Glossary
Cloud cost management
Cloud cost management is the practice of tracking, assigning, forecasting and reducing cloud spending so that every charge belongs to a team or workload. For storage, the surprises are seldom in capacity. They come from requests, retrievals, transfers out and early-deletion penalties.
How is cloud spend attributed to teams?
It starts with tagging. Each resource carries an owner, a project and an environment, and charges roll up by those labels. Showback reports what each team spent, with no money moving. Chargeback goes one step further and puts the cost into that team's budget. Both depend on tags being complete: untagged spend has no owner, and in most accounts it is the first large number a review turns up.
The FinOps cycle runs in three phases. Inform makes spend visible and attributed. Optimize acts on it, by removing idle resources or changing tiers. Operate builds the decisions into routine work. Unit costs, such as cost per terabyte stored or per customer served, make the totals comparable from one month to the next, because a bill that doubled alongside a doubling of data means something different from one that doubled with flat data.
Why is the storage line the surprising one?
The capacity price is the part everyone sees on the pricing page. The bill is assembled from more parts, and the other parts follow how data is used, not how much of it exists.
| Charge | What triggers it | Where it surprises |
|---|---|---|
| Capacity | Gigabytes stored per month | Rarely the surprise; grows with data |
| Requests | Each read, write and listing call | Applications that list or scan large buckets |
| Retrieval | Reading from cool or archive tiers | Data filed as cold that is still read |
| Egress | Data leaving the provider or a region | Restores, migrations, analytics in another region |
| Early deletion | Removing data before a tier minimum ends | Short-lived data placed in an archive tier |
What does modeling the access pattern mean?
Choosing a tier from capacity alone is the common error. Archiving data that is still read can cost more than leaving it in a warmer tier, and running compute in one region against data in another adds transfer charges that can exceed the storage itself. The inputs that matter are reads per month, the share of data touched, and where the readers sit. Egress fees and the tier rules in cloud archive storage are the two lines most often left out of a forecast.
Where does owned infrastructure change the picture?
On infrastructure an organization runs itself, reads, listings and copies are not metered per request. The cost sits in hardware, power, space and staff, and it is mostly fixed once the capacity is bought. That does not make it cheaper in every case. It changes which variable drives the bill, which is why exit planning and lifecycle policies belong in the same review, along with the hybrid question of which data sits where.
Frequently asked questions
What is the difference between showback and chargeback?
Showback reports each team's spend for awareness. Chargeback bills that spend to the team's own budget, which changes behavior more but needs accurate tagging first.
How is cloud waste usually found?
Common sources are idle or oversized resources, unattached volumes, forgotten test environments and data left in a tier that no longer matches its access pattern.
Why does the storage bill rise when the data does not?
Usage fees follow activity. A new application listing buckets, a restore, or a job reading across regions raises request, retrieval or transfer charges without adding a byte of capacity.
Related terms
- Egress fees: the per-gigabyte charge for leaving a provider.
- Cloud archive storage: a tier with retrieval fees.
- Storage lifecycle management: moving data to cheaper tiers as it ages.
- Cloud automation: enforcing tags in code.
- Hybrid cloud: splitting spend across owned and public capacity.