Security & data
DC Sim holds a model of your room, not your network. The browser receives rendered frames; the simulation engine, the importer and your data stay server-side in an isolated tenant. Live data arrives through an agent that only dials out. You can export your model and history, and ask for deletion, at any time.
Architecture, in five lines
- Browser: receives frames and diagrams; the web bundle contains no secrets and no engine code.
- Server: the simulation engine, the importer and the catalog run on the server; each customer is a separate organisation with its own sites, versions and history.
- Agent (for Zabbix, NetBox and future connectors inside your network): a single binary inside your network that polls your sources and sends metric values over one outbound WebSocket. No listening port. Credentials stay in its config file on your machine.
- Storage: model and readings in a Postgres-compatible database per deployment.
- Editions: the English edition is Sensaka’s, hosted in the EU; the Chinese-language edition is operated by CloudSino in China. The two do not share data.
What we do with your files and data
- Excel/CSV import: parsed server-side, values only. Formulas are never evaluated; macro-enabled, encrypted and zip-bomb files are refused; 10 MB per file, 50,000 rows per sheet.
- Manuals and spec sheets you send for the catalog: only the facts (size, weight, power, parts) are stored, with the source cited; the document itself is not kept beyond extraction unless you opt in.
- No training on customer data. Your model, telemetry and files are not used to train any model, ours or a third party’s. The AI design assistant (roadmap) will see only the catalog and the brief you type.
- Redaction (on request): before a shared walkthrough, a case study or a support session, device names, owners and tenant names are replaced by placeholders.
- Export and delete: download your model as the filled Excel template at any time; deletion requests for your model and files are honoured within 30 days.
Access
Today the product runs in walkthroughs on our servers, with demo data and the files you send. Self-serve accounts are on the roadmap; the product already separates organisations, users and roles, and a partner’s access to a customer’s site is by invitation and can be revoked by the customer. SSO (SAML/OIDC) is part of Enterprise (later). All traffic is TLS; the agent will use WSS.
What we do not do
- We do not require inbound ports, VPNs or appliances.
- We do not control anything: the agent reads; the twin simulates; nothing writes back to your UPS, PDU or BMS.
- We do not sell or share your data. Logos and names of customers appear only with written permission.
Hosting and compliance, plainly
EU hosting for the Sensaka edition. Backups per deployment. A written security summary, a data-processing addendum and the sub-processor list are available on request. Certifications are not claimed; ask and we will tell you exactly what is and is not in place.
Can the twin be installed on-premises?
The same container runs on-prem and is part of the Enterprise license (later). Ask if you need it earlier.
What will the agent send?
When live connectors ship: metric names, values, timestamps and the host label, nothing else. A dry-run flag prints exactly what would be sent before it connects anywhere.
Who can see my room?
Today, during a walkthrough, only you and Sensaka. In the product: your users and the partners you invite; Sensaka support only with your permission during a ticket.
Do you keep my spreadsheet?
The values are turned into your model. You can download the model as a filled template, and ask us to delete the model and the original file; we confirm in writing within 30 days.
Is customer data used to improve the catalog?
Only if you choose to contribute a spec sheet, and then only the public facts from it with the source cited.
One Excel file. Your room in 3D.
Fill in the Devices sheet and send it with your walkthrough request; we import it on the call. What each sheet does