Abstract:Erasure codes have become an integral part of distributed storage systems as a tool for providing data reliability and durability under the constant threat of device failures. In such systems, an $[n, k]$ code over a finite field $\mathbb{F}_q$ encodes $k$ message symbols into $n$ codeword symbols from $\mathbb{F}_q$ which are then stored on $n$ different nodes in the system. Recent work has shown that significant savings in storage space can be obtained by tuning $n$ and $k$ to variations in device failure rates. Such a tuning necessitates code conversion: the process of converting already encoded data under an initial $[n^I, k^I]$ code to its equivalent under a final $[n^F, k^F]$ code. The default approach to conversion is to reencode data, which places significant burden on system resources. Convertible codes are a recently proposed class of codes for enabling resource-efficient conversions. Existing work on convertible codes has focused on minimizing access cost, i.e., the number of code symbols accessed during conversion. Bandwidth, which corresponds to the amount of data read and transferred, is another important resource to optimize. In this paper, we initiate the study on the fundamental limits on bandwidth used during code conversion and present constructions for bandwidth-optimal convertible codes. First, we model the code conversion problem using network information flow graphs with variable capacity edges. Second, focusing on MDS codes and an important parameter regime called the merge regime, we derive tight lower bounds on the bandwidth cost of conversion. The derived bounds show that bandwidth cost can be significantly reduced even in regimes where access cost cannot be reduced as compared to the default approach. Third, we present a new construction for MDS convertible codes which matches the proposed lower bound and is thus bandwidth-optimal during conversion.

Reduction of Sufficient Number of Code Tables of $k$-Bit Delay Decodable Codes

Properties of k-bit Delay Decodable Codes

The Optimality of AIFV Codes in the Class of $2$-bit Delay Decodable Codes

Optimization of the H.264 Coefficient Tables Based on Variable Step-Length Group

Optimal Construction of N-bit-delay Almost Instantaneous Fixed-to-Variable-Length Codes

Research on Low Delay Low Bit Rate Speech Coding Algorithm

A Family of Low-Complexity Binary Codes with Constant Hamming Weights

Shortened LDPC codes accelerate OSD decoding performance

On the Minimum Decoding Delay of Balanced Complex Orthogonal Designs

Optimized Degree Distribution of Rateless Codes With Lower Complexity

An Approximate Coding-Rate Versus Minimum Distance Formula for Binary Codes

Divisible minimal codes

A $k^{\frac{q}{q-2}}$ Lower Bound for Odd Query Locally Decodable Codes from Bipartite Kikuchi Graphs

Design and Performance of Rate-compatible Non-Binary LDPC Convolutional Codes

Improved Explicit Near-Optimal Codes in the High-Noise Regimes

A New Class of Near-Optimal Partial Fourier Codebooks from an Almost Difference Set

P-Adic LDPC Codes at Precision k

Rate-adaptive codes for distributed source coding

Bandwidth Cost of Code Conversions in Distributed Storage: Fundamental Limits and Optimal Constructions

Strategies for Reducing Decoding Cycles in Stochastic LDPC Decoders.

An Efficient Decoding Algorithm for Low-Density Parity-Check Codes.