Linear secret-sharing scheme size problem

For an nn-party monotone access structure Γ2[n]\Gamma\subseteq 2^{[n]}, let Γ\ell_{\Gamma} be the minimum, over all linear secret-sharing schemes for one-bit secrets realizing Γ\Gamma, of the maximum number of bits in any party's share. Define L(n)=maxΓ2[n]ΓL(n)=\max_{\Gamma\subseteq 2^{[n]}}\ell_{\Gamma}. Determine the exact or asymptotically optimal growth of L(n)L(n), including the optimal exponential share-size exponent for realizing arbitrary access structures by linear secret-sharing schemes.

Progress summary

Partially solved

A 2026 e-print reports a better construction, but no matching lower bound or full solution is established.

The problem asks how small shares can be made in linear secret-sharing schemes, with the central goal of improving the best asymptotic construction. The remaining gap is between known upper and lower bounds.

Known results

  • Ball, Çakan, and Malkin (2021) prove a total-share lower bound of 2n12n-1 for 1<t<n1<t<n.
  • Over a field F\mathbb{F}, they prove a lower bound of nlogchar(F)nn\lceil\log_{\operatorname{char}(\mathbb{F})} n\rceil.
  • In characteristic 22, this matches the known O(nlogn)O(n\log n) upper bound attributed to Karchmer and Wigderson, while the general gap remains open.

August 2026 upper-bound update

An IACR e-print reports reducing the asymptotic upper-bound exponent from 0.5850.585 to 0.4960.496, advancing the best known construction. The available evidence does not establish a matching lower bound or a complete resolution.

Current status (as of August 2026): The best reported upper-bound exponent is 0.4960.496, while the general lower-bound gap and the full size problem remain open.

Sources
Sources & referencesView supporting material

Additional references

Solutions 0

No solutions have been posted yet.