Evaluating the impact of fine-scale burstiness on cloud elasticity

Elastic-PPQ: A two-level autonomic system for spatial preference query processing over dynamic data streams

“…Burstiness may happen at multiple time-scales and is measured using second-order properties such as the index of dispersion or the Hurst parameter [22,23,10]. Fig.…”

“…As demonstrated in this paper, frequent sharp rises in the traffic volume cannot always be neutralized with resource scaling only. The impact of such kind of burstiness in traditional Cloud computing infrastructures has been studied from the consumer's perspective in [10]. This demonstrated that burstiness may be detrimental for the elastic behavior of Cloud systems.…”

“…Most of the past elastic approaches have been experimented assuming workload variations in the form of deterministic trends and non-stationarities observed at timescales of tens of seconds [3] or even minutes/hours [9]. The work in [10] demonstrated that in elastic Clouds the presence of fast variabilities such as burstiness may be detrimental both for the system QoS and for the resource utilization rate. This negative effect also occurs in stream processing systems, where phases with arrivals occurring in clusters (temporal burstiness) are likely to generate a growing computational burden to the system.…”

IOTune: A G-states Driver for Elastic Performance of Block Storage

“…Then, IOTune periodically makes tuning decisions. By default, we set the tuning interval as one second, existing work [15] as well as our trace analysis show that a lot of I/O bursts last only for a few seconds. We believe a fine granularity tuning is necessary to timely satisfy the resource demands of these short I/O bursts.…”

“…First, the static provisioning cannot adapt to the workload variability and unpredictability. Production workloads are bursty and fluctuant [11,12,13,14,15,16]. Peak I/O rates are usually more than one order of magnitude higher than average rates.…”