More resources for workload balancing and distribution
Virtual workload balancing best practices
The resource needs of virtualization and virtual workloads
Things to avoid when load balancing servers
Q. How do individual workloads affect hypervisor performance? Can a single server tackle today’s applications?
A modern server can handle just about anything, but it certainly cannot handle everything. As a result, IT professionals often need to plan workload distribution and consider workload and resource balancing across virtualized servers.
Running multiple simultaneous workloads can strain a server’s resources and lead to performance problems. For example, running too many workloads may consume more memory, CPU cycles or network bandwidth than the system physically has available. Virtualization can mask some of that contention by abstracting workloads from the underlying hardware, but the hypervisor cannot provide more resources than the physical server provides.
A virtual server can handle SQL, Exchange and Web servers, but not necessarily all at the same time on the same system. Determine the resource needs of each workload, and spread them out so that complementary workloads rely on the same system. Furthermore, putting multiple memory-intensive workloads on the same server may result in a memory shortage. By mixing workloads with differing resource needs, you experience less resource contention and better consolidation while preserving performance.
This was first published in April 2013