Rebuilding information networks through organized sharing is vital – effective information sharing shields from cyberattacks
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)
These incorrect assumptions have undermined information sharing as an effective tool against cyber threats, yet policies, structures, and processes must be based on assumptions about the overall environment in order to function. As a replacement for the faulty assumptions explained above, this chapter proposes four alternative presumptions to enable effective information sharing. First, cyber threat information consists of multiple information types across different levels, with distinct value to different consumers, meaning that information sharing needs to be tailored and nuanced. Second, for this reason, relevance and comparative advantage should drive sharing activities. Third, effective information sharing efforts must overcome context-specific technical, economic, legal, and cultural barriers; and fourth, trust is a necessary component of information sharing. The rest of this section will explore these alternative presumptions in greater detail. A. Types of Cyber Threat Information Chismon and Ruks (2015) assembled a useful taxonomy of cyber information categories based on the kind of decisions the information informs. A modified version of their taxonomy is shown in Table I. As detailed in Table I, different categories of information, from technical to strategic, are intended for different consumers. However, information across the four levels—technical, tactical, operational, and strategic—is interrelated. For example, technical and tactical information can be combined to generate operational cyber threat information to improve organisational understanding of an impending attacker’s methods and capabilities (Chismon & Ruks, 2015). Similarly, post-incident analysis of technical cyber threat information often provides the foundation for the implementation of a tactical level decision. A holistic assessment of technical, tactical, and operational inputs drives the output of strategic cyber threat information. Despite these complex relationships, this taxonomy provides a useful way to think about cyber threat information and is indicative of why technical data-sharing should not be the sole focus of information sharing programmes. Smaller or less mature organisations are unlikely to find much utility in technical or tactical information sharing, while even larger organisations may miss out on key operational or strategic information insights if they focus exclusively on the technical information. For this reason, the Cyber Threat Alliance (CTA), which includes established cyber security vendors and related enterprises, shares a total of ten types of actionable cyber threat information across these four categories, as recalled by the authors and detailed in Table II. Understanding the value of these various forms of cyber threat information requires taking a more mature and nuanced view than the simplistic assumption that more information sharing means better security. This expanded conceptual framework for cyber threat information sharing reflects the diversity of information that industry leaders already know must be shared to strengthen defences. Each type informs a different aspect of cyber security and has a different value in different situations. Broad adoption of this (still high-level) extension to the framework provided by Chismon and Ruks (2015) would enable cyber security practitioners to develop more nuanced and useful policies for information sharing. B. Relevance and Comparative Advantage in Information Sharing In other disciplines, from finance to health to politics to sports, organisations do not produce and consume the same information equally. Instead, wide variation occurs based on relevance to business models, missions, and perceived benefits. Cyber security practitioners and policymakers should expect cyber threat information sharing to behave similarly. Different organisations should produce and consume different types of information based on two principles: relevance and comparative advantage. These two concepts should drive who should be sharing what information with whom, in what detail, and at what periodicity. 1) Relevance of Information Companies, non-profit organisations, and government agencies all have goals or missions and employ specific business models to achieve those goals. Information sharing should relate directly to an organisation’s goals and business model. Thus, a cyber security vendor should share technical cyber threat information at speed and at scale continuously because it is directly relevant to their business model. Conversely, a medium-sized manufacturer primarily needs strategic and operational level cyber threat information— strategic warnings, best practices, and tactical warnings (e.g., if a government learns that the business or its industry is being targeted)—all of which need only to be updated when a change has occurred. Technical cyber threat information provided at scale to this business would simply not be useful. 2) Comparative Advantage of Information Sharing Even if some organisations can produce certain information types, others might be more efficient at that work. For example, although governments can use their intelligence and law enforcement capabilities to collect, process, and produce technical cyber threat information, they do not have a comparative advantage in that information type. Private sector companies can perform that function just as efficiently. However, governments have a comparative advantage in other categories, such as attribution of cyber attacks, strategic warnings, and tactical warnings, which benefit from nation-state-level intelligence capabilities and authorities. As in other activities, the principle of comparative advantage should determine which organisations should be collecting, processing, sharing, and consuming different types of information. C. Technical, Economic, Legal and Cultural Barriers At first glance, the barriers inhibiting information sharing seem quite varied. However, a closer review shows they fall into four categories: technical, economic, legal, and cultural. While their specific manifestations and relative significance will vary across sharing contexts, these barriers can combine in various ways to create a formidable obstacle to sharing. Technical barriers prevent information from moving rapidly at scale or in easily consumable formats. For example, inconsistent definitions and terminology and difficulty in achieving interoperability and automation remain significant obstacles (Zibak & Simpson, 2019). In turn, these barriers often inhibit the usability or reliability of shared information (ENISA, 2017). Economic barriers stem from the inability to identify a clear return on investment from sharing activities. Organisations ‘participate in sharing networks when their return is more than the cost to participate’ (Vázquez et al., 2012: p. 432). This problem can be compounded by first-mover disadvantage, given that ‘establishing threat intelligence sharing infrastructure is expensive … [while] in the long run, intelligence sharing could help bring down the overall security cost’ (Zibak & Simpson, 2019: p. 7). Absent a clear and immediate prospect of a return on investment, proponents often have difficulty making the business case to establish, invest in or sustain sharing activities. Legal barriers come from uncertainty about what information can be shared under what circumstances or unanswered questions about liability, fines, or prosecution. These uncertainties deter organisations from sharing. Privacy laws can hinder sharing by inadvertently classifying certain cyber threat information as private and thereby limiting how it can be used or distributed (Panda Security, 2018). These legal concerns require sharing organisations to undertake extensive consideration of their potential implications (Borden et al., 2018; Albakri et al., 2019). Finally, cultural barriers can also impede sharing (Luiijf & Kernkamp, 2015). For cyber security companies, it can be hard to overcome the idea that retaining unique data yields a competitive advantage. For other organisations, it can be hard to overcome sentiments such as ‘no one would target me’, ‘cyber security is too complex for executives and non-technical employees to understand’, or ‘falling victim to hackers is inevitable, so why bother?’ For governments, long-standing views about the appropriate respective roles of the public and private sectors get in the way of cooperation and sharing. The good news is that, over the last twenty years, practitioners have developed ways to overcome these barriers. The bad news is that none of these methods is frictionless or cost-free. For example, adopting technical standards for information sharing may require organisations to adjust business processes or infrastructure; high initial costs may need to be met with loans that are paid back by future sharing participants; legal consultations may be needed to shape sharing rules; and reluctant executives may need the benefits of information sharing to be explained in bottom-line terms. Across the board, information sharing requires organisations to expend resources, either money or time. These costs can decrease but do not disappear. Yet, to be worthwhile, information sharing needs to be sustained and organisations have to pay a long-term, regular cost for engaging in information sharing activities. This requirement, in turn, means that information sharing requires incentives to achieve the scope, scale, and speed required for effective cyber defence. Such incentives can range from the individual (avoiding the costs of a cyber incident) to the public (government grants) to the avoidance of sticks (fines or penalties for not engaging in appropriate sharing). Regardless, information sharing laws, policies, programmes, and structures should assume that information sharing is resource-heavy and requires sustained investment to occur. D. Trust as a Necessary Component of Information Sharing Experience from previous initiatives and programmes demonstrates that information sharing only occurs when the providers and recipients have a degree of trust. As noted by Wagner et al. (2018), trust ‘plays a critical role in sharing’ (p. 5). The European Network Information Agency (ENISA) observes that in situations where trust between members of the community is diminishing or non-existent the value of information shared is undermined (ENISA, 2013). For information sharing to work, it is necessary to ‘foster confidence for stakeholders that the provided information will be acted upon as intended’ (Wagner et al., 2018: p. 5). Information providers have to understand who will receive their information, what they will do with it, and what level of information sharing-related risk to expect, while information recipients want to know where the information came from and its reliability. To reach this level of confidence, information sharing organisations should ‘provide control mechanisms to specify what information is shared, how much of it and with whom’ (Sauerwein et al., 2017: p. 845). According to ENISA (2012, cited by Vázquez et al., 2012: p. 433), the use of intentionally carefully designed trust-building mechanisms, such as ‘the policies, membership rules, requirement for security clearance and interaction type’ can be beneficial in the context of information sharing and will support the creation of trust. Absent trust, information sharing will not occur no matter what structures and incentives are put in place. Trust does not require that the participants all like each other, nor does it mean they share everything. Trust means that participants have a reasonable belief that all other participants will adhere to the agreed rules.