DCT

8:20-cv-00699

Parity Networks LLC v. Edgecore USA Corp

Log In
Key Events
Complaint

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 8:20-cv-00699, C.D. Cal., 04/09/2020
  • Venue Allegations: Venue is alleged based on Defendant's transaction of substantial business in the district, including through an office in Irvine, California, and by placing infringing products into the stream of commerce with the expectation of purchase and use by customers in the district. For the Taiwan-based defendant, venue is asserted under 28 U.S.C. § 1391(c)(3) as a foreign resident.
  • Core Dispute: Plaintiff alleges that Defendant’s networking switches infringe six patents related to packet routing, classification, multicasting, and congestion management.
  • Technical Context: The patents address foundational technologies for managing data traffic in high-speed network routers, which are critical components of modern data networks like the Internet.
  • Key Procedural History: The complaint alleges that Defendant received actual notice of the patents-in-suit via letters dated October 5, 2016, and November 28, 2016, which forms the basis for the willfulness allegations.

Case Timeline

Date Event
2000-07-26 Priority Date for U.S. Patent No. 6,553,005
2001-03-06 Priority Date for U.S. Patent Nos. 6,870,844 and 7,719,963
2001-08-22 Priority Date for U.S. Patent Nos. 6,763,394, 7,103,046, and 7,107,352
2003-04-22 U.S. Patent No. 6,553,005 Issued
2004-07-13 U.S. Patent No. 6,763,394 Issued
2005-03-22 U.S. Patent No. 6,870,844 Issued
2006-09-05 U.S. Patent No. 7,103,046 Issued
2006-09-12 U.S. Patent No. 7,107,352 Issued
2010-05-18 U.S. Patent No. 7,719,963 Issued
2016-10-05 Plaintiff allegedly sent first notice letter to Defendant
2016-11-28 Plaintiff allegedly sent second notice letter to Defendant
2020-04-09 Complaint Filed

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 6,553,005 - "Method and Apparatus for Load Apportionment among Physical Interfaces in Data Routers," Issued April 22, 2003

The Invention Explained

  • Problem Addressed: In high-speed routers, a single logical destination (egress interface) may comprise multiple physical ports. The patent addresses the problem of determining which specific physical port to use for a given packet while ensuring that packets belonging to the same "flow" (i.e., same source and destination) are always routed through the same physical path to prevent out-of-order delivery, and while efficiently balancing the traffic load across the available physical ports (’005 Patent, col. 1:21–col. 2:24).
  • The Patented Solution: The invention proposes using a common function, such as a hashing function, that processes packet characteristics (e.g., source and destination addresses) to produce a unique numerical result (’005 Patent, col. 2:35-38). This result is then used to map the packet to a specific physical egress port, ensuring that all packets in the same flow receive the same result and are thus routed to the same port (’005 Patent, col. 5:32-49; Fig. 3).
  • Technical Importance: This approach provides a deterministic, hardware-implementable method for maintaining packet order and performing load balancing, which is critical for applications sensitive to latency and packet order, such as voice and video streaming (’005 Patent, col. 1:43-52).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶38).
  • Claim 1 Elements:
    • Using a common characteristic of each packet to choose a set of egress ports from a table.
    • Processing a source/destination address pair of each packet using a hashing function, producing a binary string result.
    • Using a default subset of the binary string result to select a single egress port from the chosen set, ensuring packets with common source/destination address pairs egress by a common port.

U.S. Patent No. 6,763,394 - "Virtual Egress Packet Classification at Ingress," Issued July 13, 2004

The Invention Explained

  • Problem Addressed: Packet filtering decisions (pass/drop) have traditionally been made at both the ingress (entry) and egress (exit) ports of a router. Performing this function at the egress port is inefficient because the router has already expended resources to process and transport a packet internally, only to drop it at the final step (’394 Patent, col. 2:3-17).
  • The Patented Solution: The invention moves the egress pass/drop decision-making logic to the ingress port. The system at the ingress port consults a rule set (like an Access Control List) that includes not only packet header criteria but also the intended egress port identity (’394 Patent, col. 5:10-20). This allows the router to decide whether a packet would be dropped at its destination egress port before it is ever sent across the internal fabric, saving processing resources (’394 Patent, Fig. 2).
  • Technical Importance: This "virtual egress classification" makes routing more efficient by eliminating the wasted processing of packets that are destined to be dropped, a significant optimization in high-throughput network environments (’394 Patent, col. 5:47-51).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶45).
  • Claim 1 Elements:
    • A system at an ingress port for egress pass/drop determination.
    • A rule set comprising a first lookup table with header combinations, values, and an egress port identity for each.
    • A second lookup table for accomplishing ingress rule determinations without egress port numbers.
    • A mechanism for noting headers and egress port, comparing them in the first and second lookup tables, and returning a rule to be applied.

U.S. Patent No. 6,870,844 - "Apparatus and Methods for Efficient Multicasting of Data Packets," Issued March 22, 2005

  • Technology Synopsis: The patent addresses efficient multicasting (delivering a single packet to multiple destinations) within a router's internal fabric. It proposes a dedicated "multicast-capable port" or "M-Port" that receives a packet once, replicates it the necessary number of times, assigns new destination addresses to the copies, and re-introduces them into the router's fabric for delivery, avoiding inefficient software-based replication (’844 Patent, col. 3:24-46).
  • Asserted Claims: At least independent claim 1 (Compl. ¶52).
  • Accused Features: The complaint alleges that certain Edgecore switches implement multicast protocols like Protocol Independent Multicasting (PIM) and Internet Group Management Protocol (IGMP) in a manner that infringes (Compl. ¶53).

U.S. Patent No. 7,103,046 - "Method and Apparatus for Intelligent Sorting and Process Determination of Data Packets Destined to a Central Processing Unit of a Router or Server on a Data Packet Network," Issued September 5, 2006

  • Technology Synopsis: The patent addresses the problem of a router's Central Processing Unit (CPU) being overwhelmed by packets, particularly during a denial-of-service (DOS) attack. The invention provides a system to sort CPU-destined packets into different priority queues (e.g., trusted, suspect, unknown) before they reach the CPU, allowing the CPU to process high-priority packets while rate-limiting or dropping low-priority ones (’046 Patent, col. 2:40-49).
  • Asserted Claims: At least independent claim 1 (Compl. ¶59).
  • Accused Features: The complaint alleges that certain Edgecore switches include packet processors that categorize packets based on source, place them in a queue, and process them via a CPU based on priority (Compl. ¶60). The complaint points to the switches' support for 802.1X authentication as a means of classifying traffic into categories (Compl. ¶28).

U.S. Patent No. 7,107,352 - "Virtual Egress Packet Classification at Ingress," Issued September 12, 2006

  • Technology Synopsis: This patent is related to the ’394 Patent and further details the method for making egress-style pass/drop decisions at the ingress port. It describes using a lookup mechanism, such as a Content-Addressable Memory (CAM), at the ingress port that checks a packet's header information and intended egress port against a set of rules to determine if the packet should be passed or dropped before it consumes internal router resources (’352 Patent, col. 5:51-65).
  • Asserted Claims: At least independent claim 1 (Compl. ¶66).
  • Accused Features: The complaint alleges that certain Edgecore switches use access control lists to implement egress determinations at ingress ports (Compl. ¶67).

U.S. Patent No. 7,719,963 - "System for Fabric Packet Control," Issued May 18, 2010

  • Technology Synopsis: The patent describes a method for managing traffic congestion within a router fabric that avoids complex "flow control" messaging. It proposes a queuing system at each port that monitors its own queue level and begins to discard incoming packets at a predetermined rate when a preset threshold is reached, and increases the discard rate as the queue becomes fuller, preventing the port from being overwhelmed and propagating congestion upstream (’963 Patent, col.2:37-49).
  • Asserted Claims: At least independent claim 1 (Compl. ¶73).
  • Accused Features: The complaint alleges that certain Edgecore switches implement a Weighted Random Early Detection (WRED) algorithm on packet queues to drop packets as a function of queue size to manage congestion (Compl. ¶74).

III. The Accused Instrumentality

Product Identification

  • The accused instrumentalities are various series of Edgecore networking switches, including the ECS4620, AS5700, AS5800, AS6701, and AS7700 series, among others (Compl. ¶¶ 22, 39, 46). The complaint also references operating systems used on these switches, such as PicOS and Cumulus Linux (Compl. ¶23).

Functionality and Market Context

  • The complaint alleges Edgecore is a global provider of networking products for a wide range of customers, from small businesses to carrier-level service providers (Compl. ¶21). The accused switches are alleged to perform several technical functions relevant to the patents-in-suit:
    • Hashing and Load Balancing: Using a hash function on source/destination addresses to produce a binary string, allegedly for egress port selection (Compl. ¶24).
    • Access Control Lists (ACLs): Implementing ACLs to provide packet filtering and make egress port determinations based on packet criteria like IP addresses, MAC addresses, and protocol type (Compl. ¶25). A table from an Edgecore reference guide illustrates the command groups for configuring these ACLs (Compl. p. 6).
    • Multicasting: Supporting IP multicast forwarding, which involves duplicating packets arriving at a port and forwarding them to other output ports (Compl. ¶¶ 26-27).
    • Quality of Service (QoS) and Prioritization: Supporting 802.1X authentication to classify traffic into categories and process it based on assigned priority (Compl. ¶28). A diagram from a configuration guide shows the Client/Server architecture of this system (Compl. p. 8, Fig. 1). The switches also allegedly use Priority Flow Control (PFC) to manage traffic based on Class of Service (CoS) priorities (Compl. ¶29).
    • Congestion Management: Implementing Weighted Random Early Detection (WRED) to selectively drop packets to manage congestion before queues become full (Compl. ¶30).

IV. Analysis of Infringement Allegations

’005 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
(a) using a common characteristic of each packet, choosing a set of egress ports from a table of sets of ports... The complaint alleges Edgecore switches determine the route and egress port used by particular packets, implying a selection from available ports (Compl. ¶39). ¶39, ¶41 col. 8:21-25
(b) processing a source/destination address pair of each packet using a hashing function... The complaint alleges that in switches like the AS7700 series, "a hash function is used to process a source/destination address" (Compl. ¶24). It further alleges these products use "hashing functions to determine the route and egress port" (Compl. ¶39). ¶24, ¶39, ¶41 col. 8:26-28
...producing thereby a binary string result for each packet processed; The complaint alleges that the "result based upon the hash function is produced as a binary string" (Compl. ¶24). ¶24 col. 8:28-29
(c) using a default subset of the binary string result to select a single egress port for each packet... The complaint does not explicitly allege that a "default subset" of the binary string is used, but states that the hashing function is used to "determine the route and egress port" (Compl. ¶39), which is the ultimate outcome of the claimed step. ¶39, ¶41 col. 8:30-34
...ensuring thereby that packets having common source/destination address pairs egress by a common egress port. The complaint alleges the accused products are used "such that packets with common source/destination address pairs use a common egress port" (Compl. ¶39). ¶39, ¶41 col. 8:34-37

Identified Points of Contention (’005 Patent)

  • Technical Question: Claim 1 requires using a "default subset" of the binary string hash result. The complaint alleges a hash function produces a binary string (Compl. ¶24) and that this is used to select a port (Compl. ¶39), but provides no evidence that a subset of that string is used as required by the claim. The case may turn on whether discovery shows the accused switches perform this specific step.
  • Scope Question: The claim recites "choosing a set of egress ports from a table." The complaint's allegations are general about determining a route and egress port. A question for the court will be whether the accused devices' routing process meets the specific "choosing a set... from a table" limitation as described in the patent.

’394 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
a rule set comprising a first lookup table including header combinations and values... and an egress port identity for each... The complaint alleges Edgecore switches implement Access Control Lists (ACLs) that "provide egress port determinations" (Compl. ¶25) and use this technology to "implement egress determinations at ingress ports" (Compl. ¶46). The ACLs are configured based on header criteria (address, protocol, etc.) (Compl. p. 6). ¶25, ¶46, p. 6 col. 6:13-19
a second lookup table for accomplishing ingress rule determinations without egress port numbers; The complaint does not provide sufficient detail for analysis of this element. It focuses on the egress-related functionality of the ACLs. N/A col. 6:19-21
a mechanism noting pertinent headers... and an egress port... comparing... the headers with rules in the rule set... The functionality of the ACLs, which filter packets based on header information (Compl. p. 6) and are allegedly used for "egress determinations at ingress ports" (Compl. ¶46), is presented as this mechanism. ¶25, ¶46 col. 6:22-26
and returning a determination of a rule to be applied. The ACL function, by its nature, returns a determination to filter (i.e., pass or drop) a packet, which constitutes applying a rule (Compl. ¶25). ¶25 col. 6:26-27

Identified Points of Contention (’394 Patent)

  • Structural Question: Claim 1 structurally requires two distinct lookup tables: one for egress rules (with port identity) and a second for ingress rules (without port identity). The complaint focuses on the single concept of using ACLs for "egress determinations at ingress ports" (Compl. ¶46) but does not allege the existence of the specific two-table structure required. This creates a significant question of whether the accused ACL system meets the claim limitations.
  • Functional Question: The court will need to determine if the accused ACLs, which are generally used for security and traffic filtering, perform the specific function of "virtual egress pass/drop determination" as claimed, or if their operation is technically distinct.

V. Key Claim Terms for Construction

’005 Patent: "default subset of the binary string" (Claim 1)

  • Context and Importance: The infringement theory hinges on whether the accused hashing function is used in the specific manner claimed. The use of a "subset" is a precise implementation detail, and its construction will be critical. A narrow construction could allow the defendant to argue its method, even if it uses hashing, is different.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent's summary describes the core invention more generally as using a "numerical processing result" to select a port, without emphasizing the "subset" aspect, which may support an argument that this is not a crucial limitation (’005 Patent, col. 2:35-42).
    • Evidence for a Narrower Interpretation: The detailed description explicitly teaches using the "least significant two bits" of a 10-bit number as an example of the subset, suggesting the inventors contemplated using only a portion of the full hash result (’005 Patent, col. 5:45-49). Figure 6 also shows a hash result F(x) being used to generate an "Offset," which implies a portion of a larger value, not the entire value itself (’005 Patent, Fig. 6).

’394 Patent: "a second lookup table for accomplishing ingress rule determinations without egress port numbers" (Claim 1)

  • Context and Importance: This term defines a key structural element of the claimed system. Practitioners may focus on this term because if the accused ACL system is found to be a single, unified table or mechanism, it may not infringe this claim, which requires two distinct tables for two different functions (ingress vs. virtual egress).
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent's "Background" section broadly discusses the problem of performing pass/drop functions at both ingress and egress, and the benefit of moving the egress function to the ingress port, potentially allowing an argument that the two-table structure is just one possible implementation of the broader concept (’394 Patent, col. 2:1-17).
    • Evidence for a Narrower Interpretation: Claim 1 explicitly recites "a first lookup table" and "a second lookup table," and claim 19, on which it is based, requires comparing headers in a "first lookup table" and a "second lookup table". The plain language strongly suggests two separate structures are required (’394 Patent, col. 6:14-25; col. 8:11-16). The patent also states the invention can be accomplished by performing ingress and egress lookups "separately" (’394 Patent, col. 5:34-36).

VI. Other Allegations

  • Indirect Infringement: The complaint alleges both induced and contributory infringement for all asserted patents. Inducement is based on allegations that Edgecore instructs its customers, through specifications and user guides, to operate the accused switches in an infringing manner (e.g., Compl. ¶¶ 40, 47). Contributory infringement is based on allegations that the infringing products contain components that are a material part of the inventions, are especially adapted for infringement, and are not staple articles of commerce suitable for non-infringing use (e.g., Compl. ¶¶ 41, 48).
  • Willful Infringement: The complaint alleges willful infringement based on pre-suit knowledge. It claims that notice letters identifying the patents-in-suit were sent to and received by Edgecore on October 5, 2016, and November 28, 2016 (Compl. ¶¶ 34, 79). It is alleged that Defendant continued to infringe despite this actual notice and an "objectively high likelihood" that its actions constituted infringement (Compl. ¶¶ 80-81).

VII. Analyst’s Conclusion: Key Questions for the Case

  • A central issue will be one of "claim scope versus real-world implementation": The complaint maps general-purpose networking features described in Edgecore's marketing and technical documents (e.g., hashing, ACLs, WRED) to specific claim limitations. A key question for the court will be whether the actual, underlying operation of these features in the accused switches meets the detailed structural and functional requirements of the patent claims, such as the '005 Patent's "default subset" requirement or the '394 Patent's "second lookup table" structure.
  • A key evidentiary question will be one of "proof": The complaint relies on high-level, public-facing documentation. The case will likely turn on whether Parity Networks, through discovery of source code and internal design documents, can produce concrete evidence that the accused switches operate in a manner that satisfies every element of the asserted claims, bridging the gap between the general descriptions and the specific claim language.
  • Finally, a critical question for damages will be "willfulness": Given the allegation of pre-suit notice letters dating back to 2016, the court will have to examine Edgecore's conduct since that time. The dispute will likely focus on whether Edgecore investigated the patents in good faith and formed a reasonable, non-frivolous belief that it did not infringe or that the patents were invalid, which will be determinative for the request for enhanced damages.