As most applications are licensed by the numbers of host CPUs, there is a concern on the cost of hosting applications on fixed physical core systems. Virtual systems now provide that flexibility of changing the number of vCPUs on a running system or in most cases after a reboot.
KPIs
Due to this additional complexity, effective Capacity Management is even more important, more specifically where services are sharing both physical and virtualized resources. With the numbers of systems being virtualized on the increase year on year, there is a number of Key Performance Indicators (KPI) we should be closely monitoring:
– Reduction in physical estate (servers)
– Reduction in physical space
– Reduction of power usage, cooling, associated costs
The 1990’s and early part of the millennium witnessed a rise in the number of physical distributed systems (UNIX/Windows) to a point where physical data center space and energy demand was becoming an issue. The introduction of virtualization has helped address these issues, by allowing multiple low usage systems to be virtualized on to a single piece of hardware. This in-turn had initially started to reduce the number of physical systems hosted within a data center, which removed the need for additional space and/or allowed for that space to be reclaimed.
The reduction of physical components and servers also allowed for a reduction in power usage and if reclaiming space and/or moving systems to a small data center were feasible then the amount of power dedicated to cooling could be reduced.
On Monday I'll be taking a look at Cloud technology.
Jamie Baker
Principal Consultant
No comments:
Post a Comment