Free beta access. 5 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Status Quo Information Sharing Is Lacking – technical data is not the only form of information that is valuable

Daniel & Kenway ’20 (Michael Daniel currently serves as President and CEO of the non-profit Cyber Threat Alliance (CTA). Joshua Kenway is a Cybersecurity Associate at the Cyber Threat Alliance and a Research Fellow with the Algorithmic Justice League. “Repairing the Foundation: How Cyber Threat Information Sharing Can Live Up to its Promise and Implications for NATO,” in Cyber Threats and NATO 2030: Horizon Scanning and Analysis, December 2020, https://ccdcoe.org/uploads/2020/12/Cyber-Threats-and-NATO-2030_Horizon-Scanning-and-Analysis.pdf)

The assumption that cyber threat information is equivalent or primarily composed of technical data severely limits the potential value of information sharing. Technical data, while necessary, is not the only form of information that can provide value. For example, a warning from the US Federal Bureau of Investigation (FBI) that a specific Chinese cyber group is targeting a US company with cyber-enabled theft of intellectual property would be a useful piece of non-technical intelligence for that company. Written advisories about vulnerabilities and associated patches are critical to organisations using vulnerable software or hardware; in fact, such information is far more useful to most organisations than technical data on one of the many variations of the LockerGoGo malware. The most common interpretation of information is too narrow. B. Every Organisation Should Produce and Consume Technical Data If the underlying assumption is that information sharing means technical information, then it logically follows that most policies, infrastructure, and programmes for sharing are built around the idea that most organisations should produce and consume technical information. If everyone were to collect, share, and consume such data, the thinking goes, security would improve across the ecosystem. The problem with this logic is that most organisations are lousy at collecting, producing, and consuming technical data—and always will be. Most companies do not have the capability to identify a malware binary, analyse it, and use the resulting information, nor would they know how to handle a malicious domain name. As a practical matter, this situation will not change; no country will have enough cyber security professionals for every organisation to have this capacity. Small and medium businesses do not and will not have the resources to collect, process, share, and consume technical data regularly. This limitation does not mean such organisations would not benefit from cyber threat information sharing; rather they need different information. Neither is this approach economically efficient. Most organisations do not need access to technical data in real-time. Despite the rapidly changing nature of cyber threats in a technical sense, for most organisations, cyber security requirements and best practices do not change much from day to day. 2 For example, see the Cybersecurity Information Sharing Act of 2015 (Consolidated Appropriations Act of 2016, Division N, Cybersecurity Information Sharing Act of 2015) and the Automated Indicator Sharing Program (DHS Cybersecurity and Infrastructure Security Agency, 2020). 182 In addition, not every business has in-house technical accounting or legal skills—why should cyber security be different? Current practice does not sufficiently differentiate between organisations in terms of what information they should share under what circumstances and how frequently. C. Information Sharing is Easy In January 2008, the US government started the Comprehensive National Cybersecurity Initiative (CNCI), formalising it in National Security Presidential Directive 54 / Homeland Security Presidential Directive 23 (The White House, 2010). ‘Connect the Centers,’ one of CNCI’s twelve lines of effort, focused on information sharing with the goal of linking the US government’s cyber centres into one common operating picture; over the long-term, it was intended to incorporate the private sector. Everyone assumed that this element would be the easiest to implement and the first to be completed. However, thirteen years later, this element is arguably one part of the CNCI vision that remains unrealised as the cyber security centres are not seamlessly connected and many silos remain stubbornly in place. A similar situation has played out in the private sector with the creation of Information Sharing and Analysis Centers (ISACs). The assumption was that companies would eagerly join these organisations, share what they knew and consume the information shared by others. Yet, more than twenty years after the concept was formalised into national legislation, many sectors are just now forming an ISAC and, even in the most successful of them, the percentage of participants that actively share information is widely understood within the industry to remain low. Public sector efforts to share information with the private sector have suffered analogous problems. The US government created the Automated Indicator Sharing (AIS) programme as a free service for general businesses, but few organisations have signed up and even fewer contribute to the programme (Marks, 2018). This is unsurprising if we look at what is being shared; a US government report from 2018 suggested that just two or three out of the 11,447 indicators submitted to AIS by the Department of Homeland Security were ‘malicious and related to cyber incidents [… while] many of the indicators received were false positives or redundant information’ (DHS Office of Inspector General, 2017: p. 15). The three examples highlight that information sharing is difficult for a variety of reasons. Simply creating programmes and establishing sharing mechanisms is insufficient without addressing obstacles to sharing actionable information. These include underlying factors such as over-classification, reputational risk, and legal concerns, as well as operational hurdles around validation, standardisation, timeliness, and automation (Zibak & Simpson, 2019).

Sign in to recut, update, or save this card.