post_parent): $temp_content = $post->post_content; $temp_content = explode("

",$temp_content); $temp_content = $temp_content[1]; $temp_content = explode("

",$temp_content); $temp_content = $temp_content[0]; $temp_content = strip_tags($temp_content); $temp_content = trim($temp_content); $authors = explode(",",$temp_content); ?> post_title));?>"> $value): ?> "> "> "> post_content); while ($parser->parse()) { if (($parser->iNodeName=="a")&&(substr_count($parser->iNodeAttributes['href'],".pdf")>0)): ?>

Vovk Ye., Ksenzovets’ D.

Data centers (DC) have become widespread in our life today. It is integral part of Information technologies infrastructure (IT-infrastructure).
One of the main IT-infrastructure management problems is the one of correct scheduling of virtual machines (VMs) between physical machines (PMs).
The dominant trend in resource allocation is the use of Model Predictive Control. In the article the new based on forecast mechanism of IT-infrastructure resources and load allocation have been developed. The IT-infrastructure resources and load allocation determines the hard requirements to time and accuracy of forecast. There are a lot of forecasting models which can be used for IT-infrastructure management, for example, ARIMA, GARCH and NARX. But for their effective using for resources and load allocation one needs to know how good are they working to solve this tasks according to requirements. In the article above mentioned models of forecast have been explored on the real data of DC. Results of exploration have been summarized and propositions for integration of forecast models in mechanism of IT-infrastructure resources and load allocation have been worked out. The Google cluster data have been used to analyze the forecast models.

Download (pdf)