作者: Kapil Kumar , Nehal J. Wani , Suresh Purini
关键词:
摘要: The memory and core requirements of a virtual machine depend on the performance applications hosted it. In this paper, we propose algorithms for dynamic scaling using combination learning feedback control techniques. These work sequential parallel such as scientific computations where speedup is primary metric. Then use these to address simultaneous allocation problem, which more complex due possible correlation between resource requirements. All can be applied in black box fashion without instrumenting source code applications.