Results 1 - 10
of
19
Reliability mechanisms for very large storage systems
- IN PROCEEDINGS OF THE 20TH IEEE / 11TH NASA GODDARD CONFERENCE ON MASS STORAGE SYSTEMS AND TECHNOLOGIES
, 2003
"... Reliability and availability are increasingly important in large-scale storage systems built from thousands of individual storage devices. Large systems must survive the failure of individual components; in systems with thousands of disks, even infrequent failures are likely in some device. We focus ..."
Abstract
-
Cited by 77 (21 self)
- Add to MetaCart
(Show Context)
Reliability and availability are increasingly important in large-scale storage systems built from thousands of individual storage devices. Large systems must survive the failure of individual components; in systems with thousands of disks, even infrequent failures are likely in some device. We focus on two types of errors: nonrecoverable read errors and drive failures. We discuss mechanisms for detecting and recovering from such errors, introducing improved techniques for detecting errors in disk reads and fast recovery from disk failure. We show that simple RAID cannot guarantee sufficient reliability; our analysis examines the tradeoffs among other schemes between system availability and storage efficiency. Based on our data, we believe that two-way mirroring should be sufficient for most large storage systems. For those that need very high reliabilty, we recommend either three-way mirroring or mirroring combined with RAID.
LH*RS -- a high-availability scalable distributed data structure
"... (SDDS). An LH*RS file is hash partitioned over the distributed RAM of a multicomputer, e.g., a network of PCs, and supports the unavailability of any of its k ≥ 1 server nodes. The value of k transparently grows with the file to offset the reliability decline. Only the number of the storage nodes p ..."
Abstract
-
Cited by 59 (11 self)
- Add to MetaCart
(SDDS). An LH*RS file is hash partitioned over the distributed RAM of a multicomputer, e.g., a network of PCs, and supports the unavailability of any of its k ≥ 1 server nodes. The value of k transparently grows with the file to offset the reliability decline. Only the number of the storage nodes potentially limits the file growth. The high-availability management uses a novel parity calculus that we have developed, based on the Reed-Salomon erasure correcting coding. The resulting parity storage overhead is about the minimal ever possible. The parity encoding and decoding are faster than for any other candidate coding we are aware of. We present our scheme and its performance analysis, including experiments with a prototype implementation on Wintel PCs. The capabilities of LH*RS offer new perspectives to data intensive applications, including the emerging ones of grids and of P2P computing.
Evaluation of Distributed Recovery in Large-Scale Storage Systems
- IN PROCEEDINGS OF THE 13TH IEEE INTERNATIONAL SYMPOSIUM ON HIGH PERFORMANCE DISTRIBUTED COMPUTING (HPDC
, 2004
"... Storage clusters consisting of thousands of disk drives are now being used both for their large capacity and high throughput. However, their reliability is far worse than that of smaller storage systems due to the increased number of storage nodes. RAID technology is no longer sufficient to guarante ..."
Abstract
-
Cited by 40 (9 self)
- Add to MetaCart
Storage clusters consisting of thousands of disk drives are now being used both for their large capacity and high throughput. However, their reliability is far worse than that of smaller storage systems due to the increased number of storage nodes. RAID technology is no longer sufficient to guarantee the necessary high data reliability for such systems, because disk rebuild time lengthens as disk capacity grows. In this paper, we present FAst Recovery Mechanism (FARM), a distributed recovery approach that exploits excess disk capacity and reduces data recovery time. FARM works in concert with replication and erasure-coding redundancy schemes to dramatically lower the probability of data loss in large-scale storage systems. We have examined essential factors that influence system reliability, performance, and costs, such as failure detections, disk bandwidth usage for recovery, disk space utilization, disk drive replacement, and system scales, by simulating system behavior under disk failures. Our results show the reliability improvement from FARM and demonstrate the impacts of various factors on system reliability. Using our techniques, system designers will be better able to build multi-petabyte storage systems with much higher reliability at lower cost than previously possible.
Disk infant mortality in large storage systems
- In Proc of MASCOTS ’05
, 2005
"... As disk drives have dropped in price relative to tape, the desire for the convenience and speed of online access to large data repositories has led to the deployment of petabyte-scale disk farms with thousands of disks. Unfortunately, the very large size of these repositories renders them vulnerable ..."
Abstract
-
Cited by 16 (3 self)
- Add to MetaCart
(Show Context)
As disk drives have dropped in price relative to tape, the desire for the convenience and speed of online access to large data repositories has led to the deployment of petabyte-scale disk farms with thousands of disks. Unfortunately, the very large size of these repositories renders them vulnerable to previously rare failure modes such as multiple, unrelated disk failures leading to data loss. While some business models, such as free email servers, may be able to tolerate some occurrence of data loss, others, including premium online services and storage of simulation results at a national laboratory, cannot. This paper describes the effect of infant mortality on long-term failure rates of systems that must preserve their data for decades. Our failure models incorporate the well-known “bathtub curve, ” which reflects the higher failure rates of new disk drives, a lower, constant failure rate during the remainder of the design life span, and increased failure rates as components wear out. Large systems are vulnerable to the “cohort effect” that occurs when many disks are simultaneously replaced by new disks. Our more accurate disk models and simulations have yielded predictions of system lifetimes that are more pessimistic than existing models that assume a constant disk failure rate. Thus, larger system scale requires designers to take disk infant mortality into account. 1.
Providing high reliability in a minimum redundancy archival storage system
- Proc.14 th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems
, 2006
"... Inter-file compression techniques store files as sets of references to data objects or chunks that can be shared among many files. While these techniques can achieve much better compression ratios than conventional intra-file compression methods such as Lempel-Ziv compression, they also reduce the r ..."
Abstract
-
Cited by 12 (3 self)
- Add to MetaCart
(Show Context)
Inter-file compression techniques store files as sets of references to data objects or chunks that can be shared among many files. While these techniques can achieve much better compression ratios than conventional intra-file compression methods such as Lempel-Ziv compression, they also reduce the reliability of the storage system because the loss of a few critical chunks can lead to the loss of many files. We show how to eliminate this problem by choosing for each chunk a replication level that is a function of the amount of data that would be lost if that chunk were lost. Experiments using actual archival data show that our technique can achieve significantly higher robustness than a conventional approach combining data mirroring and intra-file compression while requiring about half the storage space. 1.
Understanding session durability in peer-to-peer storage system
- In: Proc. of the Int’l Conf. on Computational Science. 2006. 428−435. http://upstore.grids.cn/publications/iccs06_lncs.pdf
"... Abstract. This paper emphasizes that instead of long-term availability and reli-ability, the short-term session durability analysis will greatly impact the design of the real large-scale Peer-to-Peer storage system. In this paper, we use a Markov chain to model the session durability, and then deriv ..."
Abstract
-
Cited by 3 (1 self)
- Add to MetaCart
(Show Context)
Abstract. This paper emphasizes that instead of long-term availability and reli-ability, the short-term session durability analysis will greatly impact the design of the real large-scale Peer-to-Peer storage system. In this paper, we use a Markov chain to model the session durability, and then derive the session dura-bility probability distribution. Subsequently, we show the difference between our analysis and the traditional Mean Time to Failure (MTTF) analysis, from which we conclude that the misuse of MTTF analysis will greatly mislead our understanding of the session durability. We further show the impact of session durability analysis on the real system design. To our best knowledge, this is the first time ever to discuss the effects of session durability in large-scale Peer-to-Peer storage system. 1
Fault-Tolerant Real-Time Streaming with FEC thanks to Capillary Multi-Path Routing
"... Abstract – Erasure resilient FEC codes in off-line packetized streaming rely on time diversity. This requires unrestricted buffering time at the receiver. In real-time streaming the playback buffering time must be very short. Path diversity is an orthogonal strategy. However, the large number of lon ..."
Abstract
-
Cited by 3 (0 self)
- Add to MetaCart
(Show Context)
Abstract – Erasure resilient FEC codes in off-line packetized streaming rely on time diversity. This requires unrestricted buffering time at the receiver. In real-time streaming the playback buffering time must be very short. Path diversity is an orthogonal strategy. However, the large number of long paths increases the number of underlying links and consecutively the overall link failure rate. This may increase the overall requirement in redundant FEC packets for combating the link failures. We introduce the Redundancy Overall Requirement (ROR) metric, a routing coefficient specifying the total number of FEC packets required for compensation of all underlying link failures. We present a capillary routing algorithm for constructing layer by layer steadily diversifying multi-path routing patterns. By measuring the ROR coefficients of a dozen of routing layers on hundreds of network samples, we show that the number of required FEC packets decreases substantially when the path diversity is increased by the capillary routing construction algorithm. I.
An Architecture for a Scalable Distributed DBS :Application to SQL
- Server 2000. 2nd Intl. Workshop on Cooperative Internet Computing (CIC 2002
"... A table to be managed by a Scalable Distributed Database System (SD-DBS) should be able to horizontally scale over several DBMSs. The partitioning and especially its dynamic evolution should remain transparent to the application. We propose an architecture for an SD-DBS. Our architecture aims at SQL ..."
Abstract
-
Cited by 2 (1 self)
- Add to MetaCart
(Show Context)
A table to be managed by a Scalable Distributed Database System (SD-DBS) should be able to horizontally scale over several DBMSs. The partitioning and especially its dynamic evolution should remain transparent to the application. We propose an architecture for an SD-DBS. Our architecture aims at SQL Server, but generalizes to other relational DBSs. 1
Reliable Multi-Path Routing Schemes for Real-Time Streaming
"... In off-line streaming, packet level erasure resilient Forward Error Correction (FEC) codes rely on the unrestricted buffering time at the receiver. In real-time streaming, the extremely short playback buffering time makes FEC inefficient for protecting a single path communication against long link f ..."
Abstract
-
Cited by 2 (0 self)
- Add to MetaCart
(Show Context)
In off-line streaming, packet level erasure resilient Forward Error Correction (FEC) codes rely on the unrestricted buffering time at the receiver. In real-time streaming, the extremely short playback buffering time makes FEC inefficient for protecting a single path communication against long link failures. It has been shown that one alternative path added to a single path route makes packet level FEC applicable even when the buffering time is limited. Further path diversity, however, increases the number of underlying links increasing the total link failure rate, requiring from the sender possibly more FEC packets. We introduce a scalar coefficient for rating a multi-path routing topology of any complexity. It is called Redundancy Overall Requirement (ROR) and is proportional to the total number of adaptive FEC packets required for protection of the communication. With the capillary routing algorithm, introduced in this paper we build thousands of multi-path routing patterns. By computing their ROR coefficients, we show that contrary to the expectations the overall requirement in FEC codes is reduced when the further diversity of dual-path routing is achieved by the capillary routing algorithm. 1.
Reducing the Requirement in FEC Codes via Capillary Routing
- ICIS-COMSAR’06 - 5th IEEE/ACIS International Conference on Computer and Information Science
, 2006
"... In off-line packetized streaming, rateless Forward Error Correction (FEC) codes spectacularly improve the reliability of transmission over lossy networks. This success relies on time diversity, which in its turn relies on unrestricted buffering time at the receiver. In realtime streaming the playbac ..."
Abstract
-
Cited by 1 (1 self)
- Add to MetaCart
(Show Context)
In off-line packetized streaming, rateless Forward Error Correction (FEC) codes spectacularly improve the reliability of transmission over lossy networks. This success relies on time diversity, which in its turn relies on unrestricted buffering time at the receiver. In realtime streaming the playback buffering time is very limited (shorter than one second) and even strong FEC codes cannot protect single path communication against failures lasting longer than the buffering time at the receiver. Path diversity is a strategy that is orthogonal to time diversity and can make FEC applicable also in case of limited buffering time of real-time streaming. In this paper we introduce capillary routing algorithm offering layer by layer a wide range of multi-path routing topologies of increasing path diversity. We introduce Redundancy Overall Requirement (ROR), which for a given multipath routing is the coefficient of the total number of redundant FEC packets to be transmitted by the sender for protecting the communication against nonsimultaneous link failures. A dozen of capillary routing layers, built on several hundreds of network samples obtained from a random walk wireless Mobile Ad-Hoc Network (MANET), are rated with ROR. We show that flow diversity patterns built by capillary routing algorithm reduce substantially the amount of FEC codes required for protection of communication. 1.
