Second --- Cyber attacks are inevitable --- lack of resilience ensures they are successful --- killing millions and collapsing society --- increased situational awareness is key to mitigate harms
Conklin et al 17 --- William Arthur Conklin, Associate Professor and Director of the Center for Information Security Research and Education in the College of Technology at the University of Houston, & Dan Shoemaker, Professor and Director of the Masters of Science in Cyber Security Program at the Center for Cyber Security and Intelligence Studies, a NSA Center of Academic Excellence, at the University of Detroit (313) Mercy, & Anne Kohnke, College of Management, Lawrence Technological University, “Cyber Resilience: Rethinking Cybersecurity Strategy to Build a Cyber Resilient Architecture”, International Conference on Cyber Warfare and Security, 2017, https://www.proquest.com/docview/1897660614/fulltextPDF/FE6EBF8A132A40C1PQ/1?accountid=12598
Von Clausewitz might have lived through the Napoleonic wars, but he could just as well have been writing about the present state of affairs in cyberspace. The fact is that the multitude of threats that populate the virtual eco-system is so all-embracing that it is impossible to defend against every form of attack (Symantec, 2015). The unfortunate reality is that every system can eventually be compromised and no computer is safe. Furthermore, America's electronic infrastructure is so riddled with vulnerabilities and there are so many adversaries out there, that "cybersecurity" is a relative term indeed (Brasso, 2016).
Adversaries in cyberspace range from state-sponsored hackers to the professional perpetrators of cybercrime, to any person with a grudge and an internet link. Moreover, because of the proliferation of adversaries and the inescapable vulnerability of the U.S. critical infrastructure, cyberattacks on its various elements are a daily fact of life. For instance, the Industrial Control Systems-Computer Emergency Response Team (ICS-CERT) reports that the U.S. industrial control systems were attacked at least 245 times over a 12-month period (OAS, 2015). In other parts of the world, a 2015 attack on Ukraine's power grid left 700,000 people without electricity for several hours [Brasso, 2016). And the perpetrators of the Ukrainian attack were observed conducting similar attacks against the U.S. energy sector (Brasso, 2016). Though there was never any actual disruption, many experts believe that those exploits were a reconnaissance for future moves on the U.S. infrastructure (Brasso, 2016). So they are still out there.
The reason why this topic is so worrisome is that a successful digital Pearl Harbor type cyberattack on the electrical infrastructure would leave the U.S. cloaked in darkness, with its people unable to communicate with each other and without any form of twenty-first century transport. It would likely kill many thousands of citizens, perhaps millions, either through civil unrest, failure of public systems, or mass starvation (Brasso, 2016). This could be the foreseeable situation if hackers successfully shut down the U.S. power grid. Consequently, in the larger sense, the key question is, "Could a catastrophic cyberattack in the U.S. infrastructure ever occur? The National Security Agency's former Director, Mike Rodgers, made his own evaluation of the possibility of a successful attack against critical infrastructure when he said; "It's a matter of, when, not if" (Smith, 2014).
To illustrate the reason why something as unthinkable as a digital Pearl Harbor could happen, the non-profit Privacy Rights Clearinghouse reports that we have lost over one BILLION records over the past decade and these are just the losses that were REPORTED. Since most companies do not like to publicize their security failures that number could be, and probably is, much higher. The running average of 100 million records lost per year has been subject to some variation overtime and the source of breaches has changed in common-sense ways as the technology has evolved. Nonetheless, the number of reported incidents rose annually from 108 in 2005, to 607 in 2013 (PRC, 2015). The important point to note is that while losses have increased at a rate of fifty-percent per year over the same period corporate budgets for security have increased at the rate of fifty-one percent (Symantec, 2014).
Consequently, it is probably safe to say that cyberattacks are going to happen, no matter how much money we throw at the problem. At the same time, notwithstanding the clear recognition of the disastrous nature of a successful cyberattack, none of the industries in our current national infrastructure have developed an effective standard strategy to protect themselves (Brasso, 2016). This includes the Chemical Sector; Commercial Facilities: Communications; Dams; the Defense Industrial Base; Emergency Services; Energy: Financial Services; Food and Agriculture; Government Facilities; Healthcare and Public Health; Information Technology; Nuclear Reactors, Materials, and Waste; Transportation; and Water and Wastewater Systems (PPD-21, 2013, p.2).
So, what we are suggesting here is that the paradigm we adopted long before the internet needs to be seriously rethought. Right now, we secure systems at the logical points of access. Even defense-in-depth, which is the modern doctrine for doing that, simply involves several increasingly rigorous perimeters. Instead of throwing our resources into defending the system at all points of attack, perhaps we should devise ways to ensure that the system assets that we cannot afford to lose are unconditionally protected. In concept, this strategy would narrow the attack surface to the point where absolute assurance within available resources could be guaranteed, while reducing the impact and recovery time of non-essential assets. This selective versus universal approach to security is what sums up the difference between a cyberresilient strategy, and one that is based on cybersecurity.
2. Cyber resilience versus cybersecurity
It is important to make a distinction between present ideas about cybersecurity and the approach we are suggesting. Cyber resilience is a much narrower strategic vision than the more diffuse activities that fall under the generic heading of "cybersecurity." At its heart, cybersecurity ensures against unknown, or unwanted penetrations of protected space. In that respect, a cybersecurity solution deploys a comprehensive set of defenses to ensure against unauthorized access to protected objects, either from external or internal sources. To accomplish its mission, cybersecurity deploys a targeted set of protection measures at all viable points of entry, along with the appropriate detection mechanisms to respond to known forms of unauthorized access. Whereas, cyber resilience attempts to create a security architecture that will ensure the continuous effective functioning of just the organization's core functionality.
Thus, cybersecurity focuses on access control, while cyber-resilience focuses on identifying and securing just those assets that the organization can't live without. In many respects, the cyber-resilience concept is comparable to locking up just those critical resources that an organization needs to survive. Those critical resources are then put into a safe within the building. This is in direct contrast with a strategy that attempts to protect every asset of value within the entire building. It should be evident that these are two entirely different approaches, especially from a resource utilization standpoint.
Ideally these two approaches work in tandem to create a secure organization. Both are important and neither works as effectively as it should if the other is missing. Cyber resilience is an emerging concept and there are already several institutionally sponsored approaches to creating a cyber resilient organization. These methods develop comprehensive enterprise architectures with the aim of building in greater resiliency. The methodology centers on and is driven by assessment and metrics. The assessment characterizes and then seeks to improve the resiliency of the enterprise's existing architecture. The goal is to understand the organization's practical ability of uninterrupted functionality, as well as the ability to recover within an acceptable period. The metrics of this entire process are meant to ensure continuous improvement. The approach itself, is deliberate and incremental and is a broad-scale strategic development process in the sense that the requisite capabilities are phased into existing enterprise processes in a planned and rational fashion.
3. Creating a cyber resilient architecture
Cyber resilience is not just a matter of building higher and thicker walls. Cyber resilience involves the creation of a set of well-defined processes, which are designed to react to successful penetrations of the organizational perimeter, affecting secure space. Cyber resilient processes prevent the attacker from ever reaching and harming the organization's most critical assets and they do not fortify every access point beyond basic access control countermeasures.
Cyber resilience is founded on classification, prioritization, and comprehensive strategic policy-based deployment of a rigorous set of real-world security controls. The security controls strictly ensure the achievement of just the organizational purpose. Therefore, less resources are required. These controls are both electronic and behavioral and they are designed to protect key assets, as well as ensure optimum recovery of the overall system in the event of successful attack. Figure 1 below shows the cyber resiliency process. The process is embodied in seven generic principles and are as follows:
1. Classify: You can't protect things that you don't know exist. Therefore, all the organization's assets must be identified, labeled, and arrayed in a coherent baseline of "things." This baseline describes all potential protection targets and it is maintained under strict configuration management. This is comparable to the "Classification" phase of the NIST Risk Management Framework (NIST, 2014).
2. Risk: Resiliency requires appropriate situational awareness. Therefore, a broad-spectrum risk assessment must be performed that characterizes all known threat scenarios as they apply to the identified asset base. This is a risk assessment process comparable to the "Security Assessment" phase of NIST SP 800-53a (NIST SP 800-53A Rev.4, 2014).
3. Rank - the assets that the organization absolutely can't afford to lose are selected, evaluated and a provably effective countermeasure response is deployed for each of the chosen assets. Organizational resources are focused on only assuring those critical things. The resources that are left over are then allocated to ensuring the protection and recovery of the rest of the organization's functionality. This is comparable to the "Select" phase of the NIST Risk Management Framework (NIST, 2014).
4. Design/Deplov: Resilience must be baked-into the architecture so that critical functions are assured given a generally successful attack. This is a design/control deployment exercise comparable to the "Implement" phase of the NIST Risk Management Framework (NIST, 2014).
5. Test: Architectural resilience needs to be assured. This is a planning and oversight function that characterizes critical control performance against stated mission goals. Methods like penetration testing apply here. This is an assessment process comparable to the "Assess" phase of the NIST Risk Management Framework (NIST, 2014).
6. Recover: Well-defined processes need to be documented and established to ensure that all enterprise functions are fully restored within requisite parameters. This is comparable to the "Plan-Purpose-Scope-Relationship" recommendations embodied in NIST SP 800-34 Rev. 1 (NIST, 2010).
7. Evolve: the organization dynamically adjusts its cyber-resilient architecture based on lessons learned. This is comparable to the implementation of the NIST Cyber Security Framework process (NIST-CSF, 2014).
4. Principle 1: Classify
In some respects, cyber resilience is nothing more than an ultimate defense-in-depth and continuity management solution rolled into a highly-limited protection mission. The core aim of cyber-resilience is to maintain critical business functionality at all costs. Thus, the decisions that come out of the cyber-resilience design process will determine how the business will invest its precious time and resources and build its architecture. The key to cyber resilience is understanding what constitutes core functionality. Cyber resilience assumes that all systems will eventually be compromised. Given this assumption, the cyber resilience function ensures a robust array of specifically targeted controls to ensure that only the subset of functions essential to the continuing operation of the business are completely protected, even if all other system activities are compromised. In conjunction with the goal of ensuring the survival of core functionality, the cyber resilience process also defines clear, straightforward and practical paths to restore any lower priority functions that might have been lost in the actual compromise. The seven stages of cyber resilience are designed to achieve those two specific goals.
All computerized systems are complex and highly interdependent so, the essence of success lies in the identification of just those fundamental processes and relationships that deliver a given desired outcome. The identification process is perhaps the most important step in creating a cyber-resilient organization because the outcome of the classification process will drive every subsequent protection action. In practical terms, if an organization does not understand what assets they have, it is almost impossible to intelligently protect them. Only eighteen percent of organizations reported being effective in classifying mission critical information assets (EY, 2014). The majority of surveyed organizations reported that they were minimally effective or did not know how effective they were (EY, 2014). Thus, a deliberate, formally executed and documented classification activity is the key starting point. This is an organization-wide exercise whose aim is to understand the criticality, sensitivity, and priority of all items in the asset base. It involves all stakeholders because buy-in is an essential condition for embedding changes in the organization.
5. Principle 2: Risk
Risk assessment provides timely and accurate understanding of the threat status of all components of the asset base and is essentially a situational awareness function. The Risk stage employs situational awareness practices to drive the decisions about the best way to ensure critical assets and services will continue to function as desired. The aim is to fully understand every conceivable hazard in the threat environment that might affect a critical asset. The term "hazard" denotes a threat or an incident, natural or manmade, that warrants action to protect against harm and to minimize disruptions of the mission. This includes natural disasters, cyber incidents, acts of terrorism, sabotage, and destructive criminal activity targeting critical components of the enterprise infrastructure (PDD-21, 2014).
The outcome of this phase is a detailed map of the threat environment sufficient to support decision making with respect to organizational priorities. No decisions are made until the entire threat assessment is mapped. There is obviously a potential that meaningful threats might be missed, or that a new threat might appear after the original risk assessment is completed. Therefore, there also should be a comprehensive plan to ensure subsequent systematic risk assessments against any potential attack that might occur against assets viewed as critical. Since the architecture will be altered to respond to those threats it is also crucial that a process exists to rapidly respond to those risks. The outcome is reasonable confidence that all conceivable risks to a given asset and its dependencies have been identified, characterized and ranked for likelihood and impact.
6. Principle 3: Rank
This is the second most important aspect of the cyber-resilience process. Once the organization's assets have been identified and baselined and the threat environment characterized, the criticality of all assets in the asset baseline is ranked. This is an organization-wide ranking process involving all stakeholders. "Assets" comprise all of the people, processes, technologies, and facilities required to achieve the organizational purpose. However, some assets are more critical than others. The ranking process identifies, documents, and assures only those assets ranked as "critical" to the organization's mission, vision, values and purposes (PDD-21, 2014).
Unfortunately, ranking can often turn into a political free-for-all where various stakeholders attempt to enforce their own agendas. Obviously, this can't be allowed to happen, if the eventual architectural solution is going to be truly resilient. Therefore, criticality must be understood based on a clear map of functions and dependencies, which are referenced in an objective and rational way to the mission and goals of the organization. Therefore, an asset can only be labeled "critical" if it provably underwrites some aspect of the organization's core functionality.
A rigorous set of protection requirements are specified for just those assets that directly enable the organizational mission. Rigor is defined as the ability to resist any known or conceivable method of attack (PDD-21, 2014). Relevant stakeholders are assigned to supervise and maintain each asset. Effective communications linkages are established between those stakeholders and documented. Then the protection requirements, access links, and the requisite permissions are enabled as a coherent set of electronic and behavioral controls.
7. Principle 4: Design/deploy
Design deploys the controls required to ensure a critical asset. This is a strategic governance process. Design creates an infrastructure of substantive controls to effectively satisfy its stated mission, goals and objectives. Thus, this phase identifies the explicit control objectives for each critical asset. It prioritizes those objectives and implements targeted control actions to most effectively achieve priority objectives. Then it analyzes and assesses the deployed control set to ensure that the resultant infrastructure satisfies the critical purpose. If documented control objectives are not satisfied, then the Design process undertakes the necessary analysis to modify controls, or plug gaps.
8. Principle 6: Recover
Organizations need to understand how resilient its critical services are. This is essentially the continuity management principle. The goal of recovery planning is to ease the impact of disruptive events by using well-established plans to ensure predictable and consistent continuity of the critical services (PDD-21, 2014). To do this, the critical service's operating environment is studied to identify all potential failure modes and then a proper strategy to recover from all possible breakdowns, or disruptions is devised.
The goal is to create a complete and consistent recovery process that will address all conceivable types of system compromise. The plan for incident recovery must be explicit for every asset and lessons learned are compiled to develop improvement strategies. This requires an operational plan capable of identifying, analyzing, responding to, escalating and learning from all adverse incidents. A well-defined process for assigning roles and responsibilities and managing and tracking resolutions.
9. Principle 7: Evolve
The Evolve stage serves as the formal basis for identifying and deploying process and technology responses and improvements across the organization. This evolution is required to meet the organization's cyber resilience goals as the threat picture changes. In this stage, measurable improvements that could increase the resilience of critical assets are identified, analyzed, and systematically deployed. The effects of currently deployed processes and technology improvements are measured and the effectiveness of the selected process improvement is characterized. The five functions that must be executed mirror those of the NIST Cyber Security Framework (NIST-CSF, 2014).
Evolution is driven by the collection and analysis of data from lessons learned about the operation of the day-to-day execution of the resilience process. Improvement recommendations are supported by data obtained from the deployment of prior process and technology controls. Nevertheless, because this is essentially a "maintenance," activity this type of analysis involves ongoing testing and risk estimation. Lessons-learned typically involve objectively evaluating the performance of deployed processes against plans, objectives, standards, and procedures; as well as the outcomes of organizational innovation and deployment process.
10. Conclusion
We understand that systems will never be 100% secure and there are simply too many threats to ever make that declaration. Organizations are far more exposed to a wider range of unknown, or even unknowable, attacks (EY, 2014). Given the existence of a myriad of unknowable elements in the cyber-threat environment, organizations must be able to detect and respond creatively to threats to their basic operation. The response encompasses all forms of attack; electronic, behavioral, and physical. Thus, the concept of cyber-resilience goes far beyond the classic boundaries of better hardware and software access controls (EY, 2014). Instead, organizations need to establish a "cyber resilience strategy" that gives them ability to withstand and recover rapidly from disruptive events (EY, 2014).
Practically speaking, organizations have to start rethinking strategically and given the large number of threats and the limited resources available to counter them, concentrate their resources where they will make the most difference. This means that businesses should stop viewing security as an access control and incident response problem. Instead they should approach the problem as a comprehensive, fully integrated, strategic management commitment; aimed at ensuring survival and easing recovery. This goal requires that a robust set of concrete controls are embedded in the enterprise architecture. The umbrella term for this approach is, "cyber resilience"
Cyber resilience recognizes that adversaries will succeed. There are too many advanced hacking tools to prevent sophisticated attackers from finding the cracks in even the most robust cyber-security system (Lois, 2015).
Accordingly, cyber resilience must be an organizational condition rather than a deployed set of explicit countermeasures. Cyber resilience should be integrated into the "bones" of the organization and requires the organization to spend whatever it takes to develop a well-defined, explicit architecture of controls to ensure cyber resiliency. The controls must assure provable protection of core functionality and its various interdependencies to include every form of electronic, human and physical threat inside and outside the enterprise's eco-system (EY, 2014).
Cyber resilience is not a traditional defensive stance focused on detecting network intrusions or fixing malware. Instead, the resilience concept emphasizes practical survival and recovery of the system in the face of all conceivable forms of attack. Effective implementation of a cyber-resilient enterprise architecture requires strategic vision. It also requires day-to-day engagement across the enterprise to assure and optimize that vision. Unfortunately, this perspective is new, and it will require a change in culture to accomplish.
In general, our belief is that the only way organizations are going to effectively prevent the inevitable digital Pearl Harbor is to amalgamate cyber security and cyber resilience concepts into a single strategy. The increased presence of advanced cyber threats makes it inevitable that all targeted organizations will ultimately be compromised (Trend Micro, 2015). Nevertheless, at its fundamental root the organization needs to survive and every organization's critical systems must be assured safe, assuming any realistic set of conditions. Lost data or ancillary functionality can be restored after the attack, but the key functions of the organization must be protected. Cyber resilient organizations can "tough-out" cyber-attacks and continue to accomplish the fundamental mission. Therefore, it is incumbent on decision makers to redesign or update their essential systems for resilience. This condition should make attacks less likely to succeed, minimize the consequences when they do succeed, increase the work-factor cost and uncertainty for the adversary, and possibly act as a deterrent against future attacks. Consequently, investment in building cyber resilience is strategic investment in organizational survival.