How should I compare shared and dedicated CPU VPS plans?
Shared CPU plans schedule your workload alongside other guests, while dedicated or reserved CPU offers depend on the provider’s exact definition. Compare the documented allocation and sustained workload requirements instead of assuming the names imply identical guarantees everywhere.
Instructions
1. Describe the workload clearly
Record expected concurrency, peak duration, latency sensitivity and whether the application needs sustained compute. A mostly idle administration service and a continuous encoder have different allocation needs.
2. Ask precise allocation questions
Confirm whether the plan reserves CPU time, pins virtual processors, shares sibling hardware threads, applies burst credits or imposes fair-use limits. These details are host-side policy and cannot be proven solely from a guest command.
3. Compare an agreed representative measurement
Use application metrics over a meaningful period and note simultaneous memory or storage pressure. If Linux sysstat is already installed, collect a bounded CPU sample.
mpstat 1 5This samples CPU statistics five times. Review guest demand and any exposed steal time; one sample is not a provider-wide performance guarantee.
Check the result
- The chosen plan’s written allocation matches the workload’s sustained and peak needs.
- Measurements cover relevant busy periods and application response, not only synthetic CPU counts.
Troubleshooting
- The advertised terms are ambiguous
- Request a written explanation of reservation, scheduling and any throttling policy.
- A dedicated allocation is still slow
- Inspect application efficiency, disk latency, networking and guest configuration as well as the host allocation.
Official references
Use the documentation for your installed version when a flag, service name or platform feature differs.