Life at Eclipse

Europe has proposed new legislation intended to improve the state of cybersecurity for software and hardware products made available in Europe. The Cyber Resilience Act (“CRA”) will mandate that all manufacturers take security into account across both their development processes and the lifecycle of their products once in the hands of consumers. 

This document discusses the legislation and the potential impact it may have on the Eclipse Foundation and its 400+ projects and community. Many of the issues noted could have a similar impact on other open source organizations and projects. It is written based on our reading of the current draft legislation and a number of assumptions stated below. Note that is consciously does not include a discussion of possible revisions to the legislation, although we may post a followup which does. It also does not include any discussion concerning the warranty and product liability provisions of the legislation as we have not yet analyzed the impact those may have on us.

We are sincerely looking for comments and feedback, as it is quite possible that we have misunderstood or misinterpreted the documents.

It is important to stress that the Eclipse Foundation is better positioned to deal with the fallout from the CRA than many other open source organizations. We have staff. We have some resources. We have common community processes and practices shared across our many projects. We have CI/CD infrastructure shared by most (but not all) of our projects. We have a security team, written security policies and procedures, and are a CVE numbering authority. Despite being in a better position than most, we fear that the obligations set forth by the legislation will cripple the Eclipse Foundation and its community. 

There are a number of other excellent summaries of the worrisome impact of this legislation on the open source ecosystem. We highly recommend reading:

• Open-source software vs. the proposed Cyber Resilience Act by Maarten Aertsen.
• The EU’s Proposed Cyber Resilience Act Will Damage the Open Source Ecosystem by Olaf Kolkman.

Both of those articles primarily focus on the potential impact of the CRA on individual open source projects. Olaf’s document in particular suggests improvements to the draft. In this document we want to focus on the impact on an organization such as the Eclipse Foundation and its open source projects if the CRA was approved in its current form. How the CRA should or could be amended is being discussed elsewhere. The purpose of this document is to provide a resource explaining the impact of the legislation as it stands today.

It is important to note that the CRA does make a laudable attempt to carve out free and open source software but only “…outside the course of a commercial activity…”. Maarten Aertsen does an excellent job of summarizing the problems with this carve out. In particular he references a definition of commercial activity used in EU Blue guide to the implementation of EU product rules which states:

Commercial activity is understood as providing goods in a business related context. Non-profit organisations may be considered as carrying out commercial activities if they operate in such a context. This can only be appreciated on a case by case basis taking into account the regularity of the supplies, the characteristics of the product, the intentions of the supplier, etc. In principle, occasional supplies by charities or hobbyists should not be considered as taking place in a business related context.

Assumptions

• The CRA references the term “product” over 600 times but does not appear to define it. The act does define the term ‘product with digital elements’. For the purposes of this document we will assume that for the purposes of the CRA, any Eclipse Foundation project software made generally available to the public as a downloadable, installable, and executable binary would be considered a ‘product with digital elements’ under the regulation.
  • In addition, there are at least some EF projects which may be considered ‘critical product with digital elements’ (e.g. Kura, Keyple, ioFog, fog05) or ‘highly critical product with digital elements’ (e.g. Oniro, Leda, 4diac) .
• The CRA defines ‘manufacturer’ as “any natural or legal person who develops or manufactures products with digital elements or has products with digital elements designed, developed or manufactured, and markets them under his or her name or trademark, whether for payment or free of charge”. For the purposes of this document, we will assume that the Eclipse Foundation would be considered the manufacturer of the binaries produced by its projects. Among other reasons justifying this assumption, the Eclipse Foundation asserts that it owns the trademark rights for each of its projects and the binaries they release and (resources permitting) we market them as works of the Eclipse Foundation. 
• As mentioned above there is an attempt to exclude free and open source software produced outside the course of a commercial activity from the scope of the legislation. For the purposes of this document we will assume that Eclipse Foundation project software would be considered as produced under the course of a commercial activity, and would therefore be subject to the legislation. This assumption is based on the following:
  • The Eclipse Foundation is not a charity. It is a Belgian-incorporated international nonprofit association of hundreds of business members. 
  • Eclipse Foundation projects are not, generally speaking, developed by hobbyists. While some are, our projects are commonly developed by full-time employees of our member companies or by individuals who are making a living from consulting services related to their project work. 
  • Eclipse Foundation projects provide goods in a business related context. By that we mean that EF projects are largely intended to provide software which is immediately ready for adoption by businesses either as a component within a commercial product or by use by employees in their daily work.
  • Eclipse Foundation projects provide a regularity of supply. As one extreme example, the Eclipse IDE takes great pride in having not missed a single release date in over 15 years.
  • Eclipse Foundation projects deliver high quality software, equivalent to the quality found in commercial products. So the “characteristics of the product” are equivalent to commercial products. 

Having said all of the above it is important to remind the reader that all Eclipse Foundation projects provide their software for free, on a non-profit basis, and under OSI-approved open source licenses which permit further use, study, modification, and distribution. 

Impact Assessment

CE Markings for Software Products

Fundamentally, the core of the proposed legislation is to extend the CE Mark regime to all products with digital elements sold in Europe. Our assumption based on the current text is that this process will be applied to open source software made available under open source licenses and provided free of charge, ostensibly under licenses which disclaim any liability or warranty. We are deeply concerned that the CRA could fundamentally alter the social contract which underpins the entire open source ecosystem: open source software provided for free, for any purpose, which can be modified and further distributed for free, but without warranty or liability to the authors, contributors, or open source distributors. Legally altering this arrangement through legislation can reasonably be expected to cause unintended consequences to the innovation economy in Europe.

Without a clearer exemption for open source, in order to comply with the legislation the Eclipse Foundation will be required to:

• Develop, document, and implement policies and procedures for every project at the Eclipse Foundation to ensure they are conformant with the requirements of the CRA including:
  • All of the development and post-release security requirements set forth in Annex I, including providing notification and update mechanisms. 
  • All of the user documentation requirements set forth in Annex II.
  • All of the product technical documentation set forth in Annex V, including “…complete information on the design and development of the product…including, where applicable, drawings and schemes and/or a description of the system architecture explaining how software components build on or feed into each other and integrate into the overall processing.”
• For each EF project release, prepare the project-specific documentation required by Annex V, including “…an assessment of the cybersecurity risks against which the product with digital elements is designed, developed, produced, delivered and maintained…”.
• Determine for each project whether it meets the definition of ‘product with digital elements’, ‘critical product with digital elements’, or ‘highly critical product with digital elements’.
  • For each project which is a ‘product with digital elements’, establish, complete, and document a CE mark self assessment process.
  • For each ‘critical product with digital elements’ or ‘highly critical product with digital elements’ engage with an external CE auditing body and complete the additional processes required to get the CE mark approval. Note that it is not clear to us what the costs in time, resources, and money would be to implement these external audit processes. Our assumption is that they would be substantial.
    
    It is also important to note that in most other domains regulated with CE markings they are done where there are well known standards, specifications, and/or certification processes in place. These are not in place for most Eclipse Foundation open source projects. This could significantly increase the costs and risks associated with conformance.
• For each single project release, document that the relevant CE mark process is followed (as described above), that an EU declaration of conformity is written and signed by an officer of the foundation, that the CE mark is affixed, and that the technical documentation and EU declaration of conformity is made available for at least 10 years after the release. Note that we estimate that in any given year the Eclipse Foundation’s projects make available several hundred releases.

Article 4(3)

Member States shall not prevent the making available of unfinished software which does not comply with this Regulation provided that the software is only made available for a limited period required for testing purposes and that a visible sign clearly indicates that it does not comply with this Regulation and will not be available on the market for purposes other than testing.

Many Eclipse Foundation projects make integration, nightly, weekly, and milestone builds available under their open source licenses available indefinitely. The intent is to provide for community testing and for traceability. These binaries are marked as such, but the terms under which they are provided do not require that they be used for testing purposes only. 

It is not clear how this requirement could be implemented by any open source project using modern CI/CD infrastructure and operating under the principle of transparency. Even if the binaries were marked as “testing purposes only”, the open source licenses they are provided under do, in fact, permit uses other than testing. Further, it is common practice to provide intermediate builds for extended periods of time (often permanently) to provide testers with access to past builds for problem identification and remediation. Discontinuing that practice would be significantly disruptive. And any solution based on providing intermediate builds under non-open source licenses would be impossible for Eclipse Foundation projects, as the EF does not own the copyright and obtaining the approval of all contributors would be impractical. In summary, compliance with this CRA requirement would represent a significant blow to open source development best practices. 

Article 5(1) and Section 1 of Annex I

(1) Products with digital elements shall be designed, developed and produced in such a way that they ensure an appropriate level of cybersecurity based on the risks

At a minimum this would require the development and enforcement of written policies requiring every project to assess their level of cybersecurity risk and to implement processes to ensure that there is a determination of the risk level and a justification for the development processes adopted.

(2) Products with digital elements shall be delivered without any known exploitable vulnerabilities
(3) On the basis of the risk assessment referred to in Article 10(2) and where applicable, products with digital elements shall:
(a) …(j)

These would require a material change to our community’s release processes to require attestations that there are no known vulnerabilities and to comply with the many requirements listed. 

(k) ensure that vulnerabilities can be addressed through security updates, including, where applicable, through automatic updates and the notification of available updates to users.

With a few exceptions, EF projects do not “call home”, require any sort of user registration, and do not provide a mechanism for notifying all users that an update is either available or required. Implementing these requirements would require a whole new infrastructure to be mandated across all projects. 

Article 5(2) and Section 2 of Annex I “Vulnerability Handling Requirements”

In general, the Eclipse Foundation is in decent shape to deal with many of the stated requirements. As noted above we have a security team, written security policies and procedures, and are a CVE numbering authority. However, there are two notable elements in the requirements. 

(1) identify and document vulnerabilities and components contained in the product, including by drawing up a software bill of materials in a commonly used and machine-readable format covering at the very least the top-level dependencies of the product

This would impose a legal requirement to produce SBOMs for all EF projects. Although it is something we aspire to, this is a very significant effort. It would also require actively monitoring all project dependencies for known vulnerabilities in dependencies. This is generally considered an unsolved problem within the open source ecosystem with no known path to implementation.

(3) apply effective and regular tests and reviews of the security of the product with digital elements;

These would require a material change to our community’s development processes to mandate a whole class of testing which is not currently mandated for our projects. This is a very significant effort both to implement and to maintain.