Open research questions in Distributed systems and fault tolerance
88 gap statements mined from Distributed systems and fault tolerance papers in our 4.5M-paper local library, which holds 594 papers on the topic — drawn mostly from each paper's own stated research gap, future-work, challenge and limitation notes, and its abstract. The ones listed below are a selection still marked open; each names the study that raised it, with a DOI link where the paper has one.
Choosing where to publish on Distributed systems and fault tolerance? See the ranked Computer Science journals.
What the literature leaves open
The paper does not provide a comprehensive comparison with other Byzantine-resilient algorithms. The experimental evaluation is limited to a simulated environment. The paper does not discuss the computational complexity of FABA.
The paper suggests further research on the application of reflective parallel algorithms. The paper suggests further research on the development of programming languages and systems that support reflection.
The paper addresses the need for a behavioural theory of reflective parallel algorithms.
The lack of effective concurrency control mechanisms in high-contention environments. The need for a method to predict and prevent transaction failures. The potential for machine learning to improve concurrency control and reduce resource waste.
We conclude the paper by listing a number of open questions we consider to be of interest. (1) Is the achieved bit complexity (nearly) optimal? (2) Is it possible to interleave our approach with randomized techniques to get both stabilization time 𝑂 (𝑓 + 1) deterministically and 𝑂 (1) in expectation?
Early-Stabilizing Counting · 2026The problem of storage access in autonomous AI systems, where storage errors are not just implementation concerns but part of the governance boundary. The need for a governed storage substrate that ensures storage safety and determinism in AI systems.
This architecture has several limitations. No independent distributed‑consistency guarantee. VFS defines semantic storage boundaries. It does not, by itself, solve distributed transactions, object‑store consistency, or network partition be‑ havior. Capability granularity and access‑control boundary.
Lack of admission control in LLM-based systems. Need for deterministic admission-control model.
Traditional offline auction frameworks have limitations such as geographic reach and time-bound schedules. Current commercial software has a trade-off between implementation complexity, runtime cost efficiency, and horizontal system scalability.
BidSphere: A Secure, Real-Time Web-Based Smart Online Auction Platform with Role-Based Access Control and High-Concurrency Bidding Architecture · 2026 · DOIEliminating the need for trust in a central authority. Achieving decentralized verification of Service Level Agreements.
The durable contributions are ActiveGraph as an auditable substrate that makes controlled improvement loops tractable, the held-out-gated loop it supports, the failure-regime taxonomy routing each failure to a pipeline location (whose marginal value over an unrouted baseline is the primary open question), and the prompt-as-discovery-probe hypothesis.
Regimes: An Auditable, Held-Out-Gated Improvement Loop Demonstrated on LongMemEval with ActiveGraph · 2026Google Safe Browsing (GSB) and DNS resolution operate concurrently during browser navigation, yet their packet-level synchronization remains understudied.
Poster: Mind the Gap -- Characterizing the Temporal Blind Spot Between GSB and DNS Resolution · 2026While our two case studies demonstrate the feasibility and scalability of our approach, the evaluation remains limited to these specific scenarios. This improves analytical precision, but the modeling effort scales with system complexity and is performed manually, which limits adoption to systems which are not well understood.
Traditional Event Sourcing architectures have limitations in handling validation rules that span across multiple aggregates. Existing solutions either expand aggregate boundaries or rely on Process Managers, sacrificing immediate consistency.
Dynamic Consistency Boundaries in Event Sourcing via Multi-Stream Optimistic Concurrency Control · 2026 · DOIThe lack of a framework for coherent transport in driven open physical systems. The need for a method to select Lambda_obs or separately validated Lambda_dyn before scoring. The absence of a locally frozen secondary-data test of P3 on a public Rijke-tube benchmark.
The unknown players challenge, where the set of participants is unknown at the time of protocol deployment and is of unknown size. The player inactivity challenge, where participants can start or stop running the protocol at any time. The sybil challenge, where one participant may masquerade as many by using many identifiers.
The paper suggests that future research should focus on developing more economically secure permissionless consensus protocols. The paper suggests that future research should explore the application of the EAAC protocol concept to real-world blockchain protocols.
To explore the application of FABA to other machine learning algorithms. To evaluate the performance of FABA in real-world distributed environments. To design more robust Byzantine-resilient algorithms.
Ensuring fault tolerance within very limited timing constraints. Dealing with transient faults, burst faults, and complete state loss in distributed real-time systems. Providing fast and stable recovery in critical applications.
A Time-Entangled Self-Reconstructing Framework for Fault Tolerance in Distributed Real-Time Systems · 2026 · DOIThe system presently removes older round accounts after a limited duration to reclaim rent costs. The protocol introduces additional system complexity and coordination overhead.
Evaluating the effectiveness of the hybrid approach in real-world scenarios. Exploring the use of other machine learning algorithms for concurrency control. Investigating the application of the hybrid approach to other database systems and workloads.
Further confirmation of the effectiveness of the CW framework through comparison with state-of-the-art methods, - Exploration of the application of CW to various types of faults, such as transient disturbances, burst faults, and complete state…
A Time-Entangled Self-Reconstructing Framework for Fault Tolerance in Distributed Real-Time Systems · 2026 · DOIDistributed transaction management in multi-tenant cloud platforms is challenging due to the lack of a single global commit protocol - The research problem concerns architectural control over subscription lifecycle operations that cross service…
Architectural Patterns for Distributed Transaction Management in Multi-Tenant Cloud Platforms · 2026 · DOIEmpirical performance benchmarks comparing execution speeds and hardware efficiency under high concurrency. Further evaluation of the proposed approach in a variety of application domains.
Dynamic Consistency Boundaries in Event Sourcing via Multi-Stream Optimistic Concurrency Control · 2026 · DOITo apply the Lambda-Flight framework to different physical systems. To evaluate the framework using different types of coherence factors. To develop new methods for selecting Lambda_obs or separately validated Lambda_dyn.
Most-cited papers in Distributed systems and fault tolerance
- Time, clocks, and the ordering of events in a distributed system · Communications of the ACM · 1978 · 6,102 citations
- On data-limited and resource-limited processes · Cognitive Psychology · 1975 · 1,617 citations
- Self-stabilizing systems in spite of distributed control · Communications of the ACM · 1974 · 1,553 citations
- The notions of consistency and predicate locks in a database system · Communications of the ACM · 1976 · 1,291 citations
- Monitors · Communications of the ACM · 1974 · 1,150 citations
- Eventually consistent · Communications of the ACM · 2008 · 560 citations
- A first order approximation to the optimum checkpoint interval · Communications of the ACM · 1974 · 491 citations
- CAP twelve years later: How the "rules" have changed · Computer · 2012 · 436 citations
- On-the-fly garbage collection · Communications of the ACM · 1978 · 419 citations
- The process group approach to reliable distributed computing · Communications of the ACM · 1993 · 389 citations
Most recent work
- Mechanistic Circuit-Breakers Generalize Across Irreversible Agent Actions and Architectures · Zenodo (CERN European Organization for Nuclear Research) · 2026
- Graduated STA Control: Authority Governors and Runtime Deployment Architecture · Zenodo (CERN European Organization for Nuclear Research) · 2026
- SAFE-Matter™ State-Machine Schema: Deterministic Runtime State Governance for Admissibility, Transition Integrity, and Execution Authority in Life-Critical Systems · Zenodo (CERN European Organization for Nuclear Research) · 2026
- High-Threshold, Low-Overhead and Single-Shot Decodable Fault-Tolerant Quantum Memory · PRX Quantum · 2026
- Recurrence Sets for Proving Fair Non-termination under Axiomatic Memory Consistency Models · Proceedings of the ACM on Programming Languages · 2026
- A Verified High-Performance Composable Object Library for Remote Direct Memory Access · Proceedings of the ACM on Programming Languages · 2026
- Verifying Almost-Sure Termination for Randomized Distributed Algorithms · Proceedings of the ACM on Programming Languages · 2026
- Behavioural theory of reflective algorithms II: Reflective parallel algorithms · Science of Computer Programming · 2026
- Improved Bounds for Coin Flipping, Leader Election, and Random Selection · 2026
- Grid Checkpointing in Speculative Simulation · 2026
Find a gap in your own Distributed systems and fault tolerance sub-topic
This page shows what the Distributed systems and fault tolerance literature already flags as unresolved. To narrow it to your specific question, search the Research Gap Finder: the search is free with a free account and lists the papers closest to your topic first. Unlocking that topic (50 credits, charged once) fills the comparison table from our 4.5M-paper local library and writes the gaps from its rows.
Open the Research Gap Finder →