Operations managers Intelligent workload placement used to be a challenge. Systems such as VMware’;s DRS provided simplicity. Operations managers take workload placement to another level. Ops Managers maintain a holistic view of DC performance and health. They have a large base of information to make decisions on. The more sources of data the better the output. Detection of possible failure conditions can trigger an automatic evacuation. OPs managers provide a range of capabilities outside of workload placement.
Traditional application architecture is not optimal for cloud services. They are big, singular and designed to be always on. Cloud requires small, mobile applications which handle failure without impact. Cloud is constantly failing. Systems are crashing, misconfigurations etc. We see it as stable through application design. When we interact with cloud systems, we are interacting with systems designed for failure. When there is a failure, the application is able to remain functional.
Data gravity imposes a significant constraint on assessing where a workload should run. Workloads perform better when close to the source of data. Data latency has a tangible cost associated. The timing might be in milliseconds, but the impact can be significant. A business loses money when employees are waiting. Customers are not patient and will look elsewhere if services are slow. Storage vendors make significant investments to lower storage latency.