Workloads are computing processes that run in different environments and work to accomplish a task. Workloads run on both physical and virtual servers and can dynamically move between environments depending on computing needs. Workloads can be different sizes and serve different purposes: some are smaller processes, and some manage entire servers; some are entire applications, and some are components of one. Because they differ considerably, workloads run best in a variety of environments. Some workloads run on-premises in a physical server and some exist entirely in the cloud.
Two common types of workloads are static and dynamic workloads. A static workload stays relatively constant over a long period of time, using very similar amounts of compute resources on a steady schedule. Dynamic workloads, however, adjust dramatically as demands arise; they’re also known as temporary workloads. If a compute process arose that required multiple servers or cloud environments, a dynamic workload would require greater flexibility and availability to run efficiently.
Workloads can also run in containers. Containers provide isolation for applications and processes that are running on the same server. Containers can save resources by running more workloads and applications on one server; they’re much lighter than virtual machines, for example, and can be transferred more easily between environments.
Running workloads in a cloud environment provides added flexibility and agility for computing processes. Running workloads in multiple clouds adds another layer of compute resources: if a server is overloaded or the workload suddenly requires more resources, workloads that run in multiple clouds can move to another server or cloud that has more availability. However, not all workloads run best in the cloud; some may be better suited to on-premises servers.
Workloads require space and energy, and intelligent workload management transfers them to a server that can run them most efficiently. Load balancers, which can be found in hardware or software, distribute workloads across servers and hybrid cloud environments. This maximizes computing resources, transferring workloads to an environment where they will run most effectively.