Technical Summary: Toward a Common Understanding of Cryptographic Agility — A Systematic Review

Problem Statement

Cryptographic agility has emerged as a critical concept in both academic and industrial discourse, driven by the necessity to adapt cryptographic systems to evolving threats, including side-channel attacks, cryptanalysis advancements, and the impending transition to post-quantum cryptography (PQC). Despite its recognized importance, the term suffers from conceptual ambiguity. The literature lacks a consensus definition, often conflating agility with related notions such as "cryptographic versatility" and "cryptographic interoperability." Definitions vary widely, ranging from abstract design paradigms to specific engineering practices or system properties. This inconsistency hinders effective research and practical implementation, leading to fragmented strategies for cryptographic migration and potential security risks during transitions.

Methodology

The authors conducted the first systematic literature review of cryptographic agility, adhering to the PRISMA 2020 guidelines.

• Search Strategy: The study queried four sources: IEEE Xplore, ACM Digital Library, Google Scholar, and the Google search engine. Search queries included variations of "cryptographic agility" and "crypto-agility."
• Selection Criteria: From an initial pool of 84 sources, 48 were selected for analysis based on strict inclusion and exclusion criteria.
  • Inclusion: Sources must explicitly provide a definition of cryptographic agility (either original or composite) and be available in English. The corpus included 24 peer-reviewed papers and 24 gray literature sources (e.g., technical reports, white papers, RFCs, theses).
  • Exclusion: Duplicates, older versions, inaccessible sources, non-English texts, and sources that merely quoted another definition without offering their own. Videos, podcasts, and webinars were excluded.
• Data Analysis: The authors employed an iterative classification process, mapping extracted aspects of the definitions into six distinct categories: Context, Mode, Desired Capabilities, Quality Attributes, Cryptographic Entities, and Drivers. This mapping was performed to identify shared characteristics and synthesize a unified understanding.

Key Contributions and Results

1. Classification of Existing Definitions (RQ1)

The review revealed a significant lack of consensus. While all definitions relate to facilitating cryptographic migration, they diverge on scope and nature. The authors categorized the literature into six dimensions:

• Context: Ranges from abstract (hardware/software) to specific (IoT, cloud, specific devices).
• Mode: Described variously as an engineering practice, an approach, an objective, or a property.
• Desired Capabilities: Primarily the ability to set up, identify, and modify cryptographic entities.
• Quality Attributes: Often cited as flexible, efficient, and business-continuous.
• Cryptographic Entities: Includes keying material, encryption methods, parameters, and libraries.
• Drivers: Motivations such as quantum threats, algorithm obsolescence, or regulatory compliance.

2. Literature-Based Definition (RQ2)

Synthesizing the recurring characteristics, the authors propose a Literature-Based Definition:

"Cryptographic Agility is a theoretical or practical approach, objective, or property which provides capabilities for setting up, identifying, and modifying encryption methods and keying material in a flexible and efficient way while preserving business continuity."

3. Context-Independent Definition (RQ3)

To establish a precise conceptual foundation, the authors refined the literature-based definition by removing context-dependent elements (such as specific quality attributes like "efficiency" or "flexibility," which vary by scenario) and isolating the core property. This yields the Context-Independent Definition:

"Cryptographic Agility represents the changeability of cryptographic entities."

This definition frames agility not as a goal to be achieved once (like a migration) but as an intrinsic property of a system that enables the addition, removal, enabling, disabling, and replacement of cryptographic entities.

4. Layer Model and Domain Clarification (RQ4)

The paper introduces a Layer Model to map where agility manifests, organizing contexts into four overarching domains:

• Conceptual Domain: Primitives and algorithms (theoretical layer).
• Software Domain: Libraries, applications, platforms, and operating systems.
• Hardware Domain: Devices and infrastructure (e.g., HSMs, TPMs).
• Organizational Domain: Policies, standards, and governance.

Crucially, the paper clarifies the boundaries between agility and neighboring concepts:

• Cryptographic Versatility: Defined as the ability of hardware to support multiple schemes by design (a static capability), distinct from agility which implies the ability to change or update those schemes over time.
• Cryptographic Interoperability: Defined as the ability of a system to communicate securely using a variety of mechanisms simultaneously. The authors argue that agility facilitates interoperability (by allowing systems to adapt to peers), but interoperability does not imply agility. A system can be interoperable with many peers using a fixed set of algorithms without being agile (unable to change those algorithms).

5. Practical Application

The authors validated their framework by analyzing three real-world software exemplars:

• OpenSSL: Demonstrates agility via its Provider architecture (v3.0+), allowing algorithm substitution via the EVP API without recompiling applications. Constraints include legacy low-level API usage and lack of centralized policy enforcement.
• NGINX: Realizes agility through runtime configuration and compile-time library linking. It is constrained by its dependency on the underlying cryptographic library and the need for recompilation when switching backends.
• GitLab CI/CD: Illustrates agility at the platform level through pipeline extensibility and orchestration of internal (NGINX, GitLab Shell) and external cryptographic tools. Constraints arise from the diversity of orchestrated environments and dependencies on external components.

Significance and Claims

The paper claims to establish the first unified terminology and conceptual foundation for cryptographic agility. Its primary significance lies in:

1. Resolving Ambiguity: By distinguishing agility from versatility and interoperability, the work prevents terminological conflation that has previously obscured research and practice.
2. Providing a Baseline: The context-independent definition ("changeability of cryptographic entities") offers a precise, measurable property for future research, distinct from specific migration objectives like PQC transition.
3. Guiding Practice: The layer model helps practitioners identify where agility is realized and where it is constrained (e.g., hardware lifecycle vs. software updates).
4. Enabling Assessment: The authors argue that agility cannot be universally quantified by a single metric (like a maturity score) but must be assessed relative to specific contexts and requirements.

The paper concludes that while cryptographic agility is essential for sustainable security, it introduces engineering complexity and trade-offs. It is not a panacea but a necessary property that enables systems to adapt securely to future threats, particularly the transition to post-quantum cryptography. The work serves as a guide for designing systems that can accommodate cryptographic change, rather than prescribing a specific implementation strategy.