Analyzing Eclipse Attack on Proof of Stake (PoS) Network

Source

INTRODUCTION

An eclipse attack means an intrusion into blockchain systems where all the nodes of a certain target node or nodes are systematically closed off from the entire network. An eclipse attack may work as a denial of service (DoS) attack to a node due to repositioning of that nodes victim. This happens in Proof of Stake (PoS) networks where participants are relied on to endorse transactions and blocks for their verification. Out of all the alternative consensus events, the PoS seems to attract the eclipse attacks as well as other forms of sybil attacks.

The attack generally consists control the incoming and outgoing connection of a node to the outside world. Doing this will distance the node from the accurate view of the network and feed the node with incorrect information only. In PoS networks where the validators are elected by the magnitude of the stake, this management can be used in a pernicious way and induce the nodeto endorse malicious blocks or to participate in double spend attacks. This does not bode well cost wise for PoS systems’ security and efficiency as well.

The most important risk of an eclipse attack in a PoS network is that it can be executed in a destructive manner with minimal resources or finances. As PoS is more dependent on communication as opposed to brute-forcing the network the success of an eclipse attack depends on controlling the network as a layer, which can be easier to exploit than attacking the consensus mechanism directly. Understanding how this attack works is crucial for improving the security of PoS systems.

• ISOLATING THE TARGET NODE:

The first step in an eclipse attack consists in cutting the targeted node off from the rest of the network. An attacker has to take control of each and every connection of the target node in such a manner that communication with any benign nodes is no longer possible. This can be achieved by saturating the target with lots of fake IP addresses or utilizing some network loopholes to drop honest peers’ connections.

Just as the node has been cut off from its peers and made passive, that node cannot and may never until new instructions are given by the attacker receive any additional new blocks or transactions, thus being cut off from new transactions. Now, the isolated node is only accustomed to the news from the attacker, which places it at risk of information deception. This isolation makes it simple for the attacker to distort the node’s understanding of how the network infrastructure’s state is.

In a PoS network, where the validators are chosen in terms of the stakes put across, an isolated node for instance by another node may be spoon-fed information about the consensus status at the moment and the spread of such fakes may lead the node to comit validating untrue blocks and not validating true blocks, or some honest nodes may even decide not to be validators altogether. This can hamper the ultimate consensus that speaks to the aspect of eliminating excessive or fraudulent information from the control of the blockchain.

• MANIPULATING CONSENSUS:

After the remainder of the network is severed from the target node which has now been isolated, the attacker is then able to determine whether this node will be active in the consensus or not. In PoS networks, validators are chosen according to their stake being met, and they adduce a major responsibility for the attestation of new blocks application and the overall security of the network. The connections of an attacker who controls a node can shove improper blocks to that node as a result of the node authorising some transaction which is not part of the central blockchain.

This manipulation can lead to several issues, including double-spending attacks, whereby the same assets are incurred more than once. Since the compromised node has a false sense of working with useful information, such further false attacks can be foisted upon this innocent node. Such propagation can also allow the attacker to use the node for building forks ad which can affect the network consensus and the attack chain could split into strands.

This ability to control the participation of the validator in the consensus process places the attacker in a comfortable position deep within a PoS network. Even though other validators are left to work normally, totally honest, the acts performed by a compromised node still have a potential reaction throughout the network , rather degrade the confidence level attached on the system.

Source

• EXPLOITING NETWORK DELAYS:

An additional important feature of the eclipse attack is the use of latency as a force multiplier. Day by day, nodes in a decentralized network must communicate to reach an agreement within a given time-frame. If an attacker eliminates a node and then determines the time of the attacking node, the attacker can delay block or transaction validation, which would reduce the speed of the network.

In any PoS network, which is designed to eliminate the inefficiencies of PoW systems, these delays can be extremely detrimental. A delayed node may abuse such an opportunity and as a result, painlessly skip some consensus rounds resulting in subject’s staying outside as far as contributing to the network is concerned. This can also provide the attacker with further margins within which to undertake other malevolent activities such as double spending or fork network creation.

By organizing methods that induce hurdles, the victim can create further obstacles to the node that limits it from active involvement in the consensus process. Eventually, this can undermine the efficiency of the consensus, making validators and ordinary people unbeneficial from the network.

• PREVENTING ECLIPSE ATTACKS

Reducing the threats of eclipse attacks is related to increasing the degree of network defences against isolation. One means is to make the maximum number of peers available to a node. The limitation of this method is that one assailant would not be able to monopolise all peers. Similarly, attacks will not be successful when a random peer selection is adopted.

By employing additional measures, it is also possible to improve resilience to such attacks when abusing system features, for example, strong cryptography of communications between nodes prevents even isolated nodes from accepting all received data as correct. A PoS network also enables such monitoring where certain thresholds are applied to detect mutant activities, loss of connections to honest peers being one of such activities.

Based on the above it is possible to conclude that investing effort to prevent eclipse attacks altogether is rather imprudent, for one layer of security can mitigate only one risk. Use it as a supplement to existing theories targeting these insidious risks. Improvements are expected at both network level security and consensus protocols in PoS systems to defend against these unsought onslaughts.

CONCLUSION

Source

Eclipse attack is a well-known issue in PoS networks, which allows manipulation of a particular node's view and consequently the consensus outcome. As a result of such concentration and misinformation, attackers can distort and break the security of the network.