Performance impacts of autocorrelated flows in multi-tiered systems

Mi, Ningfang; Zhang, Qi; Riska, Alma; Smirni, Evgenia; Riedel, Erik

doi:10.1016/j.peva.2007.06.016

Cited by 65 publications

(54 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, once a job is assigned to a computing site, the associated load information (e.g., the present queue length) of that site cannot be updated immediately at the load balancer. As a result, the load balancer always submits the bursty arrivals to that top-ranked site within the delay period 2 . Consequently, significant load is incurred on that particular site, resulting in the performance degradation under bursty workloads.…”

Section: Motivationmentioning

confidence: 99%

See 1 more Smart Citation

ArA: Adaptive resource allocation for cloud computing environments under bursty workloads

Tai

Zhang

et al. 2011

30th IEEE International Performance Computing and Communications Conference

Self Cite

View full text Add to dashboard Cite

Abstract-Cloud computing nowadays becomes quite popular among a community of cloud users by offering a variety of resources. However, burstiness in user demands often dramatically degrades the application performance. In order to satisfy peak user demands and meet Service Level Agreement (SLA), efficient resource allocation schemes are highly demanded in the cloud. However, we find that conventional load balancers unfortunately neglect cases of bursty arrivals and thus experience significant performance degradation. Motivated by this problem, we propose new burstiness-aware algorithms to balance bursty workloads across all computing sites, and thus to improve overall system performance. We present a smart load balancer, which leverages the knowledge of burstiness to predict the changes in user demands and on-the-fly shifts between the schemes that are "greedy" (i.e., always select the best site) and "random" (i.e., randomly select one) based on the predicted information. Both simulation and real experimental results show that this new load balancer can adapt quickly to the changes in user demands and thus improve performance by making a smart site selection for cloud users under both bursty and non-bursty workloads.

show abstract

Section: Motivationmentioning

confidence: 99%

“…Bursty workloads are often found in multi-tier architectures, large storage systems, and grid services [1], [2], [3]. Internet flash-crowds and traffic surges are familiar examples of bursty traffic, where bursts of requests are aggressively clustered together during short periods and thus create spikes with extremely high arrival rate.…”

Section: Introductionmentioning

confidence: 99%

ArA: Adaptive resource allocation for cloud computing environments under bursty workloads

Tai

Zhang

et al. 2011

30th IEEE International Performance Computing and Communications Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…When the number of EBs is beyond 100, it becomes not obvious which server is responsible for the bottleneck: the average CPU utilizations of two servers are about the same, differing by 10%. In presence of burstiness in the service times, this may suggest that a phenomenon of bottleneck switching occurs between the front and the database servers across time [13]. That is, a server may become the bottleneck while processing consecutively large requests, while being lightly loaded during the other periods.…”

Section: Bottleneck Switch In Tpc-wmentioning

confidence: 99%

“…However, can we really use queueing models based only on mean service times to predict performance if the workloads are bursty or highly-variable? The answer is often negative: burstiness and high-variability can critically degrade performance to an extent that cannot be captured using mean service times only [13,2]. Therefore, describing burstiness in performance models of contemporary systems and finding measurement and parameterization techniques that remain simple and practical to use are important open challenges.…”

Section: Introductionmentioning

confidence: 99%

How to parameterize models with bursty workloads

Casale

Cherkasova

et al. 2008

SIGMETRICS Perform. Eval. Rev.

Self Cite

View full text Add to dashboard Cite

Although recent advances in theory indicate that burstiness in the service time process can be handled effectively by queueing models (e.g., MAP queueing networks [2]), there is a lack of understanding and of practical results on how to perform model parameterization, especially when this model parameterization must be derived from limited coarse measurements as is often encountered in practice. We propose a new modeling methodology based on the index of dispersion of the service process at a server, which is inferred by observing the number of completions within the concatenated busy periods of that server. The index of dispersion together with other measurements that reflect the "estimated" mean and the 95th percentile of service times are used to derive a MAP process that captures well burstiness and variability of the true service process, despite inevitable inaccuracies that result from inexact measurements. Detailed experimentation on a TPC-W testbed where all measurements are obtained via a commercially available tool, the HP (Mercury) Diagnostics, shows that the proposed technique offers a simple yet powerful solution to the difficult problem of inferring accurate descriptors of the service time process from coarse measurements. Experimental and model prediction results are in excellent agreement and argue strongly for the effectiveness of the proposed methodology under bursty or simply variable workloads.

show abstract

“…Replication is commonly employed to improve the performance [1] of such multi-tier architectures [2], [3], as well as the performance of other distributed environments, including databases [4]. In all cases, multiple replicas of the same data are maintained to tolerate failures, or concurrently serve requests and consequently balance the load among multiple machines.…”

Section: Introductionmentioning

confidence: 99%

Consistent Replication in Distributed Multi-Tier Architectures

Repantis¹,

Iyengar²,

Kalogeraki³

et al. 2011

Proceedings of the 7th International Conference on Collaborative Computing: Networking, Applications and Worksharing

View full text Add to dashboard Cite

Abstract-Replication is commonly used to address the scalability and availability requirements of collaborative web applications in domains such as computer supported cooperative work, social networking, e-commerce and e-banking. While providing substantial benefits, replication also introduces the overhead of maintaining data consistent among the replicated servers. In this work we study the performance of common replication approaches with various consistency guarantees and argue for the feasibility of strong consistency. We propose an efficient, distributed, strong consistency protocol and reveal experimentally that its overhead is not prohibitive. We have implemented a replication middleware that offers different consistency protocols, including our strong consistency protocol. We use the TPC-W transactional web commerce benchmark to provide a comprehensive performance comparison of the different replication approaches under a variety of workload mixes.

show abstract

Performance impacts of autocorrelated flows in multi-tiered systems

Cited by 65 publications

References 13 publications

ArA: Adaptive resource allocation for cloud computing environments under bursty workloads

ArA: Adaptive resource allocation for cloud computing environments under bursty workloads

How to parameterize models with bursty workloads

Consistent Replication in Distributed Multi-Tier Architectures

Contact Info

Product

Resources

About