Trust Factors and Insider Threats in Permissioned Distributed Ledgers

Abstract

Permissioned distributed ledgers have recently captured the attention of organizations looking to improve efficiency, transparency and auditability in value network operations. Often the technology is regarded as trustless or trust-free, resulting in a false sense of security. In this work, we review the various trust factors present in distributed ledger systems. We analyze the different groups of trust actors and their trust relationships to the software layers and inherent components of distributed ledgers. Based on these analyses, we investigate how insiders may conduct attacks based on trust in distributed ledger components. To verify practical feasiblity of these attack vectors, we conduct a technical study with four popular permissioned distributed ledger frameworks: Hyperledger Fabric, Hyperledger Sawtooth, Ethereum and R3 Corda. Finally, we highlight options for mitigation of these threats.

Detecting Blockchain Security Threats

“…Internal threats. Distributed Ledgers are subject to a variety of threats by insiders and external insiders [17], who may attempt to exploit the blockchain for personal or organizational gain. Besides inadvertent misconfigurations, insiders may intentionally manipulate the configuration of blockchain peers and threaten network security.…”

“…A major category of dependencies are database systems used for storage of blockchain state. Most blockchain frameworks use self-sufficient DBMS such as LevelDB, CouchDB or Postgres [17]. As dependencies, their versions are often updated infrequently and insufficient default configurations are used.…”

“…Reentrancy [19], delegatecall [13], Dependency injection [17] scanned potential vulnerabilities threat intelligence on vulnerabilities [20], [22], Sybil attacks [18], [20], [23], Boycott attack [20], Blacklisting attack [20] certificate requests (successful/denied) certificate revocations transactor identities…”

“…Reactive authentication and non-repudiation. If SHA256 were to be affected by a collision-resistance attack similar to the one discovered for SHA-1 (CVE 2005-4900), most major blockchain frameworks would be affected [17]. Detecting such an attack is only possible by monitoring threat intelligence, i.e.…”

“…To achieve DoS, an outsider may attempt to flood specific or multiple blockchain peers with TCP syn packets. In particular, DoS attacks targeting the consensus leader can significantly reduce or stall consensus entirely [17]. Internal attackers with access to the network can simply send a large number of transactions (or transactions with a large size), which must all be processed by endorsing peers [4].…”