DCT

1:25-cv-01229

SAP Se v. Celonis Se

Log In
Key Events
Amended Complaint
complaint Intelligence

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 1:25-cv-01229-MN, D. Del., 09/09/2026
  • Venue Allegations: Venue is alleged to be proper in the District of Delaware because Defendant Celonis, Inc. is a Delaware corporation that resides in the district, and Defendant Celonis SE is a foreign corporation.
  • Core Dispute: Plaintiff alleges that Defendant’s process mining software, including the Celonis Platform, PQL Engine, and Object Link features, infringes four patents related to database graph traversal, business intelligence tool connectivity, and dynamic data storage architecture.
  • Technical Context: The lawsuit concerns enterprise software for process mining, which analyzes business process data from systems like Enterprise Resource Planning (ERP) to discover and improve operational workflows.
  • Key Procedural History: Plaintiff SAP and Defendant Celonis had a business relationship from 2016 to 2021, under which SAP resold Celonis's process mining software. Plaintiff alleges that during this relationship, Celonis gained a deep technical understanding of SAP’s database architecture and data extraction techniques.

Case Timeline

Date Event
2012-12-13 ’492 and ’254 Patents Priority Date
2015-03-31 ’492 Patent Issued
2015-01-01 SAP launches S/4HANA ERP software
2016-01-01 SAP and Celonis begin business relationship
2017-07-04 ’254 Patent Issued
2018-11-13 ’034 Patent Priority Date
2020-10-14 Earliest alleged release date of accused Celonis Platform infringing the ’034 Patent
2021-01-01 SAP/Celonis business relationship terminated
2021-06-24 ’007 Patent Priority Date
2021-09-14 ’034 Patent Issued
2022-11-29 ’007 Patent Issued
2023-10-01 Earliest alleged release date of accused Celonis Object Link infringing the ’254 and ’492 Patents
2023-10-01 Earliest alleged release date of accused Celonis Object-Centric process mining system infringing the ’007 Patent
2025-10-03 Date from which SAP seeks damages for ’254 and ’492 Patent infringement
2026-09-09 Complaint Filing Date

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

U.S. Patent No. 9,697,254 - “Graph Traversal Operator Inside a Column Store,” Issued July 4, 2017

The Invention Explained

  • Problem Addressed: The patent addresses the inefficiency of conventional computer graph traversal techniques when applied to very large graph datasets stored in modern, memory-centric database systems (Compl. ¶26; ’254 Patent, col. 1:38-47).
  • The Patented Solution: The invention proposes an optimized graph traversal engine that uses a specialized in-memory data structure, a column-oriented "edge table," combined with traversal operations that use "path predicates" to pre-filter edges and limit the search scope. This allows the system to restrict searches to relevant portions of the data and enables parallel processing of subgraphs on multi-core processors, improving performance (Compl. ¶26; ’254 Patent, col. 9:65-10:3).
  • Technical Importance: This approach aimed to improve computer performance for complex graph queries by enhancing memory efficiency and minimizing redundant processing in large-scale, memory-resident columnar databases (Compl. ¶27).

Key Claims at a Glance

  • The complaint asserts independent claim 1 (a method) and independent claim 8 (a system), along with dependent claims 12 and 13 (Compl. ¶49; Compl. ¶80).
  • Independent Claim 1 (Method) Essential Elements:
    • Receiving an input parameter with a set of path step descriptions.
    • Recursively traversing a column-oriented edge table in a memory-centric database.
    • This traversal further comprises: selecting a path step description, extracting a path predicate, evaluating the predicate to generate valid edges and reduce a scan range, determining a scanning direction, and generating a set of traversed vertices.
    • Returning the set of traversed vertices.
  • The complaint also asserts dependent claims 12 and 13, which add limitations related to an optimization module for parallel traversal of subgraphs and topological ordering (Compl. ¶80).

U.S. Patent No. 8,996,492 - “Graph Traversal Operator Inside a Column Store,” Issued March 31, 2015

The Invention Explained

  • Problem Addressed: The patent identifies the inefficiency of traversing large graph datasets stored in memory, particularly where unnecessary memory retrievals consume significant memory bandwidth in systems using columnar data structures stored in cache memory (Compl. ¶29; ’492 Patent, col. 1:31-39).
  • The Patented Solution: The invention describes storing the graph’s edges in a column-oriented "edge table" within a cache memory. Traversal operations use "path predicates," constraints, and termination conditions to restrict searches to relevant portions of the edge table. This focused approach is designed to improve cache locality, reduce memory bandwidth consumption, and enable parallel processing on multi-core processors (Compl. ¶¶29-30; ’492 Patent, col. 9:47-52).
  • Technical Importance: This design sought to create an optimized graph traversal engine specifically for columnar databases stored in cache memory, improving computer performance by reducing computation and memory overhead during complex queries (Compl. ¶30).

Key Claims at a Glance

  • The complaint asserts independent claim 1 (a method) and independent claim 8 (a system), along with dependent claims 12 and 13 (Compl. ¶136; Compl. ¶171).
  • Independent Claim 1 (Method) Essential Elements:
    • Receiving an input parameter with path step descriptions and a set of start vertices.
    • Recursively traversing a column-oriented edge table stored in a cache memory.
    • This traversal further comprises: identifying a path predicate, generating valid edges to reduce a scan range, and determining a scanning direction.
    • Generating traversed vertices by scanning until a terminating condition is met.
    • Terminating the generation of vertices when the condition is met.
    • Returning the set of traversed vertices.
  • The complaint also asserts dependent claims 12 and 13, which add limitations related to an optimization module for parallel traversal of subgraphs and topological ordering (Compl. ¶171).

U.S. Patent No. 11,120,034 - “Automatic Routing of Connection Requests in Business Intelligence (BI) Tools,” Issued September 14, 2021

Technology Synopsis

The ’034 Patent describes a method for centrally managing data access for business intelligence (BI) systems (Compl. ¶32). The invention aims to solve problems of "thick-client" complexity, where BI tools are bundled with numerous database drivers, by using a middleware architecture. In this architecture, the BI tool connects to a central BI platform, which then provides the necessary database drivers on demand, thereby reducing client-side storage, improving security, and simplifying maintenance (Compl. ¶¶33-34; ’034 Patent, col. 3:28-4:22).

  • Asserted Claims: Claim 9 (Compl. ¶236).
  • Accused Features: The complaint accuses the Celonis Platform user interface and the Celonis Execution Management System (EMS) platform of infringement, alleging they perform a data integration workflow that connects to data sources and retrieves data connections in a manner that practices the claimed method (Compl. ¶224; Compl. ¶230; Compl. ¶242).

U.S. Patent No. 11,514,007 - “Dynamic Data Processing For a Semantic Data Storage Architecture,” Issued November 29, 2022

Technology Synopsis

The ’007 Patent claims a method for storing and maintaining both raw and curated data in a semantic data store (Compl. ¶36). The invention uses a dual-storage architecture with a "raw storage portion" and a "curated storage portion" (Compl. ¶37; ’007 Patent, col. 1:33-55). It describes a real-time processing pipeline that receives event-based data, stores it in raw form, updates a "semantic index," automatically transforms the data's schema into a normalized form, and stores the enriched result in the curated portion, enabling system resilience and low-latency analytics (Compl. ¶39).

  • Asserted Claims: Claim 8 (Compl. ¶269).
  • Accused Features: The complaint accuses the Celonis Object-Centric process mining system, which implements the Object-Centric Data Model (OCDM), of infringement. It is alleged that this system stores data in data pools, transforms it, and aligns it with semantic constructs in a way that infringes the claimed method (Compl. ¶262; Compl. ¶270; Compl. ¶279).

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are components of the Celonis software platform, primarily the "Celonis PQL Engine," the "Object Link" feature with its "LINK PATH" operator, the "Celonis Platform" user interface for data integration, and the "Celonis Object-Centric process mining system" which uses an "Object-Centric Data Model (OCDM)" (Compl. ¶42; Compl. ¶130; Compl. ¶224; Compl. ¶262).

Functionality and Market Context

  • The Celonis platform provides process mining software that analyzes log data from enterprise systems (like SAP's ERP systems) to discover, validate, and improve business processes (Compl. ¶14).
  • The "Celonis PQL Engine" is described as an "analytical column-store main memory database system" that executes queries written in Celonis's domain-specific Process Query Language (PQL) (Compl. ¶52; Compl. ¶64).
  • The "Object Link" feature is alleged to be a framework for modeling and analyzing flows between object instances as a graph, using "dedicated mapping tables" to build the graph structure. The "LINK PATH" operator is used to traverse and analyze this graph (Compl. ¶43; Compl. ¶53). The complaint includes a diagram from Celonis's documentation showing how mapping tables connect different data tables to form a graph structure. (Compl. ¶63).
  • The "Celonis Platform" user interface is alleged to provide a data integration workflow for creating data pools, connecting to data sources, and managing the extraction and transformation of data (Compl. ¶¶224-225). A data flow diagram from Celonis's documentation illustrates this process. (Compl. ¶80).

IV. Analysis of Infringement Allegations

’254 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
A computer-implemented method for improved graph traversal in a memory-centric database, comprising: The Celonis Object Link method is alleged to be a computer-implemented method used in the Celonis PQL Engine, a memory-centric database. ¶52 col. 13:14-17
receiving, by at least one processor, an input parameter, wherein the input parameter includes a set of path step descriptions; The Celonis LINK_PATH operator, executed by the PQL Engine on at least one processor, allegedly receives input parameters (e.g., "direction", "CONSTRAINED BY" clause) that serve as path step descriptions. ¶¶55-58 col. 13:18-20
recursively traversing an edge table for a data graph, wherein the edge table is a column oriented table in the memory-centric database... The Celonis PQL Engine, an analytical column-store main memory database, allegedly recursively traverses "mapping tables" that function as a column-oriented edge table to build and analyze a data graph. A diagram illustrates this configuration of mapping tables. ¶¶59-64; ¶63 col. 13:21-24
selecting a path step description from the set of path step descriptions; The PQL Engine allegedly selects one or more arguments from the LINK_PATH operator syntax (e.g., "direction", "CONSTRAINED BY") to be processed. ¶¶66-67 col. 13:25-26
extracting a path predicate from the selected path step description; The "CONSTRAINED BY" clause in the LINK_PATH operator is alleged to be a path predicate that is extracted from the path step description. ¶¶68-70 col. 13:27-28
evaluating the extracted path predicate against the edge table, wherein the evaluation generates a set of valid edges, and the set of valid edges reduces a scan range during traversing; The constraints in the "CONSTRAINED BY" clause are allegedly evaluated against the objects/edges in the graph, generating a set of valid edges and limiting the traversal to "critical areas," thereby reducing the scan range. ¶¶71-73 col. 13:29-32
determining a direction for scanning the set of valid edges based on the selected path step description; The "direction" parameter of the LINK_PATH operator allegedly specifies the traversal direction as "FORWARDS" or "BACKWARDS". ¶¶74-75 col. 13:33-35
generating the set of traversed vertices by scanning the set of valid edges using the determined direction; and The PQL Engine allegedly generates a set of traversed vertices by scanning the valid edges in the direction determined by the "direction" parameter. ¶¶76-77 col. 13:36-37
returning the set of traversed vertices... The LINK_PATH operator allegedly returns a table of traversal activity, representing the "individual paths calculated by traversing the Object Link graph," which constitutes the set of traversed vertices. A diagram illustrates the generated tables. ¶¶78-79; ¶79 col. 13:38-40

’492 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
A computer-implemented method comprising: receiving an input parameter, wherein the input parameter includes at least a set of path step descriptions that set criteria for a data graph traversal and a set of start vertices; The Celonis LINK_PATH operator allegedly receives input parameters such as the "START" argument (for start vertices) and other clauses ("CONSTRAINED BY") that set criteria for the traversal. ¶¶143-146 col. 12:59-62
recursively traversing an edge table in a data graph that is stored in a column oriented table in a cache memory, based on the input parameter... The Celonis PQL Engine allegedly traverses "mapping tables," which are considered a column-oriented table in cache memory. The complaint cites a paper on the PQL Engine that states it achieves "good cache locality." ¶¶147-150; ¶149 col. 12:63-67
identifying a path predicate included in a path step description; The "CONSTRAINED BY" clause or the "ALL" expression within it is alleged to be a path predicate identified from the path step description. ¶¶153-156 col. 13:1-2
generating a set of valid edges based on the path predicate, wherein the set of valid edges reduces a scan range during the traversing; The "ALL" constraint within the "CONSTRAINED BY" clause allegedly limits valid edges based on the specified predicate, which generates a set of valid edges and reduces the scan range. ¶¶157-158 col. 13:3-5
determining a direction for scanning the set of valid edges... wherein the direction determines whether the scan begins at a vertex source column or a vertex target column in the edge table; The "direction" parameter ("FORWARDS" or "BACKWARDS") of the LINK_PATH operator allegedly determines the traversal direction, which in turn determines whether the scan begins at a source or target object in the edge table. ¶¶159-162 col. 13:6-10
generating the set of traversed vertices by scanning the set of valid edges against a set of vertices... until at least one terminating condition is met...; and The PQL Engine allegedly scans valid edges against a set of vertices (including start vertices) until a terminating condition, such as the "END" constraint, is met. ¶¶163-165 col. 13:11-16
terminating the generating of the set of traversed vertices when the at least one terminating condition is met; and The traversal allegedly terminates when the "END" condition is met (i.e., the end object is reached). ¶¶166-168 col. 13:17-19
returning the set of traversed vertices. The output of the LINK_PATH operator is alleged to be a column containing all objects of all calculated paths, which constitutes the returned set of traversed vertices. ¶¶169-170 col. 13:20-20

Identified Points of Contention

  • Scope Questions: A central issue for both the ’254 and ’492 patents will be whether Celonis’s "CONSTRAINED BY" clause, described as a way to "limit the result of the operator" (Compl. ¶72), functions as the claimed "path predicate" used to "pre-filter edges" (Compl. ¶26). The analysis may turn on whether Celonis's constraints are evaluated in the same manner and achieve the same scan-range reduction as required by the claims.
  • Technical Questions (’492 Patent): A key question for the ’492 Patent is whether the accused system stores its "edge table" in a "cache memory" as the claim requires. The complaint supports this by referencing a paper stating the Celonis PQL Engine achieves "good cache locality" (Compl. ¶149). The dispute may focus on whether achieving "good cache locality" is equivalent to storing the entire edge table in a "cache memory" for the purpose of the claimed traversal.
  • Functional Equivalence: The court may need to determine if the "mapping tables" described in Celonis's documentation (Compl. ¶63) are structurally and functionally equivalent to the claimed "column oriented edge table."

V. Key Claim Terms for Construction

’254 and ’492 Patents

  • The Term: "path predicate"
  • Context and Importance: This term is at the core of the infringement allegations for the traversal patents. The complaint alleges that Celonis's "CONSTRAINED BY" clause and its arguments (e.g., "ALL") function as this claimed element (Compl. ¶69; Compl. ¶154). The case may turn on whether Celonis's user-defined "constraints" for filtering results are equivalent to the claimed "predicate" used for pre-filtering edges and reducing scan ranges during traversal.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specifications state the predicate is an "arbitrary complex relational expression" (’254 Patent, col. 6:1-2), which may support construing the term to cover a wide range of logical conditions, potentially including Celonis's constraints.
    • Evidence for a Narrower Interpretation: The patents describe the predicate's function as evaluating against the edge table to "generate a set of valid edges" that "reduces a scan range during traversing" (’254 Patent, cl. 1). This functional language may support a narrower construction tied specifically to an operational step of pre-filtering the set of edges to be scanned, rather than a general-purpose result filter.

’492 Patent

  • The Term: "cache memory"
  • Context and Importance: This term distinguishes the ’492 Patent from the more general "memory-centric database" of the ’254 Patent. The complaint alleges infringement by pointing to evidence that the Celonis PQL engine uses cache to achieve "good cache locality" (Compl. ¶149; Compl. ¶182). Practitioners may focus on this term because the viability of the infringement claim depends on whether this alleged functionality meets the claim's requirement of storing the "edge table in a data graph that is stored in a column oriented table in a cache memory" (’492 Patent, cl. 1).
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specification does not appear to provide an explicit, limiting definition of "cache memory," which could leave the term open to its plain and ordinary meaning, potentially encompassing systems optimized for high cache utilization.
    • Evidence for a Narrower Interpretation: The patent repeatedly emphasizes improvements in "cache locality" and reduction of "memory-bandwidth waste" by storing the graph's edges in cache (Compl. ¶29; Compl. ¶30). This context may support an interpretation requiring the primary traversal structure (the edge table) to reside substantially or entirely within a CPU cache, as opposed to merely being accessed in a cache-friendly manner from main memory.

VI. Other Allegations

  • Indirect Infringement: The complaint alleges induced infringement for all four patents. The allegations are based on Celonis providing documentation, training through the "Celonis Academy," and user interfaces that allegedly instruct and encourage customers to use the accused features (e.g., Object Link, data integration workflows) in an infringing manner (Compl. ¶47; Compl. ¶53; Compl. ¶54; Compl. ¶134; Compl. ¶232; Compl. ¶265).
  • Willful Infringement: The complaint alleges that the Defendants' infringement has been willful, but the factual basis is tied to knowledge of the patents acquired "at least since the service of the present Complaint" (Compl. ¶47; Compl. ¶134; Compl. ¶232; Compl. ¶265). This frames the willfulness claim as being based on post-suit conduct.

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

  1. Definitional Scope (Graph Traversal): A primary issue for the ’254 and ’492 patents will be one of claim construction: does the "CONSTRAINED BY" clause in Celonis’s "LINK_PATH" operator, which filters traversal results, meet the functional and structural requirements of the claimed "path predicate", which is described as reducing the scan range by generating a set of valid edges during traversal?

  2. Architectural Equivalence (Cache vs. Memory): A key technical question for the ’492 patent will be whether the accused Celonis PQL Engine, described as a "main memory database" that achieves "good cache locality," practices the specific limitation of storing its "edge table" in "cache memory". The court's interpretation of this term will be critical to determining infringement.

  3. System Functionality (BI and Data Storage): For the ’034 and ’007 patents, the case will likely focus on whether the accused Celonis systems perform the specific methods claimed. This raises two evidentiary questions: (1) Does the Celonis EMS platform function as the claimed middleware-based "BI platform" that automatically connects and serves "database coupling components"? (2) Does the Celonis Object-Centric Data Model implement the claimed method of transforming data from a "raw storage portion" to a "curated storage portion" based on a "semantic index"?