English

Fault-Tolerant Distance Oracles Below the $n \cdot f$ Barrier

Data Structures and Algorithms 2026-03-26 v1

Abstract

Fault-tolerant spanners are fundamental objects that preserve distances in graphs even under edge failures. A long line of work culminating in Bodwin, Dinitz, Robelle (SODA 2022) gives (2k1)(2k-1)-stretch, ff-fault-tolerant spanners with O(k2f1212kn1+1k+kfn)O(k^2 f^{\frac{1}{2}-\frac{1}{2k}} n^{1+\frac{1}{k}} + k f n) edges for any odd kk. For any k=O~(1)k = \tilde{O}(1), this bound is essentially optimal for deterministic spanners in part due to a known folklore lower bound that \emph{any} ff-fault-tolerant spanner requires Ω(nf)\Omega(nf) edges in the worst case. For fnf \geq n, this Ω(nf)\Omega(nf) barrier means that any ff-fault tolerant spanners are trivial in size. Crucially however, this folklore lower bound exploits that the spanner \emph{is itself a subgraph}. It does not rule out distance-reporting data structures that may not be subgraphs. This leads to our central question: can one beat the nfn \cdot f barrier with fault-tolerant distance oracles? We give a strong affirmative answer to this question. As our first contribution, we construct ff-fault-tolerant distance oracles with stretch O(log(n)loglog(n))O(\log(n)\log\log(n)) that require only O~(nf)\widetilde{O}(n\sqrt{f}) bits of space; substantially below the spanner barrier of nfn \cdot f. Beyond this, in the regime nfn3/2n \leq f \leq n^{3/2} we show that by using our new \emph{high-degree, low-diameter} decomposition in combination with tools from sparse recovery, we can even obtain stretch 77 distance oracles in space O~(n3/2f1/3)\widetilde{O}(n^{3/2}f^{1/3}) bits. We also show that our techniques are sufficiently general to yield randomized sketches for fault-tolerant ``oblivious'' spanners and fault-tolerant deterministic distance oracles in bounded-deletion streams, with space below the nfnf barrier in both settings.

Cite

@article{arxiv.2603.24530,
  title  = {Fault-Tolerant Distance Oracles Below the $n \cdot f$ Barrier},
  author = {Sanjeev Khanna and Christian Konrad and Aaron Putterman},
  journal= {arXiv preprint arXiv:2603.24530},
  year   = {2026}
}