Skip to main content

Redundancy Layers

NetActuate's infrastructure is designed with multiple independent redundancy layers. Each layer addresses a different class of failure, and together they provide end-to-end resilience for your workloads.

Physical Layer​

The physical layer covers hardware redundancy within each data center location.

  • Redundant power feeds -- dual power supplies connected to independent power distribution units (PDUs) and backed by UPS and generator systems
  • Redundant storage -- Ceph clusters with replication across multiple storage nodes; a single disk or node failure does not cause data loss
  • Hardware sparing -- spare compute and network hardware is staged at each location for rapid replacement
  • RAID and ECC -- bare metal servers use RAID configurations for disk redundancy and ECC memory for error correction

A physical layer failure (single disk, single PSU, single NIC) is handled transparently without impact to running workloads.

Hypervisor Layer​

The hypervisor layer provides redundancy for virtualized workloads.

  • Live migration -- VMs can be migrated between physical hosts without downtime for planned maintenance
  • Automatic restart -- if a hypervisor host fails unexpectedly, VMs are automatically restarted on a healthy host
  • Resource isolation -- CPU, memory, and I/O resources are isolated between tenants so that a noisy neighbor does not affect your workloads
  • Distributed storage -- VM disks are stored on the distributed Ceph cluster, not on the local hypervisor host, so a host failure does not lose disk data

Hypervisor-level redundancy protects against individual server failures within a location.

Data Center Layer​

The data center layer addresses facility-level redundancy.

  • Multiple facilities per region -- NetActuate operates in multiple data centers across key regions, allowing you to deploy in geographically distinct facilities
  • Carrier-neutral facilities -- data centers are connected to multiple network providers, so a single carrier outage does not isolate the facility
  • Physical security -- biometric access, 24/7 security, and CCTV monitoring at all facilities
  • Environmental controls -- redundant cooling systems with N+1 or 2N configurations

You can deploy across multiple data centers within a region for facility-level redundancy, or across regions for geographic redundancy.

Network Layer​

The network layer provides redundancy for connectivity.

  • Redundant uplinks -- each location has multiple upstream transit providers and peering connections
  • BGP multi-homing -- traffic is automatically rerouted through alternate paths if a transit provider or peering session fails
  • Redundant switching -- network switches are deployed in redundant pairs with link aggregation
  • DDoS mitigation -- multi-layered DDoS protection absorbs volumetric attacks at the network edge (see DDoS Mitigation)

Network-layer redundancy ensures that connectivity to your workloads survives individual link, switch, or provider failures.

Edge Layer​

The edge layer is where NetActuate's global anycast and BGP capabilities provide application-level redundancy.

  • Anycast routing -- the same IP prefix is announced from multiple locations globally. If one location goes down, traffic automatically routes to the next nearest location.
  • ECMP load balancing -- within a single location, traffic is distributed across multiple servers. If one server fails, traffic is redistributed to the remaining servers. See ECMP Load Balancing.
  • BFD fast failover -- Bidirectional Forwarding Detection provides sub-second failure detection for BGP sessions, enabling rapid traffic rerouting.
  • Tiered anycast groups -- primary, secondary, and tertiary anycast groups provide structured failover across locations. See Redundant Anycast Groups.

The edge layer is what makes NetActuate's architecture uniquely suited for latency-sensitive, globally distributed workloads like DNS, CDN, and API services.

How the Layers Work Together​

Consider a DNS service deployed on NetActuate:

  1. Physical: each server has redundant power and ECC memory.
  2. Hypervisor: if a host fails, the VM restarts on another host automatically.
  3. Data center: the service is deployed in multiple facilities.
  4. Network: each facility has redundant uplinks and DDoS protection.
  5. Edge: anycast routes users to the nearest healthy location; ECMP distributes load within each location; BFD detects failures in milliseconds.

A failure at any single layer is absorbed by that layer's redundancy mechanism. A catastrophic failure at one location (data center + network) is absorbed by the edge layer, which routes traffic to the next nearest location.

Next Steps​

Need Help?​

Contact support@netactuate.com or open a support ticket from the portal.