PTAB

IPR2019-00992

Cisco Systems Inc v. NetFuel Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Managing Computer Network Resources
  • Brief Description: The ’659 patent discloses a system for automated management of computer networks using software agents. These agents operate within runtime environments to monitor network parameters, detect abnormalities, and apply corrective policies in a hierarchical structure that may include local, regional, and global modelers.

3. Grounds for Unpatentability

Ground 1: Obviousness over Lindskog in view of Turek - Claims 1-3, 7-9, 13-15 are obvious over Lindskog in view of Turek.

  • Prior Art Relied Upon: Lindskog (Patent 6,665,262) and Turek (Patent 6,460,070). Petitioner noted that Lindskog incorporates by reference Lindskog II (Patent 6,370,572) and Lindskog III (Patent 6,349,325), treating their disclosures as part of Lindskog.
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Lindskog teaches the core elements of the challenged claims, including a hierarchical, agent-based system for distributed fault management in a communications network. In Lindskog, fault agents monitor network resources, detect faults, and escalate requests for corrective action (a "corrective policy") to higher-level agents if a fault cannot be handled locally. Petitioner contended that to the extent Lindskog does not explicitly teach "threads" and a "runtime environment," Turek remedies this. Turek discloses using software agents within a multi-threaded runtime process to diagnose and correct network faults. Turek explicitly teaches a "multi-threaded runtime environment" and using software agents written in Java, a well-known multi-threaded language. For dependent claims, Petitioner asserted Lindskog II teaches load balancing operations and Lindskog III teaches comparing monitored parameters to known thresholds to detect abnormalities.
    • Motivation to Combine: A POSITA would combine Lindskog's hierarchical fault management system with Turek's multi-threaded agent architecture to achieve a more robust, efficient, and scalable network management system. Petitioner argued this combination would achieve predictable results, such as conserving network bandwidth (a goal of both references) and improving processor utilization through multitasking. Implementing Lindskog’s network resources as threads would allow for uninterrupted monitoring while other threads are engaged in corrective actions.
    • Expectation of Success: A POSITA would have had a high expectation of success, as the combination involved applying a well-known technique (multi-threading from Turek) to a known system architecture (agent-based fault management from Lindskog) to achieve the predictable benefits of improved performance and robustness.

Ground 2: Obviousness over Lindskog, Turek, and Dijkstra - Claims 6 and 18 are obvious over Lindskog in view of Turek and Dijkstra.

  • Prior Art Relied Upon: Lindskog (Patent 6,665,262), Turek (Patent 6,460,070), and Dijkstra ("Self-stabilizing systems in spite of distributed control," Communications of the ACM, Nov. 1974).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds upon the combination of Lindskog and Turek from Ground 1 to address the specific limitation in claims 6 and 18 requiring that "creating the corrective policy comprises using Dijkstra's Self-Stabilization Algorithm." Petitioner argued that Lindskog's system, which implements corrective policies until a network reconfiguration is fully implemented, is inherently self-stabilizing. Dijkstra provides the explicit teaching of a well-known, foundational algorithm for achieving self-stabilization in distributed networks. The algorithm allows a system to recover from an illegitimate state through local actions, which is directly analogous to the claimed creation of a "corrective policy" to resolve a network abnormality.
    • Motivation to Combine: A POSITA would have been motivated to incorporate the well-known Dijkstra algorithm into the combined Lindskog/Turek system. The goal of all the references is to create robust, self-correcting distributed networks. Using Dijkstra's algorithm would have been a natural and obvious way to implement the corrective policies in the agent-based system to enhance its stability and ensure it could automatically recover from faults, a key objective of network management.
    • Expectation of Success: Petitioner asserted a high expectation of success because Dijkstra's algorithm was specifically designed for and widely known to be applicable to stabilizing distributed networks of the type disclosed by Lindskog and Turek.

4. Key Claim Construction Positions

  • "global modeler": Petitioner argued for the construction "a module that enforces policies upon subordinate modelers." This construction was adopted by the district court in co-pending litigation and is based on explicit language in the ’659 patent's specification. Petitioner contended this construction is necessary to give meaning to the word "global" and to distinguish it from other "modelers" in the patent's hierarchy, contrasting it with the Patent Owner's allegedly overbroad construction.
  • "first making a request... if the corrective policy is not available...": Petitioner argued this phrase requires a conditional, two-step process: (1) it is first determined that the corrective policy is not available to the local agent, and (2) then a request is made to an external entity. Petitioner asserted this sequence is mandated by the plain language ("if") and consistently described in the specification and figures. This construction is critical to mapping the hierarchical escalation taught in the prior art to the claim language.

5. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1-3, 6-9, 13-15, and 18 of the ’659 patent as unpatentable.