Guide
How much downtime and data loss your business can accept
How long can you go without your tools, and how much work can you afford to lose? These two figures are a management decision. Here is how to set them, tool by tool, and what they change in your server architecture.
Updated on 18 October 2026 · Figures are sourced at the bottom of the page; opinions are those of IT CyberWall.
A management decision
How long can your business go without its tools? How much work can it afford to lose? These are not technical settings: they are management choices, which then determine the server architecture and the budget. The SGDSN guide on business continuity requires the impacts of an interruption to be measured on a common scale approved by management[1].
Two figures to set
- The maximum tolerable downtime (RTO, known in French as DMIA): how long a tool can be unavailable before the business suffers[1][2].
- The maximum tolerable data loss (RPO, known in French as PDMA): the point in time to which data must be restorable after a failure; if the last copy dates from last night, the day’s work is lost[1][3].
These two concepts are put into practice by your backup and recovery provider: WeDoBack explains them in detail in RPO and RTO explained.
Which tools are critical?
The SGDSN suggests one sheet per critical activity: its acceptable interruption time, its criticality, the resources it depends on and its interdependencies, with human, financial, legal and reputational impacts[1]. For an SME, a table is enough to start with. An example to adapt:
| Tool | If the tool stops | Tolerable downtime | Acceptable loss |
|---|---|---|---|
| No more orders or quotes | Half a day | None | |
| Management or production software | Workshop or invoicing at a standstill | A few hours | One hour of data entry |
| File server | Folders inaccessible | One day | The current day |
| Accounting | Delayed closing | Several days | One day |
Fictitious values, given as an example: only your management can set them.
What your answers change in the architecture
| Tolerable downtime | Server architecture |
|---|---|
| Several days | One server, and a backup that allows it to be rebuilt. |
| A few hours | Virtualised servers, which restart on another machine without reinstalling everything. |
| Almost none | Several physical servers, with automatic restart on another machine in the event of a failure, and often recovery servers at another site. |
These architectures are described on our Virtualised architecture page. Acceptable data loss, for its part, translates into backup frequency and a recovery plan: that is the job of WeDoBack, which describes the DRP and BCP in its IT continuity guide.
Who does what
| IT CyberWall | WeDoBack |
|---|---|
| Helps management list the critical tools and set the two figures. | Implements the disaster recovery plan and the business continuity plan. |
| Designs the server architecture accordingly. | Sets the backup frequency, organises restoration tests and recovery servers. |
Let's set your two figures together
We list your critical tools with you and propose the architecture that matches.
Sources
Sources consulted on 18 October 2026. Supplier prices are shown excluding tax, in the currency and region of the published price list, and may change: always check the official page before deciding.
- SGDSN – Guide pour réaliser un plan de continuité d’activité (2013), sheets 6 and 8 — armement.defense.gouv.fr/sites/default/files/2022-04/20130612_NP_SGDSN_guide_pca.pdf
- NIST – Glossary, Recovery Time Objective (SP 800-34 Rev. 1) — csrc.nist.gov/glossary/term/rto
- NIST – Glossary, Recovery Point Objective (SP 800-34 Rev. 1) — csrc.nist.gov/glossary/term/recovery_point_objective
Backup and restoration are operated by WeDoBack.