PTAB

IPR2017-01734

Cavium Inc v. Alacritech Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Network Interface Device That Fast-Path Processes Solicited Session Layer Read Commands
  • Brief Description: The ’205 patent describes a system for accelerating network data transfers. The invention uses a network interface device to perform "fast-path" processing on responses to solicited session layer read commands, thereby bypassing the host computer's conventional protocol stack and reducing CPU load.

3. Grounds for Unpatentability

Ground 1: Obviousness over Thia and SMB - Claims 1, 4, 5, 8, 11, and 13 are obvious over Thia in view of SMB.

  • Prior Art Relied Upon: Thia ("A Reduced Operation Protocol Engine (ROPE) for a multiple-layer bypass architecture," 1995) and SMB ("Protocols for X/Open PC Interworking: SMB, Version 2," a 1992 technical standard).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Thia disclosed the core architecture of the challenged claims. Thia taught a "multiple-layer bypass architecture" for the OSI protocol model, featuring a host computer with a standard protocol stack (SPS or "slow-path") and a network interface device with a specialized "Bypass Stack" ("fast-path") on a Reduced Operation Protocol Engine (ROPE) chip. Thia’s network interface device performed a "bypass test" on incoming packets; packets passing the test were processed on the fast-path, bypassing the host’s network and transport layer processing, while packets failing the test were routed to the host’s slow-path SPS for conventional processing. Petitioner asserted that SMB, a widely used session layer file-sharing protocol compatible with the OSI model, provided the claimed "solicited read command" (e.g., an "SMBread" request). The combination thus allegedly taught an apparatus (claim 1) where a network interface device performs fast-path processing on a response to a session layer read command, placing data into host memory without using the host computer's network or transport layers. Claim 8's recitation of both fast-path and slow-path processing was argued to map directly onto Thia's dual-path architecture.
    • Motivation to Combine: Petitioner contended a POSITA would combine Thia and SMB to implement a practical, high-performance file-sharing system. Thia provided a high-speed OSI-compatible bypass architecture that required a session layer protocol to be useful. SMB was a well-known and natural choice for this purpose. Combining them would predictably add valuable file-sharing functionality to Thia while simultaneously improving the performance of SMB communications by offloading significant processing to the network device.
    • Expectation of Success: A POSITA would have a high expectation of success, as both references operated within the OSI model framework. Thia explicitly stated its design could be adapted to existing OSI stack implementations with minor software modifications, providing an easy migration path for protocols like SMB.

Ground 2: Obviousness over Thia, SMB, and Carmichael - Claims 6 and 7 are obvious over Thia in view of SMB and Carmichael.

  • Prior Art Relied Upon: Thia (a 1995 article), SMB (a 1992 technical standard), and Carmichael (Patent 5,894,560).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground built upon the Thia/SMB combination from Ground 1 and added Carmichael to address specific data handling limitations in dependent claims 6 and 7. Petitioner asserted that a file transfer response would naturally comprise multiple packets. For claim 6's requirement that data from first and second packets be placed in memory "in a substantially contiguous manner," Petitioner pointed to Carmichael. Carmichael taught efficient Direct Memory Access (DMA) transfers using physical region descriptor (PRD) tables to transfer single, contiguous blocks of data to host memory. For claim 7's limitation that the first data is placed into destination memory "before the second packet is received," Petitioner again relied on Carmichael, which disclosed that blocks of data were placed into destination memory sequentially.
    • Motivation to Combine: Petitioner argued that a POSITA, having combined Thia and SMB, would be motivated to incorporate Carmichael's teachings to improve the efficiency of the system's DMA operations. Thia, SMB, and Carmichael were all directed at improving computer system performance and were compatible with common operating systems like UNIX. Adding Carmichael’s method for efficient, contiguous block data transfers via DMA was argued to be a predictable way to further enhance the performance of the Thia/SMB file-sharing system at a minimal cost.
    • Expectation of Success: The combination was asserted to be a predictable aggregation of known elements. A POSITA would expect that applying Carmichael’s improved DMA techniques to the data transfers in the established Thia/SMB architecture would result in a faster, more efficient system without altering the fundamental fast-path/slow-path operation.

4. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1, 4-8, 11, and 13 of the ’205 patent as unpatentable.