is the “virtual WWN.” EX1001, 7:28-67. The virtual I/O server 106
`
`
`
`- 15 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`uses the virtual WWN to facilitate and process storage commands for application
`
`server 102. Id. Specifically, the same virtual WWN corresponding to the
`
`application server’s virtual HBA 208 is presented to the SAN 114. EX1001, 7:30-
`
`36. Rather than using different names or identifiers, “virtual I/O server 106
`
`exposes these virtual WWN on SAN I/O subsystems 114.” Id., 7:32-36. As
`
`confirmed by Petitioner’s expert, a POSITA would have viewed both [1.1] and
`
`[1.2] as being executed by an intermediary system positioned between application
`
`server 102 and SAN 114. EX2003, 25:5-26:20 (Petitioner’s expert agreeing that
`
`the “maintaining” and “presenting” steps would be performed by virtual I/O server
`
`106). Virtual I/O sever 106 is such a system. See EX1001, 7:28-67; EX2001, ¶¶56-
`
`57.
`
`Additionally, the claims also capture the multi-tier hierarchical QoS control
`
`and classification of a storage command described in the specification. EX1001,
`
`2:10-18, 8:20-44, 13:21-30, 14:12-21. For example, claim 1 also recites:
`
`• [1.4] receiving, over the connection, a storage command from the
`
`virtual storage network interface layer of the application server,
`
`wherein the storage command is a command to read data from, or write
`
`data to, a target connected to the storage area network;
`
`
`
`- 16 -
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`• [1.5] classifying the storage command relative to the hierarchical token
`
`
`
`bucket resource allocation to determine a current amount of tokens
`
`available;
`
`As seen from claim element [1.4], a storage command is first received over
`
`the claimed connection. EX1001, 17:54-58, 17:61-63. For example, this may occur
`
`when virtual I/O server 106 receives a storage command from the virtual HBA
`
`208a of an application server 102. Id., 7:30-36, 7:49-57. Subsequently, claim
`
`element [1.5] refers to “classifying the storage command” to determine an amount
`
`of available tokens. Id., 17:61-63. This active classification step refers to
`
`determining the particular type of storage command received as explained above.
`
`Id., 13:21-30, 14:12-21; EX2001, ¶¶59-60. For example, this classification
`
`discerns whether the storage command is a read or write command versus another
`
`low-bandwidth command, such as a setup or management command. EX1001,
`
`13:21-30, 14:12-21; EX2001, ¶¶59-60. This provides the additional fine-grain QoS
`
`process described in the specification. EX1001, 1:66-2:9, 10:21-29, 11:25-36.
`
`Thus, the ’818 patent’s inventive techniques are recited in the claims. As
`
`discussed in detail below, none of the Petition’s presented grounds anticipate or
`
`render obvious these features of claim 1 or the other independent claims.
`
`
`
`- 17 -
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`III. LEVEL OF ORDINARY SKILL IN THE ART
`
`
`
`A person of ordinary skill in the art (“POSITA”) would have a bachelor’s
`
`degree in electrical engineering, computer engineering, computer science, or an
`
`equivalent field as well as two or more years of relevant industry experience,
`
`including experience in distributed computer systems involving virtualization as
`
`well as quality of service (“QoS”), including class-based scheduling mechanisms
`
`such as hierarchical token bucket (“HTB”). EX2001, ¶¶28-30.
`
`IV. CLAIM CONSTRUCTION
`In inter partes review proceedings, a claim of a patent is construed using the
`
`same standard used in federal district court. See 37 C.F.R. § 42.100(b) (2018).
`
`Under that standard, claim terms are construed in accordance with their ordinary
`
`and customary meaning, as understood by one of ordinary skill in the art, and in
`
`view of the patent specification and the relevant prosecution history. Phillips v.
`
`AWH Corp., 415 F.3d 1303, 1312-1313 (Fed. Cir. 2005) (en banc). Terms that may
`
`appear clear on their face nevertheless cannot be viewed “in a vacuum” but should
`
`be based on “the ordinary meaning in the context of the written description and the
`
`prosecution history.” Medrad, Inc., v. MRI Devices Corp., 401 F.3d 1313, 1319
`
`(Fed. Cir. 2005) (quoting DeMarini Sports, Inc. v. Worth, 239 F.3d 1314, 1324
`
`(Fed. Cir. 2001)). The specification thus plays a key role in this analysis:
`
`“[i]mportantly, the person of ordinary skill in the art is deemed to read the claim
`
`
`
`- 18 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`term not only in the context of the particular claim in which the disputed term
`
`appears, but in the context of the entire patent, including the specification.”
`
`Phillips, 415 F.3d at 1313. Indeed, the specification is the “single best guide” to
`
`construing claim terms. Id. at 1315.
`
`In the Institution Decision, the Board stated that “we determine that no claim
`
`terms require express construction.” DI, 13. Patent Owner agrees that no claim
`
`terms require construction.
`
`V.
`
`SUMMARY OF THE APPLIED REFERENCES
`A.
`Srinivasa
`Srinivasa is titled “Network Traffic Control For Virtual Device Interfaces.”
`
`EX1006, Title. Srinivasa discloses a QoS scheduling module that receives a packet
`
`from a virtual device interface and applies an HTB allocation scheme to allocate
`
`bandwidth. Id., Abstract; EX2001, ¶¶63-72. In Srinivasa’s HTB, ports are roots,
`
`and the virtual device interfaces are leaves. EX1006, Abstract.
`
`Figure 3 of Srinivasa “illustrates the protocol stack and modules of a virtual
`
`I/O server 60 according to one possible implementation of the invention.” Id.,
`
`4:29-31. “[V]irtual block interface 326 [in red below], in one particular
`
`implementation may receive SCSI commands from an application server 102 over
`
`the I/O fabric, and make calls into the SCSI layer of storage driver stack 322 [in
`
`orange below].” Id., 4:57-6. “In another implementation, virtual block interface
`
`
`
`- 19 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`326 receives generic read, write and/or other commands that are translated into
`
`appropriate protocols.” EX1006, 4:61-65.
`
`EX1006, FIG. 3 (annotated).
`
`
`
`“In one implementation, the multiplexer 308 [in purple above] is a kernel
`
`module which transfers incoming requests from the application servers 102 to the
`
`appropriate native HBA” and may also “be configured to provide . . . [QoS] to the
`
`client application and servers.” EX1006, 3:67-4:15. Management console 399 (in
`
`blue above) allows an operator to configure one or more virtual block devices
`
`
`
`- 20 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`and/or virtual network interfaces for an application server, including the portion of
`
`a virtual or physical Host Bus Adapter (“HBA”) available to a particular
`
`application server. Id., 5:42-45, 6:1-6.
`
`Figure 8A illustrates an “alternative implementation of the virtual I/O
`
`server.” Id., 7:54-56.
`
`EX1006, FIG. 8A (annotated).
`
`
`
` “In the implementation illustrated, the virtual I/O server 60 omits
`
`multiplexer 308.” EX1006, 7:56-57. “In yet other implementations, the
`
`encapsulation functions of [QoS] module 306 can be incorporated into virtual
`
`block interface module 326 and virtual network interface 346 of virtual I/O server
`
`60 . . . .” Id., 7:57-63. “In such an implementation, the corresponding virtual
`
`
`
`- 21 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`interface of application server 102 and virtual I/O server 60 can communicate
`
`directly over the I/O fabric using separate logical fabric connections . . . .” Id.,
`
`7:63-8:1.
`
`“As some of the Figures illustrate, a bandwidth management or QoS
`
`mechanism can be included in the application servers 102 and the virtual I/O
`
`servers 60 to control bandwidth utilization across the I/O switch fabric. For
`
`example, FIG. 8A illustrates an encapsulation and QoS layer 306 that can be
`
`utilized to enforce bandwidth allocations to different I/O traffic types. FIG. 8B
`
`illustrates a similar encapsulation and QoS layer 206 that controls bandwidth
`
`utilization at an application server 102.” Id., 16:32-40.
`
`Figure 9B illustrates a “hierarchical configuration where the child leaf nodes
`
`608 [in green below] of the storage node 606 represent virtual Host Bus Adapters
`
`(vHBAs).” Id., 16:63-66. “Network [in red below] and storage [in blue below]
`
`nodes 604, 606, representing network and storage I/O subsystem traffic,
`
`respectively, are appended to nodes 602. Leaf nodes 608 represent the I/O
`
`subsystem traffic associated with each virtual interface, including storage and
`
`networking traffic.” Id., 16:59-63. “As shown in FIGS. 9A and 9B, packet queues
`
`are associated with leaf nodes, though tokens might be distributed throughout the
`
`inverted hierarchy, as will be explained further below. In addition, one or more of
`
`the application servers 102 may also include a second I/O fabric interface, in which
`
`
`
`- 22 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`case the hierarchical configuration illustrated in FIG. 9A, as well as the bandwidth
`
`control functionality described herein, can be replicated for this second interface.”
`
`Id., 16:66-17:6.
`
`EX1006, FIG. 9B (annotated).
`
`
`
`In this class-based HTB approach, “[e]ach class or node has a bucket of
`
`tokens associated with it.” EX1006, 17:52-53. “[A] network administrator may
`
`then allocate bandwidth among the application servers, I/O traffic types and virtual
`
`devices/interfaces by configuring the nodes of the hierarchy with a bandwidth
`
`allocation.” Id., 17:21-25. “In one implementation, each class has a guaranteed
`
`rate, a maximum rate, an actual or observed rate, and a priority level.” Id., 17:58-
`
`
`
`- 23 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`60. “High priority classes might borrow excess resource allocation (such as
`
`bandwidth) from low priority classes.” Id., 17:60-62.
`
`EX1006, FIG. 12 (annotated).
`
`
`
`Figure 12 illustrates a process employing such QoS “according to one
`
`possible implementation of the invention.” EX1006, 18:59-60. This process “may
`
`
`
`- 24 -
`
`
`
`
`
`
`
`Case IPR2022-00096
`U.S. Patent No. RE44,818
`
`result in writing the packet to a virtual device queue or forwarding the packet to the
`
`output port of the I/O fabric interface.” Id., 18:63-66.
`
`In 1004 (in red above), the process accesses the HTB to determine whether
`
`the packet can be sent. Id.,