EMC Corp v. PersonalWeb Technologies LLC

1. Case Identification

• Case #: IPR2013-00082
• Patent #: 5,978,791
• Filed: December 15, 2012
• Petitioner(s): EMC Corporation and VMware, Inc.
• Patent Owner(s): PersonalWeb Technologies LLC
• Challenged Claims: 1-4, 29-33, and 41

2. Patent Overview

• Title: Data Processing System Using Substantially Unique Identifiers to Identify Data Items, Whereby Identical Data Items Have the Same Identifiers
• Brief Description: The ’791 patent describes a data storage system that uses "substantially unique identifiers" to identify and access data items. These identifiers are generated based on all of, and only, the data within a data item, enabling functions like eliminating duplicate files independent of their location or filename.

3. Grounds for Unpatentability

Ground 1: Anticipation by Browne - Claims 1-4, 29-33, and 41 are anticipated by Browne under 35 U.S.C. §102.

• Prior Art Relied Upon: Browne (S. Browne et al., “Location-Independent Naming for Virtual Distributed Software Repositories,” University of Tennessee Technical Report CS-95-278 (Feb. 1995)).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Browne discloses every limitation of the challenged claims. Browne describes a Bulk File Distribution (BFD) system that addresses the shortcomings of location-dependent file identifiers (like URLs) by creating a "Location Independent File Name" (LIFN). This LIFN is a substantially unique identifier computed as the MD5 hash of the entire contents of a file, thus depending on "all of" and "only" the data in the data item, as required by claim 1. Browne further teaches an "existence means" through its LIFN server, which a client can query with a file's LIFN signature to determine if the file is present on the network and, if so, receive a list of locations where it can be accessed. This mechanism is analogous to the ’791 patent’s "True File Registry." Browne’s caching system also discloses a "local existence means" for determining if a file is present locally.

Ground 2: Anticipation by Langer - Claims 1-4, 29-33, and 41 are anticipated by Langer under §102.

• Prior Art Relied Upon: Langer (Albert Langer, "Re: dl/describe (File descriptions)," post to the "alt.sources" newsgroup (Aug. 7, 1991)).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner contended that Langer, a newsgroup post predating the patent by several years, taught a virtually identical solution to the same problem. Langer proposed using a cryptographic hash function, specifically MD5, on the "entire contents of the file" to create a unique, location-independent identifier. This identifier, based on all of and only the file's content, satisfies the "identity means" of claim 1. Langer taught using these MD5 hashes with central database servers (like the Archie and WAIS search engines) to associate the content-based identifiers with physical file locations. A user could query these search engines with an MD5 hash to determine if and where a file was present in the system, thereby providing the claimed "existence means" for determining if a particular data item is present.

Ground 3: Obviousness over Browne in view of Langer - Claims 1-4, 29-33, and 41 are obvious over Browne in combination with Langer under §103.

• Prior Art Relied Upon: Browne (a 1995 technical report) and Langer (a 1991 newsgroup post).

• Core Argument for this Ground:

  • Prior Art Mapping: Petitioner argued that to the extent Browne’s full LIFN syntax ("lifn:netlib:") was found to include protocol and server identifiers beyond the content-based hash, a person of ordinary skill in the art (POSITA) would have found it obvious to modify Browne’s system.
  • Motivation to Combine: Langer explicitly taught using an identifier consisting only of the MD5 hash to uniquely identify files. A POSITA would have been motivated to apply Langer's simpler identifier to Browne's system to simplify file access in scenarios where Browne's more complex syntax (for handling multiple protocols or servers) was not needed. This modification would yield the predictable result of a more streamlined system.
  • Expectation of Success: The combination involved applying a known technique (Langer's MD5-only identifier) to a known system (Browne's BFD package) to achieve a predictable improvement in simplicity, leading to a high expectation of success.

• Additional Grounds: Petitioner asserted numerous additional anticipation and obviousness challenges. These included standalone anticipation grounds based on Kantor (a 1993 software manual describing a "contents-signature" system using CRC hash and file length to find duplicates on bulletin board systems) and Woodhill (Patent 5,649,196, describing a backup system using content-derived "Binary Object Identifiers"). Further obviousness grounds were asserted combining these references, such as modifying Langer to include a file length value as taught by Woodhill or Kantor to reduce hash collisions.

4. Key Claim Construction Positions

• Petitioner argued that several key claim terms are means-plus-function limitations under 35 U.S.C. §112, ¶ 6, and proposed constructions based on the specification.
• "identity means for determining... a substantially unique identifier" (claim 1): Petitioner identified the function as determining a unique identifier based on all of and only the data in a data item. The corresponding structure was identified as the processor of Fig. 1(b) programmed to execute the "Calculate True Name" mechanism using an MD4, MD5, or SHA hash function.
• "existence means for determining whether a particular data item is present" (claim 1): Petitioner identified the function as determining if a data item is present by examining identifiers. The corresponding structure was identified as the processor of Fig. 1(b) storing the True File Registry and programmed to execute the "Locate Remote File" mechanism.
• "substantially unique identifier": Petitioner relied on a prior court construction defining this term as "an identity for a data item generated by processing all of the data in the data item, and only the data in the data item, through an algorithm."

5. Key Technical Contentions (Beyond Claim Construction)

• Petitioner’s central technical contention was that the ’791 patent’s premise of novelty was fundamentally flawed. The patent asserted that prior art systems always identified data items relative to a specific context (e.g., a pathname) with "no direct relationship between the data names and the data item." Petitioner argued this was incorrect, contending that the concept of creating a content-based "fingerprint" or "signature" for a data item using mathematical hash functions (like MD5) was a widely used and well-understood technique for decades prior to the invention, rendering the patent’s core concept obvious.

6. Relief Requested

• Petitioner requested institution of an inter partes review and cancellation of claims 1-4, 29-33, and 41 of the ’791 patent as unpatentable under 35 U.S.C. §§ 102 and 103.