Astra Cloud Platform 2.1 combines DR, IAM, SDN and storage

On 6 July 2026, Astra Group introduced Astra Cloud Platform 2.1 for private and public clouds. The release combines changes to virtualisation, storage, access management, networking, automation and container infrastructure. For an enterprise customer, the important point is whether these components can operate as one controlled environment.

What the announcement confirms

The Brest virtualisation component fixes CVE vulnerabilities and simplifies common operations. Administrators can deploy GPU-enabled virtual workstations for CAD, video production and heavy graphics. Disk snapshots and backup capabilities have also been added.

The main infrastructure change is a cross-site disaster-recovery scenario. Virtual machines and attached disks are replicated to a reserve data centre and restarted when the primary site or its hypervisors fail. The vendor describes automatic recovery without manual intervention.

The same platform also includes:

• TROK software-defined storage with block and file services;
• IAM for central control of permissions and access conditions;
• SDN for centralised network management;
• an initial API Gateway covering virtual-machine, disk and network lifecycles;
• the Botsman platform for customer-managed Kubernetes clusters;
• Astra Monitoring for platform-component health;
• ARM support on Baikal-S processors.

The vendor says the IAM component is undergoing FSTEC certification. This means certification was not complete on the announcement date; regulated deployments should verify the current status.

What to verify in disaster recovery

A statement about automatic recovery does not define the design parameters of a specific system. A pilot should establish:

1. target RPO and RTO for each service class;
2. replication mode and lag;
3. which objects move with the virtual machine;
4. DNS, routing, load-balancer and external-dependency behaviour;
5. the failback procedure after the incident;
6. effects on licences, secrets and integrations;
7. test frequency and evidence collection.

The announcement does not provide independent full-site-failure results or measured RPO and RTO figures. Teams need to validate those values on their own topology and data. Our disaster recovery plan explains how to structure such a test.

A practical pilot

Start with a non-critical but representative service: several virtual machines, persistent disks, a network, identities and monitoring. The team should complete the entire lifecycle—provision through the API, restrict access, create a backup, simulate failure, restore the service and return it to the primary site.

Responsibilities also need to remain explicit. A unified interface reduces manual work but does not replace redundant inter-site connectivity, protected administrator identities, configuration control or application-level recovery tests. Our IaaS shared-responsibility guide provides a useful ownership model.

Practical conclusion

Astra Cloud Platform 2.1 is moving from a collection of infrastructure components towards a unified managed environment. Cross-site recovery, central IAM, software-defined networking and storage, and an automation API are the most relevant changes for business operations.

Before production use, test them as one system with measured RPO and RTO, failed external dependencies, least-privilege access, observability and a documented failback procedure.

Source: the official Astra Group announcement dated 6 July 2026.