Abstract:We consider locally recoverable codes (LRCs) and aim to determine the smallest possible length $n=n_q(k,d,r)$ of a linear $[n,k,d]_q$-code with locality $r$. For $k\le 7$ we exactly determine all values of $n_2(k,d,2)$ and for $k\le 6$ we exactly determine all values of $n_2(k,d,1)$. For the ternary field we also state a few numerical results. As a general result we prove that $n_q(k,d,r)$ equals the Griesmer bound if the minimum Hamming distance $d$ is sufficiently large and all other parameters are fixed.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to determine the minimum possible length \(n = n_q(k, d, r)\) of linear codes with local recoverability (Locally Recoverable, LRC), where \(k\) is the dimension of the codeword, \(d\) is the minimum Hamming distance, and \(r\) is the locality parameter. Specifically, for given parameters \(k\) and \(d\), the author aims to find the shortest length \(n\) of linear codes with a specific locality \(r\).
### Background and Motivation
Local recovery codes (LRCs) are widely used in distributed storage and communication systems. These codes are characterized by the fact that each encoded symbol can be recovered by a small number of other encoded symbols (at most \(r\) symbols). Therefore, LRCs can quickly recover data when some symbols are lost or corrupted while maintaining a high error - correction capability.
### Main Contributions
1. **Exact Results**: For \(k\leq7\), the author precisely determined all values of \(n_2(k, d, 2)\); for \(k\leq6\), precisely determined all values of \(n_2(k, d, 1)\).
2. **General Results**: It is proved that when the minimum Hamming distance \(d\) is sufficiently large and other parameters are fixed, \(n_q(k, d, r)\) is equal to the Griesmer bound.
3. **Numerical Results**: For the ternary field, the author also gave some numerical results.
### Mathematical Background
- **Griesmer Bound**: For a linear code \([n, k, d]_q\), its minimum possible length \(n_q(k, d)\) satisfies:
\[
n_q(k, d)\geq g_q(k, d):=\sum_{i = 0}^{k - 1}\left\lceil\frac{d}{q^i}\right\rceil
\]
- **Locality Definition**: A linear code \(C\subseteq F_q^n\) has locality \(r\), if for each coordinate \(i\in\{1,\ldots,n\}\), there exists a set \(S_i\subseteq\{1,\ldots,n\}\) such that \(i\notin S_i\), and \(\vert S_i\vert\leq r\), and for any two codewords \(c, c'\in C\), if \(c_j = c'_j\) for all \(j\in S_i\), then \(c_i = c'_i\).
### Methods and Techniques
- **Geometric Interpretation**: The author re - interprets locality through geometric methods and uses this to derive bounds on the minimum possible length.
- **Integer Linear Programming (ILP)**: The author uses integer linear programming to determine the minimum possible length \(n_q(k, d, r)\) for specific parameters.
### Conclusions
- The author proves that for given parameters \(k\) and \(q\), when the minimum Hamming distance \(d\) is sufficiently large, the minimum possible length \(n_q(k, d, r)\) of linear codes with locality \(r\) is equal to the Griesmer bound.
- The author also provides specific numerical results, especially for some small parameter sets in the binary field and the ternary field.
Through these results, the author provides an important theoretical basis and practical guidance for understanding and designing linear codes with local recoverability.