Modern software is assembled using third-party and open source components, glued together in complex and unique ways, and integrated with original code to provide the desired functionality. Third-party (including commercially licensed, proprietary, and “source available” software) along with open source components provide the necessary building blocks that allow organizations to deliver value, improve quality, reduce risk and time-to-market. The benefits of open source are many. However, by using open source components, organizations ultimately take responsibility for code they did not write. Strategic alliances between organizations and open source projects can lead to healthy open source usage and overall risk reduction.
Component Analysis is the process of identifying potential areas of risk from the use of third-party and open-source software and hardware components. Component Analysis is a function within an overall Cyber Supply Chain Risk Management (C-SCRM) framework. A software-only subset of Component Analysis with limited scope is commonly referred to as Software Composition Analysis (SCA).
Any component that has the potential to adversely impact cyber supply-chain risk is a candidate for Component Analysis.
Common Risk Factors
Having an accurate inventory of all third-party and open source components is pivotal for risk identification. Without such knowledge, other factors of Component Analysis become impractical to determine with high confidence. Use of dependency management solutions and/or Software Bill of Materials (SBOM) can aid in inventory creation.
Components may have varying degrees of age acceptance criteria. Factors that impact acceptable age include the type of component, the ecosystem the component is a part of (Maven, NPM, etc), and the purpose of the component. The age of a component may signify use of outdated technology and may have a higher probability of being passed over by security researchers.
Newer versions of components may improve quality or performance. These improvements can be inherited by the applications that have a dependency on them. Components that are end-of-life (EOL) or end-of-support (EOS) also impact risk. Two common approaches to community supported open-source is:
- Support the latest revision of the last (x) releases - (i.e. 4.3.6 and 4.2.9)
- Support only the latest published version (i.e. 4.3.6 today and 4.4.0 tomorrow)
Depending on the risk appetite, it may be strategic to only use third-party and open source components that are supported.
With reference to Semantic Versioning terminology, API changes can be expected between major versions of a component, but are rare between minor versions and patches. Changes in a components API may result in longer remediation times if the components have not been continuously updated. Keeping components up-to-date can reduce remediation time when a rapid response is warranted.
Outdated version identification may leverage ecosystem-specific repositories and is achievable through the use of dependency managers or Package URL. Component analysis can identify outdated components as well as provide information about newer versions.
- Applications (includes libraries and frameworks)
- Operating Systems
While the NVD may be the most recognizable source of vulnerability intelligence, it’s not the only. There are multiple public and commercial sources of vulnerability intelligence. Having a known vulnerability doesn’t require the vulnerability information be present in one of these sources. Simply being documented (i.e. social media post, defect tracker, commit log, release notes, etc) classifies the vulnerability as being known.
Component analysis will commonly identify known vulnerabilities from multiple sources of vulnerability intelligence.
Frameworks and libraries have unique upgrade challenges and associated risk. Abstractions, coupling, and architectural design patterns may affect the risk of using a given component type. For example, logging libraries may be embedded throughout a large application, but replacing implementations can likely be automated. Likewise, replacing a web application framework for an alternative framework would likely be a high-risk endeavor leading to architectural changes, regressions, and code rewrites. Evaluating the type should be part of every Component Analysis strategy.
Identifying and analyzing the purpose of each component may reveal the existence of components with duplicate or similar functionality. For example, it’s unlikely an application would need multiple XML parsers or cryptographic providers. Potential risk can be reduced by minimizing the number of components for each function and by choosing the highest quality components for each function. Evaluating component function should be part of every Component Analysis strategy.
The number of third-party and open source components in a project should be evaluated. The operational and maintenance cost of using open source will increase with the adoption of every new component. The impact can be significantly higher with micro-module ecosystems when there are hundreds or thousands of applications in a given environment. In addition to increased operational and maintenance cost, a decrease in a development teams ability to maintain growing sets of components over time can be expected. This is especially true for teams with time-boxed constraints.
Components in many software ecosystems are published and distributed to central repositories. Repositories have known threats. Some of the threats against public repositories include:
- Typosquatting - naming a component in such as way as to take advantage of common misspelling
- Organization/Group abuse - pretending to be a public person or entity and abusing the perceived trust. This includes dependency confusion and namespace confusion.
- Malware through transfer - leveraging weak or absent code-signing requirements to spread malware through the transfer of an open source project from one maintainer to another
- Cross Build Injection (XBI) - Abusing dependency resolution schemes and weak infrastructure controls to inject malicious components in place of safe ones
Public repositories that have code-signing and verification requirements have some level of trust, whereas public repositories without basic countermeasures do not. For no-trust or low-trust repositories, utilizing private repositories may be advantageous. Private repositories refer to repositories where access is limited, usually software that organizations install and control, or a commercially available service. Golden repositories containing vetted or permitted components are a common use-case for private repositories. Private repository services focusing on security may additionally provide anti-malware analysis and static source code analysis requirements prior to acceptance in the repository. When leveraging private repositories, it is important to have traceability to the components’ provenance.
A component’s provenance refers to the traceability of all authorship, build, release, packaging, and distribution across the entire supply chain. In physical supply chains this is referred to as the chain of custody. Provenance may include individual and community authorship of software components, manufacturers, suppliers, software repositories, and country of origin. For high assurance applications, provenance plays an important role in determining Foreign Ownership, Control, or Influence (FOCI).
Data which describes the lineage for which software has been created or altered. Pedigree includes the ancestors, descendants, and variants which describe component lineage from any viewpoint and the commits, patches, and diffs which make a component unique. Maintaining accurate pedigree information is especially important with open source components whos source code is readily available, modifiable, and redistributable.
Formulation describes how components were built often including build system invocation and properties, SDK and compiler versions, compiler flags, and a comprehensive list of parallel and sequential steps that were taken to build, test, and deliver a component. Formulation and pedigree are complimentary concepts and are often combined and referred simply as pedigree.
Third-party and open-source software typically has one or more licenses assigned. The chosen license may or may not allow certain types of usage, contain distribution requirements or limitations, or require specific actions if the component is modified. Component Analysis can identify the license(s) for a given component and may optionally provide guidance as to the nature of the license (i.e. copyright, copyleft, OSI approved, etc). Utilizing components with licenses which conflict with an organizations objectives or ability can create serious risk to the business.
Third-party and open source components often have dependencies on other components. A transitive dependency is when an application has a direct dependency on a component and that component has a dependency on another component. Transitive dependencies are common and are expected in highly modular ecosystems which values reuse over re-invent. Like any component, transitive dependencies have their own risk which is inherited by every component and application that relies on them. Components may additionally have specific runtime or environmental dependencies with implementation details not known or prescribed by the component. Component Analysis can aggregate the risk of all direct, transitive, runtime, and environmental dependencies providing a holistic view of inherited risk.
There are many datapoints to consider when evaluating the health of an open source project.
- Quality Controls and Metrics - The overall quality and controls for achieving and maintaining high-quality components may be a factor in risk evaluation. For software components, this refers to the use of unit and integration tests, linters and static analysis tools, the percentage of coverage, and results from various tools.
- Community Engagement - The current and historical trend for a project and its maintainers to accept pull requests, answer defect and enhancement requests, and engage in productive collaboration with the community may be a factor in risk evaluation.
- Vulnerability Analysis - Analyzing current and historical security vulnerabilities for timeline trends and for root-cause patterns may signify a projects ability to protect the community from future (and similar) issues. This activity may be a factor in risk evaluation.
Applications or their direct or transitive components may rely on external services for functionality. Common examples include convenience libraries around common web services such as mapping, stock quotes, or weather, but may also include autogenerated software components built from API specifications such as TUF or OpenAPI. External services should be included in the overall inventory of components.
Software Bill of Materials (SBOM)
Multiple efforts between government and industry are attempting to define Software Transparency. Some of these efforts will lead to increased compliance or regulatory requirements. Software Transparency is often achieved through the publishing of software bill of materials. An SBOM is synonymous to the list of ingredients in a recipe. Both are an implementation of transparency.
There are multiple SBOM standards including OWASP CycloneDX and SPDX, each having their own strengths and use-cases they were designed to solve. Evaluating SBOM standards to determine which are applicable to an organizations requirements should be part of an overall C-SCRM strategy.
Component ecosystems generally devise different terminology and formats for representing components. This self-imposed fragmentation makes uniquely identifying and representing components difficult when referring to them outside of their respective ecosystems. Centralized databases such as the CPE Product Dictionary maintained by NIST adds additional fragmentation as the CPE vendor and product names often do not reflect reality.
SWID tags are an alternative approach to component identity. They have historically been used to describe commercial software for the purpose of license entitlements and ITOM discovery within a CMDB system. Software manufactures can create their own SWID tags whenever new versions are released.
Generally, a component will have a name and version. Components may optionally have a higher-level grouping identifier, commonly referred to as a groupId, organization, or vendor. When referencing components in a C-SCRM framework it is important to have a standard and uniform way to represent them. The Package URL specification provides a decentralized and uniform way to represent components.
pkg:deb/debian/[email protected]?arch=i386&distro=jessie pkg:docker/[email protected]:244fd47e07d1004f0aed9c pkg:gem/[email protected]?platform=java pkg:maven/org.apache.xmlgraphics/[email protected]?packaging=sources pkg:npm/%40angular/[email protected] pkg:nuget/[email protected] pkg:pypi/[email protected] pkg:rpm/fedora/[email protected]?arch=i386&distro=fedora-25
Multiple standards exist for the verification of software components including:
- OWASP Software Component Verification Standard (SCVS)
- Supply-chain Levels for Software Artifacts (SLSA)
Both standards aim to measure and improve the assurance of the software supply chain.
Open Source Policy
Open source policies provide guidance and governance to organizations looking to reduce third-party and open source risk. Policies typically include:
- Restrictions on component age
- Restrictions on outdated and EOL/EOS components
- Prohibition of components with known vulnerabilities
- Restrictions on public repository usage
- Restrictions on acceptable licenses
- Component update requirements
- Deny list of prohibited components and versions
- Acceptable community contribution guidelines
While the open source policy is usually filled with restrictions, it provides an organizations security, development, and legal teams an opportunity to create solutions for healthy open source usage.
- Limit the age of acceptable components to three years or less with exceptions being made for high-value, single purpose components that are still relevant
- Prohibit the use of end-of-life (EOL) components
- Prohibit the use of components with known vulnerabilities. Update components that are exploitable with high to moderate risk first.
- Reduce the attack surface by excluding unnecessary direct and transitive dependencies
- Reduce the number of suppliers and use the highest quality components from those suppliers
- Standardize on a single component for each component function
- Establish a maximum level of acceptable risk for public repositories. Utilize private repositories in lieu of untrusted ones.
- Automate component updates (from trusted repositories only)
- Provide time-boxed allowances every sprint to maintain component hygiene
- Establish a allowed list of acceptable licenses, a deny list of prohibited licenses, and seek advice from counsel for all other licenses
- Automate the creation of software bill-of-materials (SBOM) for all deliverables
- Leverage Package URL for describing components within SBOMs
- Contractually require SBOMs from vendors and embed their acquisition in the procurement process
- Automate the analysis of all third-party and open source components during Continuous Integration (CI), either by analyzing the files themselves, or by analyzing a SBOM
- Import SBOMs into systems capable of tracking, analyzing, and proactively monitoring all components used by every asset in an environment (i.e. enterprise wide, entire cloud infrastructure, etc)
|Black Duck Hub||Synopsys||Commercial||Cross Platform|
|Bytesafe Dependency Checker||Bytesafe||Free||SaaS|
|CAST Highlight||CAST Software||Commercial||SaaS|
|Clarity||Insignary||Commercial||Cross Platform / SaaS|
|ClearlyDefined||Open Source Initiative||Open Source||SaaS|
|CodeSentry||GrammaTech||Commercial||Cross Platform / SaaS|
|Dependency-Check||OWASP||Open Source||Cross Platform|
|Dependency-Track||OWASP||Open Source||Cross Platform|
|Dependabot||Dependabot||Commercial / Freemium||SaaS|
|DepShield||Sonatype||Open Source||Cross Platform / SaaS|
|DotNET Retire||Retire.NET Project||Open Source||Cross Platform|
|FlexNet Code Insight||Flexera Commercial||Cross Platform|
|Fluid Attack’s Scanner||Fluid Attacks||MPL 2.0.||SaaS|
|FOSSA||FOSSA||Commercial / Freemium||SaaS|
|FOSSology||Linux Foundation||Open Source||Cross Platform|
|Grafeas||Grafeas||Open Source||Cross Platform|
|OSS Review Toolkit||HERE||Open Source||Cross Platform|
|Ion Channel SA||Ion Channel||Commercial||SaaS|
|NPM Audit||NPM||Open Source||SaaS|
|OSS Index||Sonatype||Free / Open Source||SaaS|
|PHP Security Checker||SensioLabs||Open Source||SaaS|
|Prisma Cloud||Palo Alto Networks||Commercial||Cross Platform / SaaS|
|Renovate||Key Location||Open Source||SaaS|
|Snyk||Snyk||Commercial / Freemium||SaaS|
|SourceClear||Veracode||Commercial||Cross Platform / SaaS|
|Nexus IQ||Sonatype||Commercial||Cross Platform|
|Open Source Lifecycle Management||WhiteSource Software||Commercial||Cross Platform / SaaS|
|Retire.js||RetireJS Project||Open Source||Cross Platform|
|VulnDB||Risk Based Security||Commercial||SaaS|
|Patton||OWASP||Open Source||Cross Platform|
|Vigiles||Timesys||Commercial / Freemium||SaaS|
|Vigilant Ops InSight||Vigilant Ops||Commercial||Cross Platform / SaaS|
- Cyber Supply Chain Risk Management (C-SCRM)
- Impact of software supply chain practices: Development Waste (Sonatype)
- Content of Premarket Submissions for Management of Cybersecurity in Medical Devices (FDA)
- CycloneDX specification
- Software Package Data Exchange (SPDX)
- SWID (ISO/IEC 19770-2:2015)
- Package URL specification and reference implementations
- Using Software Bill-of-Materials to drive change and reduce risk
- The Unfortunate Reality of Insecure Libraries (Contrast Security)
- Framework for Software Supply Chain Integrity (SAFECode)
- Managing Security Risks Inherent in the Use of Third-party Components (SAFECode)
- Deliver Uncompromised (MITRE)