The Incidence-Multiplicity Bound for Linear Exact Repair in MDS Array Codes

Abstract: We study linear exact repair for $(n,k,\ell)$ MDS array codes over $\mathbb{F}_q$, with redundancy $r=n-k$, in the regime where $q$, $r$, and $\ell$ are fixed and the code length $n$ varies. A recent projective counting argument gives a general lower bound on repair bandwidth and repair I/O in this setting. While this bound is attained over a broad interval of code lengths in the two-parity case, it is not attained once $r\ge 3$ and $\ell\ge 2$. In this paper, we refine the counting argument behind this bound and establish a sharper lower bound, which we call the incidence-multiplicity bound. We prove that for every $(n,k,\ell)$ MDS array code over $\mathbb{F}_q$ with $r\ge 2$, both the average and worst-case repair bandwidth, as well as the average and worst-case repair I/O, are at least $$\ell(n-1)-(r-1)\frac{q^\ell-1}{q-1}.$$This bound agrees with the earlier projective counting bound when $r=2$, and is strictly stronger for every $r\ge 3$. We also show that the incidence-multiplicity bound is sharp in a broad parameter range. Assume that $\ell\ge 2$, $r\ge 2$, $(r-1)\mid(q-1)$, and $(q-1)/(r-1)\ge 2$. Then for every integer $n$ satisfying $$2(r-1)\frac{q^\ell-1}{q-1}\le n\le q^\ell+1,$$ there exists an $(n,n-r,\ell)$ MDS array code over $\mathbb{F}_q$ that attains the incidence-multiplicity bound simultaneously for both repair bandwidth and repair I/O. These codes arise from field reduction of a normal rational curve. Together, these results reveal incidence multiplicity as the governing geometric principle for linear exact repair in MDS array codes beyond the two-parity case.

• The paper presents a refined incidence-multiplicity bound that improves repair bandwidth and I/O lower bounds compared to previous projective counting methods.
• It recasts linear exact repair in MDS array codes through a geometric lens, leveraging dual projective geometry to characterize subspace incidences.
• Constructive designs using field reduction from normal rational curves demonstrate the bound’s sharpness under specific divisibility conditions.

The Incidence-Multiplicity Bound for Linear Exact Repair in MDS Array Codes

Introduction and Problem Context

The study of exact node repair in MDS array codes over finite fields is foundational for the design and analysis of distributed storage systems. In the practical regime where code length $n$ varies but parameters such as field size $q$, redundancy $r$, and sub-packetization level $\ell$ are fixed, there is a critical interest in determining sharp lower bounds on the repair bandwidth and I/O for linear exact repair. Existing projective counting techniques provide lower bounds that are sharp for $r=2$ (two-parity codes), but fail to provide tight results as $r$ and $\ell$ grow. This paper presents a refined counting argument—termed the incidence-multiplicity bound—which encapsulates the core geometric constraints imposed by the MDS property and yields strictly stronger results for $r \geq 3$.

Mathematical Foundations and Main Results

Intrinsic Formulation

By recasting linear repair in terms of subspaces of the ambient vector space $\mathbb{V} = F_q^{r\ell}$ and their associated parity-check blocks, the author leverages a geometric lens. Each node is associated with a subspace $\mathcal{H}_i \leq \mathbb{V}$ of dimension $q$0, and the code is MDS precisely when any collection of $q$1 such subspaces collectively span $q$2. For a given failed node $q$3, the feasible set of repair subspaces $q$4 consists of codimension-$q$5 subspaces of $q$6 avoiding $q$7, which correspond to kernel spaces of possible repair matrices.

Incidence-Multiplicity Bound

The principal theorem states: For any $q$8 MDS array code over $q$9 with redundancy $r$0, the per-node as well as average and worst-case repair bandwidth and repair I/O satisfy

$r$1

This improves upon the previous bound $r$2 when $r$3, and converges with it for $r$4. The core of the proof exploits dual projective geometry: points in the dual space of the repair subspace can be "hit" no more than $r$5 times among the node subspaces, leading to the stated combinatorial upper bound on incidences. The analysis makes critical use of the intersection pattern between the helper node subspaces and the repair space, focusing on the aggregate dimensions of intersections.

Tightness and Constructions

The bound is shown to be sharp for a broad explicit family when the divisibility constraint $r$6 holds, in the code length range

$r$7

The construction uses field reduction from a normal rational curve in projective space. For each code, node subspaces are defined as the spreads obtained from points on the normal rational curve, and repair spaces are chosen so that each hits a precise subset of helper node subspaces, maximizing the aggregate intersection dimension and realizing the bound for both repair bandwidth and I/O, simultaneously in average and worst-case forms.

Technical Significance

Improvement Over Projective Bound

The incidence-multiplicity bound identifies the correct scale for the "correction term" in the repair lower bound, aligning it with the order of $r$8 rather than $r$9. This matches the scale of the classical MDS length upper bound, which is $\ell$0, and therefore closes a substantial gap in the theory for $\ell$1.

Geometric Principle

The key insight is that optimal repair is governed not just by disjointness in projective spreads but by higher-order incidence multiplicities in dual subspaces. Specifically, the number of projective points that can be covered $\ell$2 times without violating the MDS condition determines the fundamental lower bound for repair resources.

Constructive Attainability

Constructions based on field reduction of normal rational curves provide codes attaining the bound, covering a wide range of practical parameters. The necessary and sufficient condition $\ell$3 naturally arises from the group structure of the kernel of the field norm map, aligning block partitions of the parameter set with incidence constraints.

Implications and Open Problems

Theoretical Implications

The results delineate a refined geometric obstruction to efficient linear exact repair, which acts at the correct combinatorial order for all $\ell$4. For $\ell$5 below the compositional threshold $\ell$6, the bound cannot be sharp, raising the question of what alternative mechanisms may dictate repair complexity in the short-length regime.

Practical Implications

The findings suggest the design of MDS array codes for storage systems should respect the incidence-multiplicity limits when constructing codes with fixed sub-packetization and redundancy, especially in systems where both repair bandwidth and repair I/O are concurrent bottlenecks.

Open Directions

• Extension of Attainable Length: Determining if the attainability of the incidence-multiplicity bound can persist for $\ell$7 below $\ell$8 and characterizing constructions in these regimes.
• Bandwidth/I-O Separation: Investigating if there exist settings where the bound for repair bandwidth is attained but not for repair I/O.
• Relaxation of Divisibility Constraint: Developing constructions attaining the bound when $\ell$9.
• Characterization in Short-Length Regime: Identifying the correct lower bounds and geometric principles for $r=2$0.

Conclusion

This paper establishes the incidence-multiplicity bound as the fundamental limit for linear exact repair bandwidth and I/O in MDS array codes of fixed redundancy and sub-packetization over finite fields. The theoretical advancement lies in the identification of incidence multiplicity as the geometric mechanism superseding projective packing for $r=2$1, substantiated both by tight combinatorial bounds and explicit code constructions. These results refine our understanding of the trade-offs in code parameters and inform practical code design for distributed storage applications, while posing several strong open problems for future inquiry.

---

Reference:

"The Incidence-Multiplicity Bound for Linear Exact Repair in MDS Array Codes" (2604.05692)