Emulating a shared atomic , read/write storage system is a fundamental problem in distributed computing. Replicating atomic objects among a set of data hosts was the norm for traditional implementations (e.g., [ 11 ]) in order to guarantee the availability and accessibility of the data despite host failures. As replication is highly storage demanding, recent approaches suggested the use of erasure-codes to offer the same fault-tolerance while optimizing storage usage at the hosts. Initial works focused on a fixed set of data hosts. To guarantee longevity and scalability, a storage service should be able to dynamically mask hosts failures by allowing new hosts to join, and failed host to be removed without service interruptions. This work presents the first erasure-code -based atomic algorithm, called Ares , which allows the set of hosts to be modified in the course of an execution. Ares is composed of three main components: (i) a reconfiguration protocol , (ii) a read/write protocol , and (iii) a set of data access primitives (DAPs) . The design of Ares is modular and is such to accommodate the usage of various erasure-code parameters on a per-configuration basis. We provide bounds on the latency of read/write operations and analyze the storage and communication costs of the Ares algorithm.

Emulating a shared atomic, read/write storage system is a fundamental problem in distributed computing. Replicating atomic objects among a set of data hosts was the norm for traditional implementations (e.g., [6]) in order to guarantee the availability and accessibility of the data despite host failures. As replication is highly storage demanding, recent approaches suggested the use of erasure-codes to offer the same fault-tolerance while optimizing storage usage at the hosts. Initial works focused on a fix set of data hosts. To guarantee longevity and scalability, a storage service should be able to dynamically mask hosts failures by allowing new hosts to join, and failed host to be removed without service interruptions. This work presents the first erasure-code based atomic algorithm, called ARES, which allows the set of hosts to be modified in the course of an execution. ARES is composed of three main components: (i) a reconfiguration protocol, (ii) a read/write protocol, and (iii) a set of data access primitives. The design of ARES is modular and is such to accommodate the usage of various erasure-code parameters on a per-configuration basis. We provide bounds on the latency of read/write operations and analyze the storage and communication costs of the ARES algorithm.

In this work, we propose FLECKS, an algorithm which implements atomic memory objects in a multi-writer multi-reader (MWMR) setting in asynchronous networks and server failures. FLECKS substantially reduces storage and communication costs over its replication-based counterparts by employing erasure-codes. FLECKS outperforms the previously proposed algorithms in terms of the metrics that to deliver good performance such as storage cost per object, communication cost a high fault-tolerance of clients and servers, guaranteed liveness of operation, and a given number of communication rounds per operation, etc. We provide proofs for liveness and atomicity properties of FLECKS and derive worst-case latency bounds for the operations. We implemented and deployed FLECKS in cloud-based clusters and demonstrate that FLECKS has substantially lower storage and bandwidth costs, and significantly lower latency of operations than the replication-based mechanisms.

Codes with locality are a class of codes introduced by Gopalanet al.to efficiently repair a failed node, by minimizing the number of nodes contacted during repair. An$[n,k]$systematic code is said to have information locality$r$, if each message symbol can be recovered by accessing$\leq r$other symbols. An$[n,k]$code is said to have all-symbol locality$r$, if each code symbol can be recovered by accessing$\leq r$other symbols. In this paper, we consider a generalization of codes with all-symbol locality to the case of handling two erasures. We study codes with locality that can recover from two erasures via a sequence of two local, parity-check computations. We refer to these codes as sequential-recovery locally repairable codes (denoted by 2-seq LR codes). Earlier approaches to handling multiple erasures considered recovery in parallel; the sequential approach allows us to potentially construct codes with improved minimum distance. We derive an upper bound on the rate of 2-seq LR codes. We provide constructions based on regular graphs which are rate-optimal with respect to the derived bound. We also characterize the structure of any rate-optimal code. By studying the Generalized Hamming Weights of the dual code, we derive a recursive upper bound on the minimum distance of 2-seq LR codes. We also provide constructions of a family of codes based on Turán graphs, that are optimal with respect to this bound. We also present explicit distance-optimal Turán graph based constructions of 2-seq LR codes for certain parameters. Our approach also leads to a new bound on the minimum distance of codes with all-symbol locality for the single-erasure case.

Motivated by applications to delivery of dynamically updated, but correlated data in settings such as content distribution networks, and distributed file sharing systems, we study a single source multiple destination network coded multicast problem in a cache-aided network. We focus on models where the caches are primarily located near the destinations and the source has no cache. The source observes a sequence of correlated frames and is expected to do frame-by-frame encoding with no access to prior frames. We present a novel scheme that shows how the caches can be advantageously used to decrease the overall cost of multicast, even though the source encodes without access to past data. Our cache design and update scheme works with any choice of network code designed for a corresponding cache-less network, is largely decentralized, and works for an arbitrary network. We study a convex relation of the optimization problem that results from the overall cost function. The results of the optimization problem determine the rate allocation and caching strategies. Numerous simulation results are presented to substantiate the theory developed.

We study a generalization of the setting of regenerating codes, motivated by applications to storage systems consisting of clusters of storage nodes. There are n clusters in total, with m nodes per cluster. A data file is coded and stored across the mn nodes, with each node storing α symbols. For availability of data, we require that the file be retrievable by downloading the entire content from any subset of k clusters. Nodes represent entities that can fail. We distinguish between intra-cluster and inter-cluster bandwidth (BW) costs during node repair. Node-repair in a cluster is accomplished by downloading β symbols each from any set of d other clusters, dubbed remote helper clusters, and also up to α symbols each from any set of I surviving nodes, dubbed local helper nodes, in the host cluster. We first identify the optimal trade-off between storage-overhead and inter-cluster repair-bandwidth under functional repair, and also present optimal exact-repair code constructions for a class of parameters. The new trade-off is strictly better than what is achievable via space-sharing existing coding solutions, whenever ℓ > 0. We then obtain sharp lower bounds on the necessary intracluster repair BW to achieve optimal trade-off. Under functional repair, random linear network codes (RLNCs) simultaneously optimize usage of both inter- and intra-cluster repair BW; simulation results based on RLNCs suggest optimality of the bounds on intra-cluster repair-bandwidth. Our bounds reveal the interesting fact that, while it is beneficial to increase the number of local helper nodes I in order to improve the storage-vs-inter-cluster-repair-BW trade-off, increasing I not only increases intracluster BW in the host-cluster, but also increases the intra-cluster BW in the remote helper clusters. We also analyze resilience of the clustered storage system against passive eavesdropping by providing file-size bounds and optimal code constructions.

We consider a communication problem in which an update of the source message needs to be conveyed to one or more distant receivers that are interested in maintaining specific linear functions of the source message. The setting is one in which the updates are sparse in nature, and where neither the source nor the receiver(s) is aware of the exact difference vector, but only know the amount of sparsity that is present in the difference vector. Under this setting, we are interested in devising linear encoding and decoding schemes that minimize the communication cost involved. We show that the optimal solution to this problem is closely related to the notion of maximally recoverable codes (MRCs), which were originally introduced in the context of coding for storage systems. In the context of storage, MRCs guarantee optimal erasure protection when the system is partially constrained to have local parity relations among the storage nodes. In our problem, we show that optimal solutions exist if and only if MRCs of certain kind (identified by the desired linear functions) exist. We consider point-to-point and broadcast versions of the problem and identify connections to MRCs under both these settings. For the point-to-point setting, we show that our linear-encoder-based achievable scheme is optimal even when non-linear encoding is permitted. The theory is illustrated in the context of updating erasure coded storage nodes. We present examples based on modern storage codes, such as the minimum bandwidth regenerating codes.

We study the trade-off between storage overhead and inter-cluster repair bandwidth in clustered storage systems, while recovering from multiple node failures within a cluster. A cluster is a collection of m nodes, and there are n clusters. For data collection, we download the entire content from any k clusters. For repair of t ≥ 2 nodes within a cluster, we take help from ℓ local nodes, as well as d helper clusters. We characterize the optimal trade-off under functional repair, and also under exact repair for the minimum storage and minimum inter-cluster bandwidth (MBR) operating points. Our bounds show the following interesting facts: 1) When t|(m - ℓ) the tradeoff is the same as that under t = 1, and thus there is no advantage in jointly repairing multiple nodes, 2) When tł(m - ℓ), the optimal file-size at the MBR point under exact repair can be strictly less than that under functional repair. 3) Unlike the case of t = 1, increasing the number of local helper nodes does not necessarily increase the system capacity under functional repair.

Motivated by emerging applications to the edge computing paradigm, we introduce a two-layer erasure-coded fault-tolerant distributed storage system offering atomic access for read and write operations. In edge computing, clients interact with an edge-layer of servers that is geographically near; the edge-layer in turn interacts with a back-end layer of servers. The edge-layer provides low latency access and temporary storage for client operations, and uses the back-end layer for persistent storage. Our algorithm, termed Layered Data Storage (LDS) algorithm, offers several features suitable for edge-computing systems, works under asynchronous message-passing environments, supports multiple readers and writers, and can tolerate f1 < n1/2 and f2 < n2/3 crash failures in the two layers having n1 and n2 servers, respectively. We use a class of erasure codes known as regenerating codes for storage of data in the back-end layer. The choice of regenerating codes, instead of popular choices like Reed-Solomon codes, not only optimizes the cost of back-end storage, but also helps in optimizing communication cost of read operations, when the value needs to be recreated all the way from the back-end. The two-layer architecture permits a modular implementation of atomicity and erasure-code protocols; the implementation of erasure-codes is mostly limited to interaction between the two layers. We prove liveness and atomicity of LDS, and also compute performance costs associated with read and write operations. In a system with n1 = Θ(n2), f1 = Θ(n1), f2 = Θ(n2), the write and read costs are respectively given by Θ(n1) and Θ(1) + n1 I(δ > 0). Here δ is a parameter closely related to the number of write operations that are concurrent with the read operation, and I(δ > 0) is 1 if δ > 0, and 0 if δ = 0. The cost of persistent storage in the back-end layer is Θ(1). The impact of temporary storage is minimally felt in a multi-object system running N independent instances of LDS, where only a small fraction of the objects undergo concurrent accesses at any point during the execution. For the multi-object system, we identify a condition on the rate of concurrent writes in the system such that the overall storage cost is dominated by that of persistent storage in the back-end layer, and is given by Θ(N).

Erasure codes are increasingly being studied in the context of implementing atomic memory objects in large scale asynchronous distributed storage systems. When compared with the traditional replication based schemes, erasure codes have the potential of significantly lowering storage and communication costs while simultaneously guaranteeing the desired resiliency levels. In this work, we propose the Storage-Optimized Data-Atomic (SODA) algorithm for implementing atomic memory objects in the multi-writer multi-reader setting. SODA uses Maximum Distance Separable (MDS) codes, and is specifically designed to optimize the total storage cost for a given fault-tolerance requirement. For tolerating f server crashes in an n-server system, SODA uses an [n, k] MDS code with k = n - f, and incurs a total storage cost of n/n-f. SODA is designed under the assumption of reliable point-to-point communication channels. The communication cost of a write and a read operation are respectively given by O(f2) and n/n-f(δw+1), where δwdenotes the number of writes that are concurrent with the particular read. In comparison with the recent CASGC algorithm [1], which also uses MDS codes, SODA offers lower storage cost while pays more on the communication cost. We also present a modification of SODA, called SODAerr, to handle the case where some of the servers can return erroneous coded elements during a read operation. Specifically, in order to tolerate f server failures and e error-prone coded elements, the SODAerr algorithm uses an [n, k] MDS code such that k = n - 2e - f. SODAerr also guarantees liveness and atomicity, while maintaining an optimized total storage cost of n/n-f-2e.

Erasure codes offer an efficient way to decrease storage and communication costs while implementing atomic memory service in asynchronous distributed storage systems. In this paper, we provide erasure-code-based algorithms having the additional ability to perform background repair of crashed nodes. A repair operation of a node in the crashed state is triggered externally, and is carried out by the concerned node via message exchanges with other active nodes in the system. Upon completion of repair, the node re-enters active state, and resumes participation in ongoing and future read, write, and repair operations. To guarantee liveness and atomicity simultaneously, existing works assume either the presence of nodes with stable storage, or presence of nodes that never crash during the execution. We demand neither of these; instead we consider a natural, yet practical network stability condition N1 that only restricts the number of nodes in the crashed/repair state during broadcast of any message. We present an erasure-code based algorithm RADON_{C} that is always live, and guarantees atomicity as long as condition N1 holds. In situations when the number of concurrent writes is limited, RADON_{C} has significantly improved storage and communication cost over a replication-based algorithm RADON_{R}, which also works under N1. We further show how a slightly stronger network stability condition N2 can be used to construct algorithms that never violate atomicity. The guarantee of atomicity comes at the expense of having an additional phase during the read and write operations.

In this paper, we consider the setting of exact repair linear regenerating codes. Under this setting, we derive a new outer bound on the storage-repair-bandwidth trade-off for the case when d = k = n - 1, where (n; k; d) are parameters of the regenerating code, with their usual meaning. Taken together with the achievability result of Tian et. al. [1], we show that the new outer bound derived here completely characterizes the tradeoff for the case of exact repair linear regenerating codes, when d = k = n-1. The new outer bound is derived by analyzing the dual code of the linear regenerating code.

In this paper, we study codes with locality that can recover from two erasures via a sequence of two local, parity-check computations. By a local parity-check computation, we mean recovery via a single parity-check equation associated with small Hamming weight. Earlier approaches considered recovery in parallel; the sequential approach allows us to potentially construct codes with improved minimum distance. These codes, which we refer to as locally 2-reconstructible codes, are a natural generalization along one direction, of codes with all-symbol locality introduced by Gopalan et al, in which recovery from a single erasure is considered. By studying the generalized Hamming weights of the dual code, we derive upper bounds on the minimum distance of locally 2-reconstructible codes and provide constructions for a family of codes based on Turán graphs, that are optimal with respect to this bound. The minimum distance bound derived here is universal in the sense that no code which permits all-symbol local recovery from 2 erasures can have larger minimum distance regardless of approach adopted. Our approach also leads to a new bound on the minimum distance of codes with all-symbol locality for the single-erasure case.

Regenerating codes and codes with locality are two coding schemes that have recently been proposed, which in addition to ensuring data collection and reliability, also enable efficient node repair. In a situation where one is attempting to repair a failed node, regenerating codes seek to minimize the amount of data downloaded for node repair, while codes with locality attempt to minimize the number of helper nodes accessed. This paper presents results in two directions. In one, this paper extends the notion of codes with locality so as to permit local recovery of an erased code symbol even in the presence of multiple erasures, by employing local codes having minimum distance >2. An upper bound on the minimum distance of such codes is presented and codes that are optimal with respect to this bound are constructed. The second direction seeks to build codes that combine the advantages of both codes with locality as well as regenerating codes. These codes, termed here as codes with local regeneration, are codes with locality over a vector alphabet, in which the local codes themselves are regenerating codes. We derive an upper bound on the minimum distance of vector-alphabet codes with locality for the case when their constituent local codes have a certain uniform rank accumulation property. This property is possessed by both minimum storage regeneration (MSR) and minimum bandwidth regeneration (MBR) codes. We provide several constructions of codes with local regeneration which achieve this bound, where the local codes are either MSR or MBR codes. Also included in this paper, is an upper bound on the minimum distance of a general vector code with locality as well as the performance comparison of various code constructions of fixed block length and minimum distance.

Regenerating codes and codes with locality are two schemes that have recently been proposed to ensure data collection and reliability in a distributed storage network. In a situation where one is attempting to repair a failed node, regenerating codes seek to minimize the amount of data downloaded for node repair, while codes with locality attempt to minimize the number of helper nodes accessed. In this paper, we provide several constructions for a class of vector codes with locality in which the local codes are regenerating codes, that enjoy both advantages. We derive an upper bound on the minimum distance of this class of codes and show that the proposed constructions achieve this bound. The constructions include both the cases where the local regenerating codes correspond to the MSR as well as the MBR point on the storage-repair-bandwidth tradeoff curve of regenerating codes.

Node failures are inevitable in distributed storage systems (DSS). To enable efficient repair when faced with such failures, two main techniques are known: Regenerating codes, i.e., codes that minimize the total repair bandwidth; and codes with locality, which minimize the number of nodes participating in the repair process. This paper focuses on regenerating codes with locality, using pre-coding based on Gabidulin codes, and presents constructions that utilize minimum bandwidth regenerating (MBR) local codes. The constructions achieve maximum resilience (i.e., optimal minimum distance) and have maximum capacity (i.e., maximum rate). Finally, the same pre-coding mechanism can be combined with a subclass of fractional-repetition codes to enable maximum resilience and repair-by-transfer simultaneously.

Let X1, ..., Xmbe a set of m statistically dependent sources over the common alphabet Fq, that are linearly independent when considered as functions over the sample space. We consider a distributed function computation setting in which the receiver is interested in the lossless computation of the elements of an s-dimensional subspace W spanned by the elements of the row vector [X1, ..., Xm]Γ in which the (m × s) matrix Γ has rank s. A sequence of three increasingly refined approaches is presented, all based on linear encoders. The first approach uses a common matrix to encode all the sources and a Korner-Marton like receiver to directly compute W. The second improves upon the first by showing that it is often more efficient to compute a carefully chosen superspace U of W. The superspace is identified by showing that the joint distribution of the {Xi} induces a unique decomposition of the set of all linear combinations of the {Xi}, into a chain of subspaces identified by a normalized measure of entropy. This subspace chain also suggests a third approach, one that employs nested codes. For any joint distribution of the {Xi} and any W, the sum-rate of the nested code approach is no larger than that under the Slepian-Wolf (SW) approach. Under the SW approach, W is computed by first recovering each of the {Xi}. For a large class of joint distributions and subspaces W, the nested code approach is shown to improve upon SW. Additionally, a class of source distributions and subspaces are identified, for which the nested-code approach is sum-rate optimal.

Motivated by applications to distributed storage, Gopalan et al recently introduced the interesting notion of information-symbol locality in a linear code. By this it is meant that each message symbol appears in a parity-check equation associated with small Hamming weight, thereby enabling recovery of the message symbol by examining a small number of other code symbols. This notion is expanded to the case when all code symbols, not just the message symbols, are covered by such “local” parity. In this paper, we extend the results of Gopalan et. al. so as to permit recovery of an erased code symbol even in the presence of errors in local parity symbols. We present tight bounds on the minimum distance of such codes and exhibit codes that are optimal with respect to the local error-correction property. As a corollary, we obtain an upper bound on the minimum distance of a concatenated code.