Windows Azure Storage (WAS) is a cloud storage system that provides customers the ability to store seemingly limitless amounts of data for any duration of time WAS customers have access to their data from anywhere, at any time, and only pay for what they use and store To provide durability for that data and to keep the cost of storage low, WAS uses erasure coding

In this paper we introduce a new set of codes for erasure coding called Local Reconstruction Codes (LRC) LRC reduces the number of erasure coding fragments that need to be read when reconstructing data fragments that are offline, while still keeping the storage overhead low The important benefits of LRC are that it reduces the bandwidth and I/Os required for repair reads over prior codes, while still allowing a significant reduction in storage overhead We describe how LRC is used in WAS to provide low overhead durable storage with consistently low read latencies

/pdf/erasure-coding-in-windows-azure-storage-3hqt1rghxj.pdf

Erasure coding in windows azure storage

Consider a linear [n,k,d]q code C. We say that the ith coordinate of C has locality r , if the value at this coordinate can be recovered from accessing some other r coordinates of C. Data storage applications require codes with small redundancy, low locality for information coordinates, large distance, and low locality for parity coordinates. In this paper, we carry out an in-depth study of the relations between these parameters. We establish a tight bound for the redundancy n-k in terms of the message length, the distance, and the locality of information coordinates. We refer to codes attaining the bound as optimal. We prove some structure theorems about optimal codes, which are particularly strong for small distances. This gives a fairly complete picture of the tradeoffs between codewords length, worst case distance, and locality of information symbols. We then consider the locality of parity check symbols and erasure correction beyond worst case distance for optimal codes. Using our structure theorem, we obtain a tight bound for the locality of parity symbols possible in such codes for a broad class of parameter settings. We prove that there is a tradeoff between having good locality and the ability to correct erasures beyond the minimum distance.

/pdf/on-the-locality-of-codeword-symbols-3k0ht1w1qd.pdf

On the Locality of Codeword Symbols

Naiad is a distributed system for executing data parallel, cyclic dataflow programs. It offers the high throughput of batch processors, the low latency of stream processors, and the ability to perform iterative and incremental computations. Although existing systems offer some of these features, applications that require all three have relied on multiple platforms, at the expense of efficiency, maintainability, and simplicity. Naiad resolves the complexities of combining these features in one framework.A new computational model, timely dataflow, underlies Naiad and captures opportunities for parallelism across a wide class of algorithms. This model enriches dataflow computation with timestamps that represent logical points in the computation and provide the basis for an efficient, lightweight coordination mechanism.We show that many powerful high-level programming models can be built on Naiad's low-level primitives, enabling such diverse tasks as streaming data analysis, iterative machine learning, and interactive graph mining. Naiad outperforms specialized systems in their target application domains, and its unique features enable the development of new high-performance applications.

/pdf/naiad-a-timely-dataflow-system-1z2cah9a5l.pdf

Naiad: a timely dataflow system

Big Data computing and clouds

Distributed storage systems for large clusters typically use replication to provide reliability. Recently, erasure codes have been used to reduce the large storage overhead of three-replicated systems. Reed-Solomon codes are the standard design choice and their high repair cost is often considered an unavoidable price to pay for high storage efficiency and high reliability.This paper shows how to overcome this limitation. We present a novel family of erasure codes that are efficiently repairable and offer higher reliability compared to Reed-Solomon codes. We show analytically that our codes are optimal on a recently identified tradeoff between locality and minimum distance.We implement our new codes in Hadoop HDFS and compare to a currently deployed HDFS module that uses Reed-Solomon codes. Our modified HDFS implementation shows a reduction of approximately 2× on the repair disk I/O and repair network traffic. The disadvantage of the new coding scheme is that it requires 14% more storage compared to Reed-Solomon codes, an overhead shown to be information theoretically optimal to obtain locality. Because the new codes repair failures faster, this provides higher reliability, which is orders of magnitude higher compared to replication.

/pdf/xoring-elephants-novel-erasure-codes-for-big-data-1lxen2ntyj.pdf

XORing elephants: novel erasure codes for big data

Windows Azure Storage (WAS) is a cloud storage system that provides customers the ability to store seemingly limitless amounts of data for any duration of time. WAS customers have access to their data from anywhere at any time and only pay for what they use and store. In WAS, data is stored durably using both local and geographic replication to facilitate disaster recovery. Currently, WAS storage comes in the form of Blobs (files), Tables (structured storage), and Queues (message delivery). In this paper, we describe the WAS architecture, global namespace, and data model, as well as its resource provisioning, load balancing, and replication systems.

/pdf/windows-azure-storage-a-highly-available-cloud-storage-18wio04utc.pdf

Windows Azure Storage: a highly available cloud storage service with strong consistency

Embodiments of the present invention relate to systems, methods and computer storage media for erasure coding data in a distributed computing environment. A sealed extent is identified that is comprised of two or more data blocks and two or more index blocks. The sealed extent is optimized for erasure coding by grouping the two or more data blocks within the optimized sealed extent together and grouping the two or more index blocks within the optimized sealed extent together. The optimized extent may also be erasure coded, which includes creating data fragments and coding fragments. The data fragments and the coding fragments may also be stored in the distributed computing environment. Additional embodiments include monitoring statistical information to determine if replication, erasure coding or a hybrid storage plan should be utilized.

Erasure coding immutable data

In various embodiments, methods and systems for implementing multiple transaction logs in a distributed storage system are provided. A log stream component detects performance metrics of a plurality of log streams. The performance metrics are associated with requests from partitions in the distributed storage system. A transaction component receives a request to execute a transaction using a log stream. The request is received from a partition of the distributed storage system. The performance metrics of the plurality of log streams can be referenced, where the performance metrics indicate a performance capacity of a selected log stream to process the request. A log stream for executing the transaction is determined based on the performance capacity. The log stream selected can also factor request attributes of the request. The transaction component communicates the request to be executed, using the log stream to perform the transaction.

Multiple transaction logs in a distributed storage system

Various embodiments for implementing garbage collection in distributed storage systems are provided. A distributed storage system operates based on independent management of metadata of extent and stream data storage resources. A hybrid garbage collection system based on reference counting garbage collection operations and mark-and-sweep garbage collection operations is implemented. An extent lifetime table that tracks reference weights and mark sequences for extents is initialized and updated based on indications from extent managers and stream managers, respectively. Upon determining that an extent is to be handed-off from weighted reference counting garbage collection operations to mark-and-sweep garbage collection operations, a reference weight field for the extent is voided and a mark sequence field of the extent is updated. The mark sequence field is updated with a latest global sequence number. Mark-and-sweep garbage collection operations are utilized to reclaim the extent when it is no longer referenced.

Yikang Xu

Papers

Erasure coding in windows azure storage

Windows Azure Storage: a highly available cloud storage service with strong consistency

Erasure coding immutable data

Multiple transaction logs in a distributed storage system

Hybrid garbage collection in a distrubuted storage system