The principle of redundancy in information technology (IT) serves to increase the reliability and availability of systems. In general, redundancy can be intentional or unintentional. As a security concept, it refers to the (intentional) provision of additional resources or components that are intended to ensure the functionality of a system in the event of a failure.
The opposite of this multiple design is called a Single Point of Failure (SPoF): Here, the failure of a single component (for example, a server) leads to the failure of the entire system, as there is no backup. Unintentional redundancy often occurs when unnecessary data duplicates exist: These occupy storage space and complicate data maintenance, which can lead to inconsistencies.
Purpose
The redundancy principle aims to increase the availability, reliability, and fault tolerance of IT systems. By implementing redundant systems, companies ensure that their services remain available even in the event of hardware or software failures. Redundancy thus contributes to avoiding downtime and ensuring continuous business operations. As such, the redundancy principle should also be part of a Business Continuity Plan (BCP).
Types
Data redundancy
Redundant data are copies or duplicates of data sets that are either mirrored or distributed across different locations and servers. This practice serves to increase the availability and security of the data. Through various technologies, regular backups, virtualization, or mirroring, it is ensured that no data is lost in the event of hardware failures or other damages. Additionally, redundant data structures enable faster access to information over greater distances and support strategies such as backup and disaster recovery.
Functional redundancy
Functional redundancy means that multiple systems or components perform the same function to increase availability. An example is server redundancy: Here, multiple servers are operated in parallel, so that if one server fails, the required tasks are taken over by the other servers. Redundant servers are operated either in active or passive mode: In active mode, all servers share the load, while in passive mode, a server is only activated when another fails. This increases fault tolerance and significantly reduces the risk of downtime.
Network redundancy ensures that data transmission is not interrupted even if a network segment fails by implementing multiple network connections and paths.
Geo-redundancy and redundancy in the data center
Geo-redundancy involves distributing data and services across multiple geographically separated locations. This distribution thus protects against large-scale outages that could be caused by natural disasters, regional power outages, or other severe incidents. By spatial separation, the risk of a single event affecting the entire operation is minimized (compare also SPoF/Single Point of Failure). The principle of geo-redundancy is applied in data centers, among other things, by distributing them across multiple locations, often in different countries or even on different continents.
Redundancy is also crucial within the data center and refers to the multiple design of technical components and the design of the infrastructure. It can include several measures:
- Redundant power supply: The use of dual power supply systems, UPS systems (uninterruptible power supply), and emergency generators ensures that energy continues to be available in the event of a failure of the primary power source.
- Cooling: Redundant cooling systems ensure continuous cooling of the servers, even if one cooling system fails. Overheating can cause significant damage to hardware and interrupt operations.
- Network components: Multiple network paths and redundant network devices ensure that data traffic can continue unimpeded even if one part of the network fails.
These measures help ensure that the services of a data center remain available even in the event of individual component failures. The combination of geo-redundancy and local infrastructure redundancy thus increases security in the data center.
