Research commentary · Security economics
The price of keeping a botnet.
A compromised machine is not a free resource to its controller. The effort required to retain and organise access changes the economics of criminal networks, and the choices available to defenders.
A count of infected devices describes the size of a problem. It does not, on its own, describe the value that an offender can extract from it. A large collection of unreliable, short-lived access can be less useful than a smaller collection that remains available. The distinction connects a familiar security measure to an economic question: what does it cost to keep control?
Control has a running cost
My 2012 conference paper, Territorial behavior and the economics of Botnets, considers the relationship between the benefits of controlling a criminal network and the time and costs involved in maintaining it. The public abstract frames territorial size as a profit decision and connects that decision to defensive strategies intended to reduce criminal returns.
The underlying question differs from asking how quickly malicious software can spread. Acquiring access and retaining useful access are separate activities. Growth can increase potential revenue while also increasing the burden of maintaining the operation. A measure of scale becomes more informative when accompanied by an account of those continuing demands.
A hypothetical service comparison
Consider two fictional organisations, each of which discovers the same number of compromised endpoints during a month. The first restores affected systems promptly and closes the route through which access was regained. The second repeatedly removes visible symptoms but leaves the same systems available for renewed compromise. Their monthly discovery counts might match, yet the duration of effective criminal control can differ substantially.
This is an original illustration, not a result reported by the paper. Its purpose is to separate three observations: a device was identified as compromised, control was removed, and the same access remained unavailable afterwards. Counting the first does not establish the other two.
For a defender, the useful outcome is not a more impressive collection of discoveries. It is less harmful access over the period that matters. That can mean a shorter interval before containment, a more complete restoration, or fewer repeated compromises. Which change matters most depends on the actual service and harm involved.
Making the economic objective explicit
A defensive programme consumes resources too. Staff time spent restoring machines is time unavailable for other work. It therefore helps to state the intended mechanism rather than treating every activity labelled security as interchangeable. One measure may reduce the opportunities for initial compromise. Another may shorten the useful life of access already obtained. A third may reduce the harm that can be inflicted while recovery proceeds.
These mechanisms can complement one another without sharing a single natural unit. A faster response and a lower reinfection rate should not be collapsed into one count simply because both appear on a dashboard. Keeping their meanings distinct allows a manager to explain what changed and why the expenditure was made.
From network size to sustained harm
The economic perspective does not excuse criminal behaviour or turn a conceptual model into a complete description of every attacker. It gives the defender a more precise question. Instead of asking only how many machines appear in a network, ask how much useful control persists, how long it persists, and which defensive action reduces the resulting harm.
The practical connection between economics and security lies there: resources, incentives and duration matter alongside technical access. A network has to be sustained to be useful. Defence can be understood partly as making that sustained control harder to retain, while protecting legitimate users and restoring their systems.