Lag-Erasure for Zero-Lag Distributed Simulation

Dr. Loren Peitso Graduation

Loren Peitso, Ph.D. Dissertation, 27 March 2026

Abstract. Communications latency generates lag within distributed simulations, creating mismatched expectations and causality violations that prevent repeatable, verifiable states and reduce confidence in simulation progression and final results. To train and build trust in both humans and complex artificial intelligence–enabled systems, it is often necessary to run simulations that exceed a single cluster site’s capabilities or require specialized hardware from geographically distributed sites. Lagging simulations are incompatible with this training. Lag is documented in the literature as unavoidable, leading to weak mitigation and acceptance rather than elimination. This work shows that the assumption of unavoidable lag is implementation-dependent (not absolute), enabling the design of a simulation architecture that limits, then erases lag effects from a distributed simulation’s state. Following a root-cause analysis of lag, techniques from both the analytical and virtual environments simulation communities are hybridized to define a software architecture for delivering lag-free distributed simulations. The affected scope is widespread—from command and control systems and Live Virtual Constructive training, to online multiplayer gaming and shared massively multi-user virtual/augmented reality metaverse applications. Lag-free distributed simulations are repeatable and verifiable, enabling a level of confidence in simulation results previously unreachable.

Conclusions

  • The Consistency-Throughput Tradeoff remains relevant, but the impossibility of dynamic shared state maintenance is put in a box in the corner of a now greatly enlarged simulation operating environment.
  • Consistent distributed shared state is now possible.

Recommendations for Future Work

  • Near term:  Perfect Playback Event Stream Archiving, Networked Implementation Testing, Lag Erasure capability expansion
  • Long term:  Implement multiple use cases, Formalizing arbitrary branching What-if? explorations, Online Games – Security Through Ubiquity

Resources

Reference

Singhal, Sandeep and Michael Zyda, Networked Virtual Environments. New York, NY, USA: Addison-Wesley Professional, 1999. Available via https://dl.acm.org/doi/book/10.5555/316724