| Internet-Draft | SCITT SW Supply Chain | September 2023 |
| Birkholz, et al. | Expires 17 March 2024 | [Page] |
This document includes a collection of representative Software Supply Chain Use Case Descriptions. These use cases aim to identify software supply chain problems that the industry faces today and acts as a guideline for developing a comprehensive solution for these classes of scenarios.¶
This note is to be removed before publishing as an RFC.¶
Status information for this document may be found at https://datatracker.ietf.org/doc/draft-ietf-scitt-software-use-cases/.¶
Discussion of this document takes place on the SCITT Working Group mailing list (mailto:scitt@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/scitt/. Subscribe at https://www.ietf.org/mailman/listinfo/scitt/.¶
Source for this draft and an issue tracker can be found at https://github.com/ietf-wg-scitt/draft-ietf-scitt-software-use-cases.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 17 March 2024.¶
Copyright (c) 2023 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
Modern software applications are an intricate mix of first-party and third-party code, development practices and tools, deployment methods and infrastructure, and interfaces and protocols. The software supply chain comprises all elements associated with an application's design, development, build, integration, deployment, and maintenance throughout its entire lifecycle. The complexity of software coupled with a lack of lifecycle visibility increases the risks associated with system attack surface and the number of cyber threats capable of harmful impacts, such as exfiltration of data, disruption of operations, and loss of reputation, intellectual property, and financial assets. There is a need for a platform architecture that will allow consumers to know that suppliers maintained appropriate security practices without requiring access to proprietary intellectual property. SCITT-enabled products and analytics solutions will assist in managing compliance and assessing risk to help prevent and detect supply chain attacks across the entire software lifecycle while prioritizing data privacy.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
Supply chain security is a paramount prerequisite to successfully protect consumers and minimize economic, public health, and safety impacts. Supply chain security has historically focused on risk management practices to safeguard logistics, meet compliance regulations, demand forecasts, and optimize inventory. While these elements are foundational to a healthy supply chain, an integrated cyber security-based perspective of the software supply chains remains broadly undefined. Recently, the global community has experienced numerous supply chain attacks targeting weaknesses in software supply chains. As illustrated in Figure 1, a software supply chain attack may leverage one or more lifecycle stages and directly or indirectly target the component.¶
Dependencies Malicious 3rd-party package or version
|
|
+-----+-----+
| |
| Code | Compromise source control
| |
+-----+-----+
|
+-----+-----+
| | Malicious plug-ins;
| Commit | Malicious commit
| |
+-----+-----+
|
+-----+-----+
| | Modify build tasks or build environment;
| Build | Poison build agent/compiler;
| | Tamper with build cache
+-----+-----+
|
+-----+-----+
| | Compromise test tools;
| Test | Falsification of test results
| |
+-----+-----+
|
+-----+-----+
| | Use bad package;
| Package | Compromise package repository
| |
+-----+-----+
|
+-----+-----+
| | Modify release tasks;
| Release | Modify build drop prior to release
| |
+-----+-----+
|
+-----+-----+
| |
| Deploy | Tamper with versioning and update process
| |
+-----------+
DevSecOps often depends on third-party and open-source solutions. These dependencies can be quite complex throughout the supply chain and render the checking of lifecycle compliance difficult. There is a need for manageable auditability and accountability of digital products. Typically, the range of types of statements about digital products (and their dependencies) is vast, heterogeneous, and can differ between community policy requirements. Taking the type and structure of all statements about digital and products into account might not be possible. Examples of statements may include commit signatures, build environment and parameters, software bill of materials, static and dynamic application security testing results, fuzz testing results, release approvals, deployment records, vulnerability scan results, and patch logs. In consequence, instead of trying to understand and describe the detailed syntax and semantics of every type of statement about digital products, the SCITT architecture focuses on ensuring statement authenticity, visibility/transparency, and intends to provide scalable accessibility. The following use case illustrates the scope of SCITT and elaborates on the generic problem statement above.¶
In the IT industry it is a common practice that once a software product is released, it is evaluated on various aspects. For example, an auditing company, a code review company or a government body will examine the software product and issue authoritative reports about the product. The end users (consumers or distribution entities) use these report to make an accurate assessment as to whether the software product is deemed fit to use.¶
There are multiple such authoritative bodies that make such assessments. There is no assurance that all the bodies may be aware of statements from other authoritative entities or actively acknowledge them. Discovery of all sources of such reports and/or identities of the authoritative bodies adds a significant cost to the end user or consumer of the product.¶
A consumer of released software product wants:¶
There is no standardized way to:¶
This use case is a specialization of the use case above.¶
A released software product is often accompanied by a set of complementary statements about it's security compliance. This gives enough confidence to both producers and consumers that the released software has a good security standard and is suitable to use.¶
Subsequently, multiple security researchers often run sophisticated security analysis tools on the same product. The intention is to identify any security weaknesses or vulnerabilities in the package.¶
Initially a particular analysis can identify itself as a simple weakness in a software component. Over a period of time, a statement from another third-party illustrates that the weakness is exposed in the same software component in a way that it is an exploitable vulnerability. The producer of the software product now provides a statement that confirms the linking of software component vulnerability with the software product and also issues an advisory statement on how to mitigate the vulnerability. At first, the producer provides an updated software product that still uses the vulnerable software component but shields the issue in a fashion that inhibits exploitation. Later, a second update of the software product includes a security patch to the affected software component from the software producer. Finally, a third update includes a new release (updated version) of the formerly insecure software component. For this release, both the software product and the affected software component are deemed secure by the producer and consumers.¶
A consumer of a released software wants:¶
There is no standardized way to:¶
A software component source (e.g., a library) released by a certain original producer is becoming popular. The released software component source is accompanied by a statement of authenticity (e.g., a detached signature). Over time, due to its enhanced applicability to various products, there has been an increasing amount of multiple providers of the same software component version on the internet.¶
Some providers include this particular software component as part of their release package bundle and provide the package with proof of authenticity using their own issuer authority. Some packages include the original statement of authenticity, and some do not. Over time, some providers no longer offer the exact same software component source but pre-compiled software component binaries. Some sources do not provide the exact same software component but include patches and fixes produced by third-parties, as these emerge faster than solutions from the original producer. Due to complex distribution and promotion lifecycle scenarios, the original software component takes myriad forms.¶
A consumer of a released software wants:¶
There is no standardized way to:¶
In contrast to operating systems or user space software components of a large and complex systems, firmware components are often already executed during boot-cycles before there is an opportunity to authenticate them.¶
Authentication takes place, for example, by validating a signed artifact against a Reference Integrity Manifest (RIM). Corresponding procedures are often called authenticated, measured, or secure boot. The output of these high assurance boot procedures is often used as input to more complex verifications known as remote attestation procedures.¶
If measurements before execution are not possible, static after-the-fact analysis is required, typically by examining artifacts. When best practices are followed, in such cases measurements (e.g., a hash or digests) are stored in a protected or shielded environment (e.g., TEEs or TPMs). After finishing a boot sequence, these measurements about foundational firmware are retrieved after-the-fact from shielded locations and must be compared to reference values that are part of RIMs. A verifying system appraising the integrity of a boot sequence must identify, locate, retrieve, and authenticate corresponding RIMs.¶
A consumer of published software wants:¶
There is no standardized way to:¶
An organization has established procurement requirements and compliance policies for software use. In order to allow the acquisition and deployment of software in certain security domains of the organization, a check of software quality and characteristics must succeed. Compliance and requirement checking includes audits of the results of organizational procedures and technical procedures, which can originate from checks conducted by the organization itself or checks conducted by trusted third parties. Consecutively, consumers of statements about released software can be auditors. Examples of procedure results important to audits include:¶
Relevant compliance, requirement, and check result documents originate from various sources and include a wide range of representations and formats.¶
A consumer of a released software wants:¶
There is no standardized way to:¶
Some software is deployed on systems not connected to the Internet. Authenticity checks for off-line systems can occur at time of deployment of released software. Off-line systems require appropriate configuration and maintenance to be able to conduct useful authenticity checks. If the off-line systems in operation are part of constrained node environments, they do not possess the capabilities to process and evaluate all kinds of different authenticity proofs that come with a released software.¶
A consumer of a released software wants:¶
There is no standardized way to:¶
Firmware is a critical component for successful execution of any constrained IoT device. It is often the bedrock on which the security story of the devices it powers. For example, personal health monitoring devices (eHealth devices) are generally battery driven and offer health telemetry monitoring, such as temperature, blood pressure, and pulse rate. These devices typically connect to the Internet through an intermediary base station using wireless technologies. Through this connection, the telemetry data and analytics transfer, and devices receive firmware updates when published by the vendor. The public network, open distribution system, and firmware update process create several security challenges.¶
Consumers and other interested parties of a firmware update ecosystem wants:¶
There is no standardized way to:¶
Software Integration is a complex activity. This typically involves getting various software components from multiple suppliers, producing an integrated package deployed as part of device assembly. For example, car manufacturers source integrated software for their autonomous vehicles from third parties that integrates software components from various sources. Integration complexity creates a higher risk of security vulnerabilities to the delivered software.¶
Consumers of an integrated software wants:¶
There is no standardized way to:¶
Software producers often have multiple and concurrent supported versions of a product. The versions may represent major feature or compatibility differentiating releases (1.0, 2.0), or implementations for different Operating System Platforms and Architectures (Linux, Mac, Windows, AMD, ARM, x86, x64).¶
For each release, the software producer must be capable of providing statements, unique to that version. Producers may provide free patching to a specific version the consumer has purchased, while requiring upgrades to newer major releases. Consumers need to know which updates are compatible with their environment. Third parties that provide statements of quality need to know how to differentiate supported version bands, avoiding the recommendation to upgrade to an incompatible version.¶
As versions age, and vulnerabilities are discovered, consumers need to know the newer version of a particular product. Software producers implement versioning updates, however there are no standards for consumers and third parties to apply across software producers.¶
Consumers of related software components want:¶
There is no standardized way to:¶