Initial Findings for Provisioning Variation in Cloud Computing

Performance Variations in Profiling Mysql Server on the Xen Platform: Is It Xen or Mysql?

“…Blockback driver forwards the request to the native driver (10). Native driver reads the blocks from the disk (11,12) and send them to the blockback driver (13). Blockback driver demultiplexes blocks to the blockfront driver in dom-u (14).…”

“…In [12], [9], authors have reported performance difference between VMs of the same abstract type in data centers like Amazon EC2. Authors have credited the difference in performance to the placement of VMs and the heterogeneous hardware in evolving data centers like Amazon EC2.…”

“…Previous studies [8], [9], [10], [11], [12], [13] have reported performance variations on the Amazon EC2 cloud platform. Contention for the shared I/O resource between guest VMs hosted on the same physical machine is the primary reason for the temporal performance variations observed in a VM [8], [10], [12]. The temporal performance variations in a VM can render the profile created during one time interval to be useless in another.…”

“…The temporal performance variations in a VM can render the profile created during one time interval to be useless in another. In recent works [9], [12] authors have reported that VMs of the same abstract type can exhibit relative performance differences. Authors have attributed the difference in performance to the placement of VMs and the heterogeneity in the underlying hardware in evolving data centers like Amazon EC2.…”

Deployment Strategies for Distributed Applications on Cloud Computing Infrastructures

“…There is a distinction between those that optimize once, i.e., during the deployment phase, and those that try to optimize constantly, i.e., during the lifetime of the VMs. The focus of these papers is on different aspects of performance, such as internal communication speed [12], latency between the application and its users [13] or processing speed of the VMs [14]. The mentioned papers do not deal with the speed of the deployment process itself.…”