Why Is Cybersecurity So Hard?
AFTER NEARLY 20 YEARS OF TRYING AND BILLIONS OF DOLLARS IN INVESTMENT, WHY ARE ORGANIZATIONS ARE STILL STRUGGLING WITH CYBERSECURITY?
After nearly 20 years of trying and billions of dollars in investment, why are organizations are still struggling with cybersecurity? In fact, the problem seems to be getting worse, not better. Answering this question requires moving beyond a purely technical examination of cybersecurity.
It’s true that the technical challenges are very real; we don’t know how to write bug-free code, for example. But if you look at the challenge more broadly, even if we resolved the technical issues, cybersecurity would remain a hard problem for three reasons:
- It’s not just a technical problem
- The rules of cyberspace are different from the physical world’s
- Cybersecurity law, policy, and practice are not yet fully developed
Differing Rules in Cyberspace
Cyberspace operates according to different rules than the physical world. I don’t mean the social “rules” but rather the physics and math of cyberspace. The nodal nature of a light-speed network means that concepts like distance, borders, and proximity all operate differently, which has profound implications for security. First, with distances greatly reduced, threats can literally come from anywhere and from any actor. Second, the borders in cyberspace don’t follow the same lines we have imposed on the physical world; instead they are marked by routers, firewalls, and other gateways. Proximity is a matter of who’s connected along what paths, not their physical location.
As a result, our physical-world mental models simply won’t work in cyberspace. For example, in the physical world, we assign the federal government the task of border security. But given the physics of cyberspace, everyone’s network is at the border. If everyone lives and works right on the border, how can we assign border security solely to the federal government? In the physical world, crime is local — you have to be at a location to steal an object, so police have jurisdictions based on physical boundaries. But in cyberspace you can be anywhere and carry out the action, so local police jurisdictions don’t work very well.
The same principles of cyberspace that allow businesses to reach their customers directly also allow bad guys to reach businesses directly. Yet you can’t have governments get in the way of the latter without also getting in the way of the former. Sharing information among people at human speed may work in many physical contexts, but it clearly falls short in cyberspace. As long we continue to try to map physical-world models onto cyberspace, they will fall short in some fashion.
Legal and Policy Frameworks
Next, cyberspace is still very new from a legal and policy point of view. In the modern form, the internet and cyberspace have existed for only about 25 years and have constantly changed over that time period. Therefore, we have not developed the comprehensive frameworks we need. In fact, we don’t yet have clear answers to key questions:
- What is the right division of responsibility between governments and the private sector in terms of defense?
- What standard of care should we expect companies to exercise in handling our data?
- How should regulators approach cybersecurity in their industries?
- What actions are acceptable for governments, companies, and individuals to take and which actions are not?
- Who is responsible for software flaws?
- How do we hold individuals and organizations accountable across international boundaries?
Some answers are beginning to emerge. For example, we should not expect the federal government to protect every business from all online threats all the time — it’s simply not practical, nor is it desirable, because it would significantly impact the way we’re able to do business. On the other hand, we can hardly expect most organizations to thwart the activities of sophisticated nation-state actors. So how do we resolve this dilemma?
Perhaps we should borrow concepts from the disaster response world, and divide responsibility in a fluid manner that adapts over time in response to changing circumstances. In disaster response, preparedness and initial response reside at the local level; if a given incident overwhelms or threatens to overwhelm local responders, then steadily higher levels of government can step in. We could apply these principles to allocating responsibility in cyberspace — businesses and organizations remain responsible for securing their own networks, up to a point. But if it becomes clear that a nation-state is involved, or even if the federal government merely suspects that a nation-state is involved, then the federal government would start bringing its capabilities to bear. Fully answering these questions is the key cybersecurity policy task for the next five to 10 years.
As long as we treat cybersecurity as a technical problem that should have easy technical solutions, we will continue to fail. If we instead develop solutions like cybersecurity training that can simplify cybersecurity, then we will make progress. A little over two years ago, a group of cybersecurity practitioners from several organizations concluded that the industry’s operational model was not producing the desired results and decided to adopt a new one — to work together in good faith to begin sharing threat information in an automated fashion, with everyone contributing to the system, and with the context of threats being given a lot more weight. CTA’s structure is an attempt to deal with the known flaws in existing information sharing efforts. If we can continue to innovate in this manner, we can finally begin to make some progress against this seemingly intractable problem.
This article was originally published on HBR by Michael Daniel. He is the president of The Cyber Threat Alliance, where he focuses on improving the U.S.’s cybersecurity. He has more than 20 years of experience working with the federal government in this effort, including five years as the nation’s cybersecurity coordinator.