In IT, redundancy is having extra or "redundant" components in a system so that even if one of the components fails, the system can still keep working. Redundancy in a hosting platform like Iron Orbit's ensures that you're always able to connect to and use your hosted solutions.
Redundancies can be either active or inactive—they can be active components of the IT asset and accept the additional workload of the failed component along with its own workload, or they can be inactive components that are not currently handling any workloads but that are set up to automatically boot up and take on the workload of a failed component once a failure occurs. It is critical that the switchover from the failed component to the redundant component be automatic, instant, and seamless, so that users never lose access to the IT asset. Redundancy also provides that the manager of the IT asset with time to properly repair or replace the failed component.
Almost any IT asset you can think of can be made redundant. This includes data, servers, storage devices, storage drives, networks, routers, switches, power supply, datacenter equipment, and datacenters. Examples of some of these redundancies include:
- Data. Data can be duplicated on stored on the same storage drive, the same storage device, a different storage device in the same location, and/or on a different storage device in a different location. Copying data to multiple locations ensures that the data will still be available to users even if a storage device in one location is destroyed in a localized event like a power outage, fire, flood, or earthquake.
- Networks. At Iron Orbit, we maintain internet connections with multiple different internet service providers (ISPs) at our datacenters so that our hosted solutions remain available even when one of the ISPs suffers an outage. We also utilize redundant routers, switches, and local networks.
- Power supply. Just like with our internet connections, at Iron Orbit we maintain power supplies with multiple different power companies so that our datacenters can remain online even if one of the power companies has an outage. We also have an added layer of redundancy with battery-based Uninterruptible Power Supplies (UPS) and diesel generators, which will allow our datacenters to continue operating for several days even if all of the power companies in the area suffer outages, as in a grid failure or events like Hurricane Katrina or Hurricane Sandy.
- Datacenter equipment. Iron Orbit maintains redundant HVAC and fire sprinkler systems in our datacenters. If we didn’t and one of these systems failed, all the hardware in our datacenters would overheat or be completely unprotected if a fire were to occur.
A business can make an IT asset as redundant as it wants (it can have a backup of a backup of a backup of a backup of a…, etc.), budget permitting. Most businesses just settle for either N+1 redundancy or 2N redundancy, though. N+1 redundancy is defined as having one extra component, so if you have an N+1 datacenter with 10 active servers you will have one extra server (for a total of 11) that is set up and ready to go if one of the active servers fails. 2N redundancy, on the other hand, is when you have a complete copy of an asset, so if you have a datacenter with 10 active servers in a 2N configuration you will also have another datacenter with 10 inactive servers (for a total of 2 datacenters and 20 servers). These redundancies can also be extended further in the same form: businesses can do a N+1 redundancy that is N+2 or N+5, or a 2N redundancy that is 3N or 5N.