📘 11-03: High Availability & Redundancy¶
📌 Availability ("Nines")¶
Availability is expressed as an uptime percentage:
| Availability | Downtime/year |
|---|---|
| 99% | ~3.65 days |
| 99.9% ("three nines") | ~8.76 hours |
| 99.999% ("five nines") | ~5.26 minutes |
Higher availability requires more redundancy — eliminating single points of failure (SPOFs).
📌 Redundancy Techniques¶
| Technique | Protects against |
|---|---|
| Redundant power supplies / UPS / generator | Power loss |
| Dual uplinks / redundant switches | Link/device failure |
| First-hop redundancy (HSRP/VRRP/GLBP) | Gateway failure |
| Link aggregation (LACP/EtherChannel) | Single-link failure + more bandwidth |
| STP/RSTP | Layer 2 loops while keeping backup paths |
| Clustering (active-passive / active-active) | Server failure |
📌 First-Hop Redundancy Protocols¶
HSRP/VRRP share a virtual gateway IP/MAC across two routers; if the active one fails, the standby takes over — hosts keep their gateway.
📌 Load Balancing¶
Distributes traffic across multiple servers/paths — improving throughput and availability (one server can fail without outage).
📌 Failover¶
Automatic switching to a redundant/standby system when the primary fails, minimizing downtime.
Tip
Key idea — High availability = remove SPOFs with redundancy: dual power/links, FHRP gateways, LACP, STP, clustering, and load balancing, with automatic failover.
See also: Fault Tolerance & Recovery, WAN Topologies & Modern WAN
Previous: Network Monitoring | All Notes | Module 11 Exercise | Next: Fault Tolerance & Recovery