[Paper Review] Design discussion on the ISDA Common Domain Model

Christopher D. Clack|arXiv (Cornell University)|Nov 29, 2017

Blockchain Technology Applications and Security2 references3 citations

TL;DR

This paper provides a critical analysis of the ISDA Common Domain Model (CDM) draft, focusing on foundational design challenges such as data lineage, timestamps, consistency, operations, events, and state transitions. It argues for rigorous specification of these elements to ensure correctness and interoperability, advocating for technology-agnostic implementation to maximize adoption across centralized and distributed systems.

ABSTRACT

A new initiative from the International Swaps and Derivatives Association (ISDA) aims to establish a "Common Domain Model" (ISDA CDM): a new standard for data and process representation across the full range of derivatives instruments. Design of the ISDA CDM is at an early stage and the draft definition contains considerable complexity. This paper contributes by offering insight, analysis and discussion relating to key topics in the design space such as data lineage, timestamps, consistency, operations, events, state and state transitions.

Motivation & Objective

• To analyze and provide feedback on the ISDA CDM draft definition, which aims to standardize data and process representation across derivatives instruments.
• To identify and discuss key design challenges in foundational concepts such as data lineage, timestamps, consistency, operations, events, and state transitions.
• To advocate for rigorous specification of operations and events to ensure correctness, completeness, and consistency in ISDA CDM implementations.
• To support broad adoption by emphasizing technology-agnostic design, enabling deployment on centralized or distributed platforms including DLT.
• To stimulate debate and feedback to improve the ISDA CDM for future iterations, particularly in aligning benefits with concrete design principles.

Proposed method

• Analyzes the ISDA CDM draft definition using formal and technical scrutiny, focusing on core concepts like data lineage, timestamps, and state transitions.
• Evaluates the semantic and syntactic clarity of event and operation definitions, highlighting inconsistencies and ambiguities.
• Examines the implications of technology choices, particularly the role of distributed ledger technology (DLT), arguing that the CDM should remain agnostic to implementation platforms.
• Assesses the use of 'state' in event definitions, critiquing its misuse to describe dynamic, changing entities rather than static system properties.
• Proposes that operations manipulating multiple contracts require a more rigorous syntax and semantics to ensure correctness and consistency.
• Uses a working paper format to stimulate discussion and feedback, aiming to refine the CDM through iterative improvement.

Experimental results

Research questions

• RQ1 How can data lineage be rigorously defined and implemented in the ISDA CDM to ensure traceability of derived data?
• RQ2 What are the implications of using 'state' to describe events, and how does this affect the semantics of state transitions?
• RQ3 To what extent does the current definition of operations and events support correctness, completeness, and consistency in multi-contract scenarios?
• RQ4 Why is it problematic to tie the ISDA CDM too closely to distributed ledger technology, and how can the model remain technology-agnostic?
• RQ5 How can the benefits listed in the ISDA CDM’s feature set be translated into concrete, implementable design principles?

Key findings

• The ISDA CDM draft lacks rigorous definitions for operations and events, particularly when manipulating multiple contracts, which undermines correctness and consistency.
• The term 'state' is misused in event definitions, as events are dynamic actions that change state, not static representations of it.
• There is no clear specification of value transfer direction, units, or delivery vs. payment in events, creating ambiguity in implementation.
• The current design does not adequately specify how the economic description of a contract should evolve after each event, raising concerns about semantic consistency.
• The claim that the CDM is optimized for DLT is misleading—lineage and consistency can be achieved in non-distributed systems, and the model should remain agnostic to underlying technology.
• The absence of formal semantics for operations and events makes it impossible to verify properties like completeness or correctness, requiring substantial redesign of core components.