Cloud-computing shares a common pool of resources across customers at a scale that is orders of magnitude larger than traditional multiuser systems. Constituent physical compute servers are allocated multiple ‘virtual machines' (VMs) to serve simultaneously. Each VM user should ideally be unaffected by others’ demand. Naturally, this environment produces new challenges for the service providers in meeting customer expectations while extracting an efficient utilization from server resources. We study a new cloud service metric that measures prolonged latency or delay suffered by customers. We model the workload process of a cloud server and analyze the process as the customer population grows. The capacity required to ensure that the average workload does not exceed a threshold over long segments is characterized. This can be used by cloud operators to provide service guarantees on avoiding long durations of latency. As part of the analysis, we provide a uniform large deviation principle for collections of random variables that is of independent interest.

The clown loach Chromobotia macracanthus, endemic to Indonesia, is a major species on the international market of ornamental freshwater fish. In order to satisfy an increasing demand with a sustainable alternative to the massive capture of wild juveniles, research has been dedicated to the artificial propagation and domestication of this species. The present study, the first of a series, focused on favourable maintenance conditions for broodfish sexual maturation, criteria for identification of ripe fish, efficiency of hormone-induced breeding treatments, predictability of their latency response, and on the comparison of reproductive performances of fish from populations of Sumatra and Borneo Islands (in total, 112 females of 46 to 404 g body weight). When reared in fully controlled conditions in large water recirculation systems, broodfish originating from Sumatra had reproductive performances similar to or slightly higher than those maturing in the wild (ovulation rate of 93% vs. 82%, relative fecundity of 109 277 vs. 103 550 ova  kg-1 and fertilization rate of 73% vs. 61%, respectively). In the same rearing conditions, captive females from Borneo (n = 22) showed lower ovulation rate (77%), relative fecundity (76 262 ova kg-1) and fertilization rate (50%) than those originating from Sumatra (n = 28). By contrast, the mean individual weight of ova (around 0.8 mg) was independent from the origin or maintenance conditions of females. An initial modal follicle diameter  ≥1.02 mm generally led to high ovulation success (>80%) after hormonal treatment and is recommended as the main criterion for selecting female broodfish. Two hormonal treatments for inducing oocyte maturation and ovulation (T1: two successive injections of Ovaprim at a 6 h-interval; T2: one injection of human chorionic gonadotropin (hCG)- and one of Ovaprim 24 h later), produced similar results in terms of ovulation rate, quantity and quality of ova collected. With both treatments, the latency decreased with increasing water temperature, then increased again at temperatures  >28–29 °C. To our knowledge, such U-shaped relationship between the latency response and temperature has never been documented in teleost fishes.

Experiments were designed to investigate the suitability of a combination of a short manual teat stimulation with a short latency period before teat cup attachment to induce and maintain oxytocin release and milk ejection without interruption. In Experiment 1, seven dairy cows in mid lactation were manually pre-stimulated for 15, 30 or 45 s, followed by either 30 s or 45 s of latency period. It was shown that all treatments induced a similar release of oxytocin without interruption until the end of milking. In particular, the latency period of up to 45 s did not cause a transient decrease of oxytocin concentration. In Experiment 2, milking characteristics were recorded in seven cows each in early, mid, and late lactation, respectively. Because the course of milk ejection depends mainly on the degree of udder filling, individual milkings were classified based on the actual degree of udder filling which differs between lactational stages but also between morning and evening milkings. All animals underwent twelve different udder preparation treatments, i.e. 15, 30, or 45 s of pre-stimulation followed by latency periods of 0, 30, 45, or 60 s. Milking characteristics were recorded. Total milk yield, main milking time and average milk flow rate did not differ between treatments if the degree of udder filling at the start of milking was >40% of the maximum storage capacity. However, if the udder filling was <40%, main milking time was decreased with the duration of a latency period up to 45 s, independent of duration of pre-stimulation. Average milk flow at an udder filling of <40% was highest after a pre-stimulation of 45 s followed by a latency period of another 45 s. In contrast, average milk flow reached its lowest values at a pre-stimulation of 15 s without additional latency period. However, average milk flow after a 15-s pre-stimulation increased with increasing latency period. In conclusion, a very short pre-stimulation when followed by a latency period up to 45 s before teat cup attachment remains a suitable alternative for continuous stimulation to induce milk ejection.