Bright Data Ltd v. Metacluster LT UAB

1. Case Identification

• Case #: IPR2021-00983
• Patent #: 10,601,948
• Filed: May 21, 2021
• Petitioner(s): Bright Data Ltd.
• Patent Owner(s): Metacluster LT, UAB
• Challenged Claims: 1-20

2. Patent Overview

• Title: Smart Proxy Rotator
• Brief Description: The ’948 patent discloses a smart proxy rotator (SPR) system designed to manage and balance the use of multiple proxies for a client's web scraper. The system dynamically configures utilization thresholds for proxy providers by calculating weights based on performance and availability feedback from the web scraper, preventing proxies from being overused or failing.

3. Grounds for Unpatentability

Ground 1: Anticipation of Claims 1-9 and 17-20 by Smith

• Prior Art Relied Upon: Smith (Application # 2017/0155623).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Smith discloses every limitation of the challenged independent claims (1, 17) and their dependents. Smith’s system for "selecting proxies" teaches a computer-implemented method for dynamically configuring a utilization threshold, which corresponds to Smith's disclosure of associating and revising "weights" (expressed as percentages) with proxy subsets. Smith’s "proxy selection engine" running on a computing device was mapped to the claimed smart proxy rotator (SPR). The petition asserted that Smith’s disclosure of determining proxy "availability" and "success labels" and revising weights based on this data meets the claimed limitations of a web scraper providing performance information to the SPR, which in turn calculates weights and configures thresholds.

Ground 2: Obviousness of Claims 10-16 over Smith in view of Raja

• Prior Art Relied Upon: Smith (Application # 2017/0155623) and Raja (Application # 2008/0225710).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner contended that Smith teaches most limitations of independent claim 10, including a non-transitory medium with instructions for dynamically configuring a proxy provider's utilization threshold and accumulating performance/availability information. The key limitation allegedly not explicitly taught by Smith is rendering content unreadable to the SPR via network or application layer encryption. Raja was introduced to supply this element, as it discloses a load-balancing appliance that uses an encryption engine and tunneling protocol to create a virtual private network, thereby rendering content unreadable to an intermediary device like the SPR.
  • Motivation to Combine: Petitioner argued that both Smith and Raja address the problem of load balancing. A person of ordinary skill in the art (POSITA) would have been motivated to incorporate Raja’s well-known encryption and tunneling techniques into Smith’s proxy selection system. This combination would predictably improve security and prevent the intermediary proxy management system from reading sensitive content, which was a known goal in network communications.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success in applying Raja's established encryption methods to Smith's system, as it involved applying a known technique to a similar system to yield predictable security benefits.

Ground 3: Obviousness of Claims 1-3, 7, 10, 12, 14, 17, and 18 over Shribman in view of Raja

• Prior Art Relied Upon: Shribman (Patent 9,241,044) and Raja (Application # 2008/0225710).

• Core Argument for this Ground:

  • Prior Art Mapping: Petitioner argued that Shribman discloses a system analogous to the ’948 patent, teaching an "acceleration server" (the SPR) that manages and selects from a list of "tunnel devices" (proxies) for a client device (the web scraper) to fetch content from a web server. Shribman further discloses that the client prepares and sends a "transaction log" containing performance data (e.g., bandwidth, round-trip time) to the acceleration server. The selection of tunnel devices can be based on this past performance. While Shribman teaches selecting devices based on performance attributes, it does not explicitly teach calculating numerical "weights." Raja was introduced to teach this specific limitation, as it explicitly discloses calculating loads for servers using "weights and thresholds" for load balancing.
  • Motivation to Combine: Petitioner asserted that since Shribman’s system involves selecting devices based on load balancing principles, a POSITA would have been motivated to look to other load balancing solutions like Raja to improve Shribman’s selection mechanism. A POSITA would combine Raja’s explicit teaching of using weights with Shribman's performance-based selection system to create a more robust and quantifiable method for distributing requests among the tunnel devices.
  • Expectation of Success: A POSITA would have reasonably expected success in this combination, as incorporating a known weighting technique (from Raja) into a performance-based selection system (Shribman) was a well-understood and predictable way to enhance load balancing algorithms.

• Additional Grounds: Petitioner asserted anticipation of claims 10 and 14 by Shribman alone and obviousness of claims 1-3, 7, 10, 12, 14, 17, and 18 over Shribman alone.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §314(a) or §325(d) was not warranted. The petition asserted that the ’948 patent had not been challenged in any prior IPR. Furthermore, it argued that no prior art rejections were made during the patent's prosecution, and the prior art and arguments presented in the petition had not been previously considered by the USPTO.

5. Relief Requested

• Petitioner requests the institution of an inter partes review and the cancellation of claims 1-20 of the ’948 patent as unpatentable.