Response time analysis of parallel computer and storage systems

In cloud storage systems, hot data is usually replicated over multiple disks/servers in order to accommodate simultaneous access by multiple users as well as increase the fault tolerance of the system. Recent cloud storage research has proposed using availability codes, which is a special class of erasure codes, as a more storagee cient way to store hot data. ese codes enable data recovery from multiple, small disjoint groups of servers. e number of the recovery groups is referred to as the availability and the size of each group as the locality of the code. Up till now, we have very limited knowledge on how code locality and availability a ect data access time. Data download from these systems involves multiple fork-join queues operating in-parallel, making the analysis of access time a very challenging problem.In this paper, we present an analysis of average data access time in storage systems employing simplex codes, which are an important, in certain sense optimal, class of availability codes. We generalize the analysis for codes with locality 2 and any degree of availability. Speci cally, using a queueing theoretic approach, we derive bounds and approximations on the average response time for two di erent Poisson request arrival models. We also compare two scheduling strategies for reduced access time and load balancing.

show abstract

“…in PK formula gives (27). e tighter the bound ρ on ρ i 's is, the be er the estimates E[V (ρ)] and E[V 2 (ρ)] are, so E[T (ρ)] is.…”

Section: And (20) Same Analysis Can Be Done To Showmentioning

confidence: 99%

Simplex Queues for Hot-Data Download

Aktas

Najm

Soljanin

2017

SIGMETRICS Perform. Eval. Rev.

View full text Add to dashboard Cite

show abstract

“…In parallel processing systems, there are two useful metrics for describing performance: subtask dispersion [9,7] (difference in time between the subtask that completes first and the subtask that completes last) and task response time [5,11] (time for all the subtasks of the task to complete). If subtask dispersion and task response time are both important, a trade-off product metric can be used [10].…”

Section: Introductionmentioning

confidence: 99%

Optimising Hidden Stochastic PERT Networks

Pesu¹,

Knottenbelt²

2017

Proceedings of the 10th EAI International Conference on Performance Evaluation Methodologies and Tools

View full text Add to dashboard Cite

This paper introduces a technique for minimising subtask dispersion in hidden stochastic PERT networks. The technique improves on existing research in two ways. Firstly, it enables subtask dispersion reduction in DAG structures, whereas previous techniques have only been applicable to single-layer split-merge or fork-join systems. Secondly, the exact distributions of subtask processing times do not need to be known, so long as there is some means of generating samples. The technique is further extended to use a metric which trades off subtask dispersion and task response time.

show abstract

“…However, there are numerous analytical queueing network approximations (e.g. [5][6][7][8][9]). In [4,10,11] we have developed approximate analytical queueing models of RAID 01 and 5.…”

Section: Introductionmentioning

confidence: 99%

Modelling Zoned RAID Systems Using Fork-Join Queueing Simulation

Lebrecht

Dingle

Knottenbelt

2009

Computer Performance Engineering

View full text Add to dashboard Cite

Abstract. RAID systems are ubiquitously deployed in storage environments, both as standalone storage solutions and as fundamental components of virtualised storage platforms. Accurate models of their performance are crucial to delivering storage infrastructures that meet given quality of service requirements. To this end, this paper presents a flexible fork-join queueing simulation model of RAID systems that are comprised of zoned disk drives and which operate under RAID levels 01 or 5. The simulator takes as input I/O workloads that are heterogeneous in terms of request size and that exhibit burstiness, and its primary output metric is I/O request response time distribution. We also study the effects of heavy workload, taking into account the request-reordering optimisations employed by modern disk drives. All simulation results are validated against device measurements.

show abstract

Response time analysis of parallel computer and storage systems

Cited by 45 publications

References 22 publications

Simplex Queues for Hot-Data Download

Simplex Queues for Hot-Data Download

Optimising Hidden Stochastic PERT Networks

Modelling Zoned RAID Systems Using Fork-Join Queueing Simulation

Contact Info

Product

Resources

About