|
|
|
Virtual machine software makes it possible to ``encapsulate''
different applications in individual virtual machines so that a failure
in one virtual machine does not affect other virtual machines hosted
on the same physical hardware -- this is fault isolation.
However, performance isolation is another important goal.
Individual virtual machines are often configured with performance
guarantees and expectations, e.g., based on service level agreements.
Thus, the resource consumption of one virtual machine should not
impact the promised guarantees to other virtual machines on the same
hardware.
Unfortunately, current virtual machine technology does not provide
effective performance isolation.
The goal is to design a set of cooperating mechanisms to effectively
control total CPU consumption across hosted virtual machines.
Effective management of virtualized IT environments introduces new and unique
requirements such as dynamically resizing and migrating virtual machines in
response to changing application demands. Such capacity management methods
should work in ensemble with the underlying resource management mechanisms.
Using Xen and its three different CPU schedulers, we analyze the
impact of the choice of scheduler and its parameters on application
perfomance and discuss challenges in estimating the application resource
requirements in virtualized environments.
Related Papers and Reports
- L. Cherkasova, D. Gupta, A. Vahdat: Comparison of the Three CPU Schedulers in Xen.
ACM SIGMETRICS Performance Evaluation Review, (PER), Vol. 35, No. 2, pp. 42-51, September 2007.
- D. Gupta, L. Cherkasova, R. Gardner, A. Vahdat: Enforcing Performance Isolation Across Virtual Machines in Xen. In Proc. of the ACM/IFIP/USENIX 7th International Middleware Conference (Middleware'2006), Melbourne, Australia, Nov. 27 - Dec.1, 2006.
- L. Cherkasova, R. Gardner: Measuring CPU Overhead for I/O Processing in the Xen Virtual Machine Monitor. In Proc. of 2005 USENIX Annual Technical Conference (USENIX'2005), Anaheim, April 10-15, 2005.
- D. Gupta, R. Gardner, L. Cherkasova:
XenMon: QoS Monitoring and Performance Profiling Tool. HP Laboratories Report No. HPL-2005-187, 2005.
|