Here are my Cloud Capacity Myths. Please feel free to add more or even better challenge me on these!
- Virtualisation and cloud will allow capacity to be turned up in an instant
- Procurement of capacity is still not instant due to internal processes
- Services and applications still take time to be built and configured even if the capacity is available
- The bottleneck may not be physical capacity, it may be an application bottleneck, the cloud doesn't address that (most bottlenecks are "logical" in my experience)
- When the capacity is turned up - what is the workload that's driving it? Whose fault was it? Without forecasting, measurement and planning one may never know
- Cloud capacity is cheap
- Temptation for ‘heavy’ and inefficient applications to be built due to the perception there that capacity is cheap which will potentially lead to cloud being more expensive than physical hardware
- The cost depends on the underlying physical hardware, if cloud is using slower processors compared to a physical servers then the cost of supporting the application will increase
- There are other costs associated with capacity, e.g. software licenses
- Virtualisation still carries a large overhead - true efficiency is arrived at when services share the same operating system and not just the same hardware