Flooding Peer-to-Peer (P2P) networks form the basis of services such as the electronic currency system Bitcoin. The decentralized architecture enables robustness against failure. However, knowledge of the network's topology can allow adversaries to attack specific peers in order to, e.g., isolate certain peers or even partition the network. Knowledge of the topology might be gained by observing the flooding process, which is inherently possible in such networks, and performing a timing analysis on the observations. In this paper we present a timing analysis method that targets flooding P2P networks and show its theoretical and practical feasibility. A validation in the real-world Bitcoin network proves the possibility of inferring network links of actively participating peers with substantial precision and recall (both ∼ 40 %), potentially enabling attacks on the network. Additionally, we analyze the countermeasure of trickling and quantify the tradeoff between the effectiveness of the countermeasure and the expected performance penalty. The analysis shows that inappropriate parametrization can actually facilitate inference attacks.

Permissionless blockchains reach decentralized consensus without requiring pre-established identities or trusted third parties, thus enabling applications such as cryptocurrencies and smart contracts. Consensus is agreed on data that is generated by the application and transmitted by the system's (peer-to-peer) network layer. While many attacks on the network layer were discussed so far, there is no systematic approach that brings together known attacks, the requirements, and the design space of the network layer. In this paper, we survey attacks on the network layer of permissionless blockchains, and derive five requirements: 1) performance; 2) low cost of participation; 3) anonymity; 4) DoS resistance; and 5) topology hiding. Furthermore, we survey the design space of the network layer and qualitatively show the effect of each design decisions on the fulfillment of the requirements. Finally, we pick two aspects of the design space, in-band peer discovery and relay delay, and demonstrate possible directions of future research by quantitatively analyzing and optimizing simplified scenarios. We show that while most design decisions imply certain tradeoffs, there is a lack of models that analyze and formalize these tradeoffs. Such models could aid the design of the network layer of permissionless blockchains. One reason for the lack of models is the deliberately limited observability of deployed blockchains. We emphasize that simulation based approaches cope with these limitations and are suited for the analysis of the network layer of permissionless blockchains.

We present a simulation model of the Bitcoin peer-to-peer network, a widely deployed distributed electronic currency system. The model enables evaluations of the feasibility and cost of attacks on the Bitcoin network at full scale of 6,000 nodes. The simulation model is based on unmodified code from core segments of the Bitcoin reference implementation used by 99% of nodes. Parametrization of the model is performed based on large-scale measurements of the real-world network. We present preliminary validation results showing a reasonable correspondence of the propagation of messages in the Bitcoin network compared with simulation results. We apply the model to study the feasibility of a partitioning attack on the network and show that the attack is sensitive to the churn of the attacking nodes.

Bitcoin relies on a peer-to-peer network for communication between participants. Knowledge of the network topology is of scientific interest but can also facilitate attacks on the users' anonymity and the system's availability. We present two approaches for inferring the network topology and evaluate them in simulations and in real-world experiments in the Bitcoin testnet. The first approach exploits the accumulation of multiple transactions before their announcement to other peers. Despite the general feasibility of the approach, simulation and experimental results indicate a low inference quality. The second approach exploits the fact that double spending transactions are dropped by clients. Experimental results show that inferring the neighbors of a specific peer is possible with a precision of 71 % and a recall of 87 % at low cost. 2 Related Work Topology inference in Bitcoin has been the subject of several previous works.

Address clustering tries to break the privacy of bitcoin users by linking all addresses created by an individual user, based on information available from the blockchain. As an alternative information source, observations of the underlying peer-to-peer network have also been used to attack the privacy of users. In this paper, we assess whether combining blockchain and network information may facilitate the clustering process. For this purpose, we apply all applicable clustering heuristics that are known to us to current blockchain information and associate the resulting clusters with IP address information extracted from observing the message flooding process of the bitcoin network. The results indicate that only a small share of clusters (less than 8 %) were conspicuously associated with a single IP address. Also, only a small number of IP addresses showed a conspicuous association with a single cluster.