A logistics company starts its week on Monday morning, yet the workstations will not boot. Overnight, the vendor of the protective software had distributed a faulty update that crashes every computer at startup. Nobody attacked, and still operations stand still until every computer has been repaired by hand, one at a time. Software does exactly what the code says, and sometimes the code contains a mistake. Without any intent, it can have the same consequences as an attack.
Malfunctioning software is a technical risk source in the sense of ISO/IEC 27005. It covers unintentional malfunctions in your own and third-party software, from programming errors to faulty updates to vulnerabilities that only an attacker brings to light.
What is behind it?
This covers every program the organization works with: operating systems, business applications, in-house software, interfaces, scripts, firmware and cloud services. Software keeps getting more complex, and much of it consists of third-party components. A bug in a small library can therefore show up in many applications at once.
How bugs arise
- Complexity. Many functions, special cases and dependencies make it impossible to test every state.
- Changes. Every update, every new interface and every changed configuration can break processes that used to work.
- Time pressure in development. Tests and reviews get cut when a deadline looms.
- Dependencies. Libraries, frameworks and third-party services bring their own bugs, which you cannot fix yourself.
- Ageing. Software without vendor support no longer receives fixes, even when bugs and gaps are known.
Impact
Software bugs hit all protection goals. Crashes and hangs knock out availability. Calculation errors, wrong rounding or faulty data imports damage integrity, often unnoticed for a long time. Errors in authorization checks or caching show data to the wrong people and breach confidentiality. The most dangerous bugs are those that silently corrupt data, because the corrupted data travels unnoticed into the backups.
Practical examples
Rounded the wrong way. After a change to the interface between inventory management and the invoicing program, VAT is rounded incorrectly for one product group. The error is small and only surfaces months later during an audit. Hundreds of invoices have to be corrected.
Other people’s data in the customer portal. A misconfigured cache in an insurance broker’s customer portal serves logged-in customers the pages of other customers for several hours. Several customers see contracts that are not theirs and get in touch. For the broker, a notifiable data breach follows.
The creeping memory leak. A reporting service fails to release memory it has used. Every two weeks the ERP server’s memory fills up and the system freezes. Until the cause is found, only a scheduled restart helps.
Typical threats
These BSI elementary threats stem from this risk source:
- G 0.28 Software Vulnerabilities or Errors: the core of this risk source.
- G 0.46 Loss of Integrity of Sensitive Information and G 0.22 Manipulation of Information: data altered unnoticed by bugs.
- G 0.45 Data Loss: errors in synchronization, migration or deletion routines.
- G 0.27 Lack of Resources: memory leaks and programs that use up computing power or storage.
- G 0.11 Failure or Disruption of Service Providers: software bugs at cloud and IT service providers.
- G 0.21 Manipulation of Hardware or Software: updates that change settings or functions unintentionally.
- G 0.23 Unauthorised Access to IT Systems and G 0.47 Harmful Side Effects of IT-Supported Attacks: bugs that open doors for attackers, including for attacks on third parties.
Effective controls
Prevention:
- A.8.25 Secure Development Life Cycle: quality and security in every phase of development.
- A.8.29 Security Testing in Development and Acceptance: tests before production use, including for updates to third-party software.
- A.8.32 Change Management: introducing changes in a planned, checked and reversible way.
- A.8.8 Management of Technical Vulnerabilities: fixing known bugs and gaps promptly.
Detection:
- A.8.16 Monitoring Activities: watching error messages, utilization and unusual behaviour of applications.
Response:
- A.8.13 Information Backup: backups that reach back far enough to correct creeping errors as well.
- A.5.30 ICT Readiness for Business Continuity: plans for a central application being down for days.
Risk sources in ISO 27005
ISO/IEC 27005 names technical risk sources as a type of their own in its definition. In Table A.10, software faults appear in several places, for example in the failure of systems, in the saturation of the information system and in tampering with software, which the standard explicitly marks as possible by accident too. Among the aspects for likelihood criteria it lists technology failure, and in the assessment it names known weaknesses and the effectiveness of existing controls as factors. To find technical vulnerabilities, Annex A.2.5.3 describes methods such as automated scanning, security testing, penetration testing and code review. The standard rates code review there as the most thorough and at the same time the most expensive method (→ ISO/IEC 27005 Information Security Risk Management).
Sources
- ISO/IEC 27005:2022, definition 3.1.6, clauses 6.4.3.3 and 7.3.3, Annex A.2.5.1 and A.2.5.3
- BSI IT-Grundschutz Compendium: elementary threats G 0.11 to G 0.47
- ISO/IEC 27002:2022, clauses 5.30, 8.8, 8.13, 8.16, 8.25, 8.29 and 8.32