What Is Voting?

Voting, in the context of engineering and computing systems, is a decision mechanism used in fault-tolerant architectures to produce a single correct output from the results of multiple independent computational modules. When two or more redundant modules process the same input, a voter circuit or software component compares their outputs and selects or synthesizes a result according to a specified rule, most commonly majority agreement. The approach allows a system to continue operating correctly even when one or more of its constituent modules has failed silently or produced an erroneous result.

Voting-based redundancy has been deployed in safety-critical systems since the early days of digital avionics and space computing in the 1960s. The approach draws on reliability engineering, probability theory, and digital circuit design, and its principles apply equally to hardware voter circuits, software-implemented voters in embedded systems, and distributed consensus protocols in networked computing.

Majority Voting and Triple Modular Redundancy

The most widely deployed voting scheme is the simple majority vote used in triple modular redundancy (TMR). In a TMR system, three identical modules execute the same function in parallel, and their outputs feed into a voter that produces the result agreed upon by at least two of the three. A single module failure is fully masked because the two functioning modules outvote the faulty one. The voter circuit itself, being far simpler than the guarded modules, has a substantially lower failure rate, though safety-critical designs sometimes triple the voter as well. TMR is the basis of flight control computers in commercial aircraft and of spacecraft processors such as those described in IEEE Xplore publications on fault-tolerant voter design for TMR circuits. N-modular redundancy (NMR) generalizes the concept to any odd number of modules, tolerating up to (N-1)/2 simultaneous failures.

Voting in Distributed and Software Systems

Software-implemented voting extends fault-tolerant principles to multiprocessor and networked architectures where hardware voter circuits are impractical. In this setting, voter logic runs as a software process that collects the outputs of replicated tasks or nodes and applies a comparison function before passing the result to dependent system components. The Byzantine fault tolerance problem, formalized by Lamport, Shostak, and Pease in 1982, addresses the harder case in which faulty nodes may produce arbitrary or deceptive outputs rather than simply crashing. Byzantine-tolerant protocols require at least 3f + 1 nodes to tolerate f malicious actors, a bound that shapes the architecture of modern distributed ledgers and safety-critical control networks. Reliability analyses of redundant systems, including probabilistic failure models for TMR, are documented in peer-reviewed studies such as reliability analyses published in Nature Scientific Reports.

Electronic Voting Systems

Electronic voting encompasses direct-recording electronic (DRE) devices, optical-scan systems, and internet-based voting platforms used in civic and organizational elections. These systems must satisfy security properties including ballot secrecy, verifiability, and integrity against manipulation. End-to-end verifiable (E2E) schemes allow voters to confirm that their ballot was recorded and counted correctly without revealing how they voted. The security of electronic voting infrastructure intersects with cryptographic protocol design, side-channel analysis of embedded hardware, and network security, all active areas of IEEE research. The NIST Voting Program supports the development of Voluntary Voting System Guidelines (VVSG) and conducts research on testing, usability, and security for election technology.

Applications

Voting has applications in a wide range of disciplines, including:

• Avionics flight control computers and autopilot systems
• Space mission processors subject to radiation-induced single-event upsets
• Nuclear plant safety instrumentation and control systems
• Distributed database replication and consensus protocols
• Electronic election administration and audit systems