Searcharxiv⌕ Search

arXiv subjects

Danilo Gligoroski

Publications and source records attributed to Danilo Gligoroski.

33 records · Page 2Linked to original sources
arXiv2017

Repair Duality with Locally Repairable and Locally Regenerating Codes

We construct an explicit family of locally repairable and locally regenerating codes whose existence was proven in a recent work by Kamath et al. about codes with local regeneration but no explicit construction was given. This explicit family of codes is based on HashTag codes. HashTag codes are recently defined vector codes with different vector length $α$ (also called a sub-packetization level) that achieve the optimal repair bandwidth of MSR codes or near-optimal repair bandwidth depending on the sub-packetization level. We applied the technique of parity-splitting code construction. We show that the lower bound on the size of the finite field for the presented explicit code constructions can be lower than the one given in the work of Kamath et al. Finally, we discuss the importance of having two ways for node repair with locally regenerating HashTag codes: repair only with local parity nodes or repair with both local and global parity nodes. To the best of the authors' knowledge, this is the first work where this duality in repair process is discussed. We give a practical example and experimental results in Hadoop where we show the benefits of having this repair duality.

cs.IT↗
arXiv2017

Balanced Locally Repairable Codes

We introduce a family of balanced locally repairable codes (BLRCs) $[n, k, d]$ for arbitrary values of $n$, $k$ and $d$. Similar to other locally repairable codes (LRCs), the presented codes are suitable for applications that require a low repair locality. The novelty that we introduce in our construction is that we relax the strict requirement the repair locality to be a fixed small number $l$, and we allow the repair locality to be either $l$ or $l+1$. This gives us the flexibility to construct BLRCs for arbitrary values of $n$ and $k$ which partially solves the open problem of finding a general construction of LRCs. Additionally, the relaxed locality criteria gives us an opportunity to search for BLRCs that have a low repair locality even when double failures occur. We use metrics such as a storage overhead, an average repair bandwidth, a Mean Time To Data Loss (MTTDL) and an update complexity to compare the performance of BLRCs with existing LRCs.

cs.IT↗
arXiv2017

HashTag Erasure Codes: From Theory to Practice

Minimum-Storage Regenerating (MSR) codes have emerged as a viable alternative to Reed-Solomon (RS) codes as they minimize the repair bandwidth while they are still optimal in terms of reliability and storage overhead. Although several MSR constructions exist, so far they have not been practically implemented mainly due to the big number of I/O operations. In this paper, we analyze high-rate MDS codes that are simultaneously optimized in terms of storage, reliability, I/O operations, and repair-bandwidth for single and multiple failures of the systematic nodes. The codes were recently introduced in \cite{7463553} without any specific name. Due to the resemblance between the hashtag sign \# and the procedure of the code construction, we call them in this paper \emph{HashTag Erasure Codes (HTECs)}. HTECs provide the lowest data-read and data-transfer, and thus the lowest repair time for an arbitrary sub-packetization level $α$, where $α\leq r^{\lceil \sfrac{k}{r} \rceil}$, among all existing MDS codes for distributed storage including MSR codes. The repair process is linear and highly parallel. Additionally, we show that HTECs are the first high-rate MDS codes that reduce the repair bandwidth for more than one failure. Practical implementations of HTECs in Hadoop release 3.0.0-alpha2 demonstrate their great potentials.

cs.IT↗
arXiv2016

Families of Optimal Binary Non-MDS Erasure Codes

We introduce a definition for \emph{Families of Optimal Binary Non-MDS Erasure Codes} for $[n, k]$ codes over $GF(2)$, and propose an algorithm for finding those families by using hill climbing techniques over Balanced XOR codes. Due to the hill climbing search, those families of codes have always better decoding probability than the codes generated in a typical Random Linear Network Coding scenario, i.e., random linear codes. We also show a surprising result that for small values of $k$, the decoding probability of our codes in $GF(2)$ is very close to the decoding probability of the codes obtained by Random Linear Network Coding but in the higher finite field $GF(4)$.

cs.IT↗
arXiv2016

General Sub-packetized Access-Optimal Regenerating Codes

This paper presents a novel construction of $(n,k,d=n-1)$ access-optimal regenerating codes for an arbitrary sub-packetization level $α$ for exact repair of any systematic node. We refer to these codes as general sub-packetized because we provide an algorithm for constructing codes for any $α$ less than or equal to $r^{\lceil \frac{k}{r} \rceil}$ where $\frac{k}{r}$ is not necessarily an integer. This leads to a flexible construction of codes for different code rates compared to existing approaches. We derive the lower and the upper bound of the repair bandwidth. The repair bandwidth depends on the code parameters and $α$. The repair process of a failed systematic node is linear and highly parallelized, which means that a set of $\lceil \fracα{r} \rceil$ symbols is independently repaired first and used along with the accessed data from other nodes to recover the remaining symbols.

cs.IT↗
arXiv2016

Coded Packet Transport for Optical Packet/Burst Switched Networks

This paper presents the Coded Packet Transport (CPT) scheme, a novel transport mechanism for Optical Packet/Burst Switched (OPS/OBS) networks. The CPT scheme exploits the combined benefits of source coding by erasure codes and path diversity to provide efficient means for recovering from packet loss due to contentions and path failures, and to provide non-cryptographic secrecy. In the CPT scheme, erasure coding is employed at the OPS/OBS ingress node to form coded packets, which are transmitted on disjoint paths from the ingress node to an egress node in the network. The CPT scheme allows for a unified view of Quality of Service (QoS) in OPS/OBS networks by linking the interactions between survivability, performance and secrecy. We provide analytical models that illustrate how QoS aspects of CPT are affected by the number of disjoint paths, packet overhead and processing delay.

cs.IT↗
arXiv2016

Minimal Header Overhead for Random Linear Network Coding

The energy used to transmit a single bit of data between the devices in wireless networks is equal to the energy for performing hundreds of instructions in those devices. Thus the reduction of the data necessary to transmit, while keeping the same functionality of the employed algorithms is a formidable and challenging scientific task. We describe an algorithm called Small Set of Allowed Coefficients (SSAC) that produces the shortest header overhead in random linear network coding schemes compared with all other approaches reported in the literature. The header overhead length is 2 to 7 times shorter than the length achieved by related compression techniques. For example, SSAC algorithm compresses the length of the header overhead in a generation of 128 packets to 24 bits, while the closest best result achieved by an algorithm based on error correcting codes has a header overhead length of 84 bits in $GF(16)$ and 224 bits in $GF(256)$. We show that the header length in SSAC does not depend on the size of the finite field where the operations are performed, i.e., it just depends on the number of combined packets $m$.

cs.IT↗
arXiv2016

Balanced XOR-ed Coding

This paper concerns with the construction of codes over $GF(2)$ which reach the max-flow for single source multicast acyclic networks with delay. The coding is always a bitwise XOR of packets with equal lengths, and is based on highly symmetrical and balanced designs. For certain setups and parameters, our approach offers additional plausible security properties: an adversary needs to eavesdrop at least max-flow links in order to decode at least one original packet.

cs.IT↗
arXiv2015

Approaching Maximum Embedding Efficiency on Small Covers Using Staircase-Generator Codes

We introduce a new family of binary linear codes suitable for steganographic matrix embedding. The main characteristic of the codes is the staircase random block structure of the generator matrix. We propose an efficient list decoding algorithm for the codes that finds a close codeword to a given random word. We provide both theoretical analysis of the performance and stability of the decoding algorithm, as well as practical results. Used for matrix embedding, these codes achieve almost the upper theoretical bound of the embedding efficiency for covers in the range of 1000 - 1500 bits, which is at least an order of magnitude smaller than the values reported in related works.

cs.MM↗
arXiv2010

The Digital Signature Scheme MQQ-SIG

This document contains the Intellectual Property Statement and the technical description of the MQQ-SIG - a new public key digital signature scheme. The complete scientific publication covering the design rationale and the security analysis will be given in a separate publication. MQQ-SIG consists of $n - \frac{n}{4}$ quadratic polynomials with $n$ Boolean variables where n=160, 196, 224 or 256.

cs.CR↗
arXiv2010

Polynomial functions on the units of Z_{2^n}

Polynomial functions on the group of units Q_n of the ring Z_{2^n} are considered. A finite set of reduced polynomials RP_n in Z[x] that induces the polynomial functions on Q_n is determined. Each polynomial function on Q_n is induced by a unique reduced polynomial - the reduction being made using a suitable ideal in Z[x]. The set of reduced polynomials forms a multiplicative 2-group. The obtained results are used to efficiently construct families of exponential cardinality of, so called, huge k-ary quasigroups, which are useful in the design of various types of cryptographic primitives. Along the way we provide a new (and simpler) proof of a result of Rivest characterizing the permutational polynomials on Z_{2^n}.

math.AC↗
arXiv2008

A Public Key Block Cipher Based on Multivariate Quadratic Quasigroups

We have designed a new class of public key algorithms based on quasigroup string transformations using a specific class of quasigroups called multivariate quadratic quasigroups (MQQ). Our public key algorithm is a bijective mapping, it does not perform message expansions and can be used both for encryption and signatures. The public key consist of n quadratic polynomials with n variables where n=140, 160, ... . A particular characteristic of our public key algorithm is that it is very fast and highly parallelizable. More concretely, it has the speed of a typical modern symmetric block cipher - the reason for the phrase "A Public Key Block Cipher" in the title of this paper. Namely the reference C code for the 160-bit variant of the algorithm performs decryption in less than 11,000 cycles (on Intel Core 2 Duo -- using only one processor core), and around 6,000 cycles using two CPU cores and OpenMP 2.0 library. However, implemented in Xilinx Virtex-5 FPGA that is running on 249.4 MHz it achieves decryption throughput of 399 Mbps, and implemented on four Xilinx Virtex-5 chips that are running on 276.7 MHz it achieves encryption throughput of 44.27 Gbps. Compared to fastest RSA implementations on similar FPGA platforms, MQQ algorithm is more than 10,000 times faster.

cs.CR↗
arXiv2005

Candidate One-Way Functions and One-Way Permutations Based on Quasigroup String Transformations

In this paper we propose a definition and construction of a new family of one-way candidate functions ${\cal R}_N:Q^N \to Q^N$, where $Q=\{0,1,...,s-1\}$ is an alphabet with $s$ elements. Special instances of these functions can have the additional property to be permutations (i.e. one-way permutations). These one-way functions have the property that for achieving the security level of $2^n$ computations in order to invert them, only $n$ bits of input are needed. The construction is based on quasigroup string transformations. Since quasigroups in general do not have algebraic properties such as associativity, commutativity, neutral elements, inverting these functions seems to require exponentially many readings from the lookup table that defines them (a Latin Square) in order to check the satisfiability for the initial conditions, thus making them natural candidates for one-way functions.

cs.CR↗
arXiv2004

Stream cipher based on quasigroup string transformations in $Z_p^*$

In this paper we design a stream cipher that uses the algebraic structure of the multiplicative group $\bbbz_p^*$ (where p is a big prime number used in ElGamal algorithm), by defining a quasigroup of order $p-1$ and by doing quasigroup string transformations. The cryptographical strength of the proposed stream cipher is based on the fact that breaking it would be at least as hard as solving systems of multivariate polynomial equations modulo big prime number $p$ which is NP-hard problem and there are no known fast randomized or deterministic algorithms for solving it. Unlikely the speed of known ciphers that work in $\bbbz_p^*$ for big prime numbers $p$, the speed of this stream cipher both in encryption and decryption phase is comparable with the fastest symmetric-key stream ciphers.

cs.CR↗