A Sybil attack is an attack in which one real entity creates many fake identities and uses them to influence a distributed system.
The name comes from the idea of one person appearing as many people. In a network, forum, voting system, blockchain or peer-to-peer system, the attacker may register many accounts, nodes or addresses. If the system treats each identity as an independent participant, the attacker can gain more influence than they should.
Why It Is Dangerous
Many distributed systems assume that “more participants” means “more independent opinions”. A Sybil attacker breaks this assumption.
For example, an attacker can:
- create many fake accounts to manipulate voting;
- run many fake nodes to influence peer discovery;
- generate many blockchain addresses to farm airdrops;
- flood a reputation system with fake reviews;
- make a small group look like a large community.
The core problem is that identities are cheap to create, but the system gives each identity value.
Common Defenses
There is no universal solution. Different systems increase the cost of identity in different ways:
- proof of work: identities require computing cost;
- proof of stake: identities require locked capital;
- account verification: identities require real-world proof;
- reputation systems: influence grows slowly over time;
- rate limits: new identities cannot act too quickly;
- graph analysis: suspicious clusters can be detected.
Each defense has trade-offs. Strong verification improves resistance but hurts privacy. Proof of work wastes resources. Proof of stake favors users with more capital.
In Blockchain Systems
Sybil resistance is central to blockchain design. If one computer could create unlimited voting nodes for free, consensus would be easy to manipulate.
Bitcoin uses proof of work to make influence depend on hash power rather than account count. Proof-of-stake systems use locked stake. Airdrop projects often add behavior analysis or identity checks to reduce fake accounts.
Summary
A Sybil attack is not about exploiting a software bug. It exploits weak identity cost. When creating many identities is cheap and each identity receives trust or reward, the system is vulnerable.
The key design question is: how much does it cost to become “one participant”?