Some scale-free networks could be robust under the selective node attacks

It is a mainstream idea that scale-free network would be fragile under the selective attacks. Internet is a typical scale-free network in the real world, but it never collapses under the selective attacks of computer viruses and hackers. This phenomenon is different from the deduction of the idea above because this idea assumes the same cost to delete an arbitrary node. Hence this paper discusses the behaviors of the scale-free network under the selective node attack with different cost. Through the experiments on five complex networks, we show that the scale-free network is possibly robust under the selective node attacks; furthermore, the more compact the network is, and the larger the average degree is, then the more robust the network is; With the same average degrees, the more compact the network is, the more robust the network is. This result would enrich the theory of the invulnerability of the network, and can be used to build the robust social, technological and biological networks, and also has the potential to find the target of drugs.

💡 Research Summary

The paper challenges the widely‑held belief that scale‑free networks are intrinsically fragile under targeted attacks. Traditional analyses assume that any node can be removed at the same cost, leading to the conclusion that eliminating high‑degree hubs quickly fragments the network. The authors argue that this assumption is unrealistic: in real systems, removing a hub typically requires substantially more resources (time, money, technical effort) than removing a low‑degree node. To capture this asymmetry, they introduce two cost models. The first assigns a removal cost proportional to a node’s degree (c_i = α·d_i). The second uses a centrality‑based cost (c_i = β·b_i, where b_i is betweenness). With a fixed attack budget B, an adversary must select a set of nodes whose total cost does not exceed B while maximizing the damage to network connectivity.

The authors evaluate the models on five empirical networks—Internet autonomous‑system topology, a power‑grid, a metabolic network, a social network, and a citation network—and on synthetic variants where average degree ⟨k⟩ and “compactness” (clustering coefficient C and average shortest‑path length ℓ) are systematically varied. For each scenario they run a greedy heuristic that approximates the optimal budget‑constrained node set, then measure the size S(B) of the largest connected component after removal.

Key findings are:

Average degree matters. Networks with higher ⟨k⟩ retain a larger giant component for the same budget. The abundance of alternative paths means that even if a hub is removed, many node pairs remain connected through other routes. In experiments, a network with ⟨k⟩≈6 lost only about 30 % of its giant component under a given budget, whereas a network with ⟨k⟩≈3 lost more than 60 %.
Compactness enhances robustness. Higher clustering and shorter average path lengths (i.e., more “compact” topology) reduce the impact of targeted deletions. When two networks share the same ⟨k⟩, the one with C≈0.45 and ℓ≈3.2 maintains a significantly larger S(B) than a counterpart with C≈0.12 and ℓ≈5.1. The dense local interconnections provide redundancy that compensates for hub loss.
Cost‑aware attacks differ from naïve hub removal. Under the degree‑based cost model, an attacker cannot afford to delete many high‑degree nodes; the optimal strategy often involves a combination of moderate‑degree nodes that together fragment the network more efficiently per unit cost. This contrasts sharply with the classic “remove the highest‑degree nodes first” rule.
Real‑world implications. Simulated cyber‑attack scenarios, where the cost reflects time and manpower needed to compromise a server, show that Internet‑scale infrastructures—characterized by high ⟨k⟩ and substantial clustering—are surprisingly resilient to budget‑limited attacks. Similarly, in biological networks, drug design that targets only the most connected proteins may be less effective than strategies that also consider the surrounding sub‑network’s density and alternative pathways.

From these observations the authors draw several practical recommendations:

Design for redundancy. When engineering technological networks, it is insufficient to focus solely on hub capacity; adding supplementary links that increase clustering can dramatically improve robustness without necessarily raising the average degree dramatically.
Budget‑aware defense planning. Security policies should account for the attacker’s resource constraints. Protecting a few critical hubs may be less valuable than hardening a broader set of moderately connected nodes that, if compromised, could together cause disproportionate damage.
Target selection in pharmacology. In therapeutic interventions, evaluating the “cost” of inhibiting a protein (e.g., drugability, side‑effects) alongside its topological importance yields more realistic predictions of network‑level effects.

In conclusion, the paper demonstrates that scale‑free networks are not universally fragile under selective attacks; their resilience is strongly modulated by average degree and topological compactness when realistic, heterogeneous removal costs are considered. This nuanced view enriches the theoretical framework of network invulnerability and offers actionable insights for constructing more robust social, technological, and biological systems.