Sybil Attack

A Sybil attack is what happens when one person or entity creates many fake identities to gain disproportionate influence over a system that assumes "one identity, one participant." The name comes from a 1973 book about a woman diagnosed with dissociative identity disorder, the metaphor being one underlying person presenting as many separate ones.

This matters anywhere influence, votes, or trust scale with the number of identities a system sees, rather than with something scarce and hard to fake. A reputation system where reviews from more accounts count more is trivially gamed by one person running a thousand fake accounts. A peer-to-peer network where routing decisions trust "the majority of my neighbors" can be quietly surrounded by an attacker who controls every neighbor under different names. A blockchain where block-production rights are shared equally among "validators" is Sybil-vulnerable the moment creating a new validator identity is free, an attacker just creates enough of them to control the vote.

Who proved this, and what exactly was proved

John R. Douceur formalized the problem in a 2002 paper, "The Sybil Attack," and proved something more precise and more pessimistic than "this is hard to stop": absent a trusted central authority that vets and issues identities, or some mechanism that makes each new identity genuinely costly to obtain, a resource-bounded attacker can always create enough fake identities to overwhelm any assumption based on identity count alone. That result is why every serious peer-to-peer or blockchain system has to build in some cost to identity creation, there's no free lunch here, only different ways of paying for it.

Three broad families of payment have emerged across the industry: proof-of-work (Bitcoin's answer, identity/influence costs real energy), proof-of-stake (identity/influence costs bonded capital), and identity-bound approaches (identity/influence costs a scarce, hard-to-fake proof of being a distinct human. Proof of Humanity and Worldcoin are two prior examples, each solving it for a different purpose: a UBI-style token registry and Sybil-resistant airdrop distribution, respectively).

Where Solidus sits

Solidus runs the first two layers of defense in different states of completeness today. Stake-based resistance is live: becoming a validator on the public testnet requires posting a minimum stake, enforced in running code, though that stake is a testnet-internal accounting unit with no real economic value yet, since no $SLDS token has ever been publicly traded (see Validator for the full detail). Identity-bound resistance, the third family, and the one novel piece Solidus adds, is specified but not enforced. Proof-of-Identity (the one term in this entire Lexicon marked invented) proposes gating validator eligibility on a verified-human credential in addition to stake, so that accumulating influence requires acquiring both capital and distinct verified identities. Solidus's own architecture documentation states plainly that this identity gate is "part of the roadmap validator-admission model, not enforced on the current testnet": nothing in the running code today checks that a validator operator holds a credential distinct from any other validator's.

Solidus's own written threat model (docs/protocol.md §13, see Threat Model) names "Sybil attack on validators" as an assumed adversary capability and lists "KYC Level 2 requirement" as the mitigation, accurately describing the specified design, not yet the running one.

Check it yourself

The stake-gate is source-visible in the public solidusnetwork/protocol repository's solidus-txns crate, and rpc.solidus.network is a live, public endpoint confirming today's small validator committee. There is nothing to check for the identity-bound layer yet, because it isn't running, see Proof-of-Identity for why that gap is named directly rather than implied away.