Active-active
An architecture in which multiple service nodes actively answer traffic at the same time. In DDI, the term can describe anycast DNS, load-balanced DHCP, or multi-primary management, but the exact behavior depends on the product.
Category glossary
Search all termsAvailability
19 definitions associated with Availability.
Active-passive
An architecture in which one node serves traffic while a standby node waits to take over after a failure. State synchronization and failover detection determine how much disruption or data loss can occur.
Dead peer detection
A mechanism used to determine that a partner, server, or network path is no longer functioning. DDI products may use heartbeats, protocol messages, routing state, or interface status to trigger failover.
DHCP failover
A protocol or product feature that allows two DHCP servers to coordinate lease ownership and continue service after a partner failure. DHCPv4 failover behavior and DHCPv6 failover use different standards and implementations.
Failback
The controlled return of service from a backup or surviving node to the preferred normal topology after a failure is resolved. Failback should account for state synchronization, traffic convergence, and client caching.
Failover
The transfer or continuation of service when a component, site, process, or network path fails. DDI failover can rely on HA clustering, DHCP state sharing, anycast routing, secondary zones, or managed-service distribution.
GMC
Grid Master Candidate, an Infoblox member configured to take over the Grid Master role if the active Grid Master fails. It should be sized, licensed, synchronized, and operationally tested for takeover.
HA pair
Two nodes configured as a high-availability unit, normally presenting a shared service identity and synchronizing state. Infoblox NIOS HA pairs use active and passive nodes with HA communication and service VIP behavior.
Heartbeat
A periodic signal used to show that a node, peer, or service is alive and communicating. Missing heartbeats can trigger alarms, state changes, or failover.
High availability (HA)
A design objective and set of mechanisms that reduce service interruption when components fail. HA can include redundant nodes, state replication, anycast, routing convergence, secondary zones, and tested recovery procedures.
Hold-down timer
A waiting period during which a resource, route, lease, or state transition is not immediately reused or reversed. Hold-downs reduce flapping and allow stale information to age out.
Hot standby
A DHCP failover mode in which one server normally owns most or all active service while the partner is ready to take over. It is often chosen when servers are in separate sites or one is intended mainly for disaster recovery.
Load balance
A failover mode or traffic design in which multiple servers actively handle clients according to a hash, split, relay selection, or load balancer. For DHCP, shared lease state and consistent client handling are critical.
Maximum client lead time (MCLT)
A DHCP failover safety interval limiting how far one partner can extend a lease beyond information known to the other. It reduces the chance of duplicate allocation after communication or partner failure.
Partner down
A failover state in which one server is administratively or automatically authorized to assume broader responsibility because its peer is considered unavailable. Entering the state incorrectly can create duplicate allocation risk.
Quorum
The minimum set of nodes or votes required to make authoritative cluster decisions. Quorum prevents split-brain writes but may sacrifice management availability when too many members are unreachable.
Redundancy
Duplication of components, paths, data, or services so a single failure does not cause total outage. Redundancy only improves availability when failure modes are independent and recovery is tested.
Split brain
A failure in which peer nodes lose communication and both act as the active authority, potentially creating conflicting leases or configuration. Quorum, fencing, MCLT, and operational procedures reduce the risk.
VRRP
Virtual Router Redundancy Protocol, which lets routers or appliances share a virtual IPv4 or IPv6 address and elect an active master. Some HA DDI systems use VRRP-like service VIP behavior.