Redundancy vs Backup

Why your backup redundancy process needs redundancy and backupIn the quest for true data security, it is important to compare the system you have in place against the possible threats you are likely to experience. We hear the terms “redundancy” and “backup” tossed around a lot and used almost interchangeably, and granted, the line between them can be difficult to distinguish at times. While both these words have strong associations with data protection, it is important to understand their differences, so you can make sure to reap the security benefits they both offer.

What is Redundancy?

Beyond hardware and software, redundancy can be thought of as a strategy to ensure data remains accessible. For instance, if I have a RAID storage system and a single drive fails, I can replace the failed drive and resume working normally with no data loss. You might argue that this only occurs because the data was “backed up” on the other drives, but (except for RAID1 mirroring), this isn’t true. No other drive in the RAID system has an exact copy of the data that was on the failed drive, so there is no “backup” in the traditional sense. The missing data only exists within the context of the RAID system, where it can be reconstructed from the remaining pieces that are strewn across multiple drives.

In hardware, true redundancy can be as simple as having an extra network switch sitting on the shelf that can be manually swapped into the network in case of a failure in the primary switch. Ironically, this would be referred to as a “backup switch”, which is why the vernacular surrounding backup and redundancy can get so confusing! Using redundant hardware in this capacity is not particularly cost-effective since you are buying two pieces but only using one. Further, in the event of a failure, it requires someone to manually swap and configure the new switch, which means there is some downtime.

Clustering Hardware and Software

A better approach to hardware redundancy is to use what is called “hardware clustering”, which involves taking multiple independent servers and turning them into a group that acts like a single system. In this case, the cluster is controlled by a single server that has been outfitted with special hardware and software that enables it to monitor the activity in the individual servers. The operating system performs load balancing and failover operations to redirect network traffic to an available server node, so there is never any downtime when a failure occurs, as long as there are enough working servers to handle the cumulative load.

Because of the specialized hardware requirements, hardware clustering can be expensive, so it is often replaced by the more cost-effective “software clustering”, which uses special software installed on each server in the group to perform the same tasks once delegated to the controlling server.

Note that in all of these cases, “redundancy” really has more to do with keeping things running after a failure than recovering lost data.

What is Backup?

A backup is an identical collection of data that represents the state of a storage system at some point in time. It is a physical collection of storage media – be it disks, drives, tapes, or cloud-based storage -- that exists completely independently of the source data. A backup guarantees you can restore the full data set to its exact state at some point in time, even if the original hardware is damaged. Backups are essential in the event of a catastrophic event, like a fire, flood, earthquake, tornado, or hurricane that can physically destroy network hardware and its redundant components. Of course, the backup needs to be protected at all costs, which is why it should not be stored on the same premises as the hardware… and ideally in a different geographic location.

While always having a current backup is essential, don’t confuse “being backed up” with “being operational”. A backup often provides the only pathway back to “operational”, but the data needs to be restored onto the local hardware before that happens. How long that process takes depends on how and where the backup is stored. If you have to get a hundred terabytes downloaded off the cloud, that could take a while, and time is money.

How to Cover All the Bases of Data Security

Thus far, we have outlined the key safety objectives, but how does redundancy and backup fit together? In daily practice, you won’t be recovering from catastrophic events (yet you always want to be prepared for such). More likely you will be recovering from a failed hard drive, bad software update, or an annoying malware attack that wreaked havoc on the network. So, what is the best way to navigate these common storms?

For mid-tier corporations, a storage area network (SAN) can provide an excellent combination of tools to keep your data safely backed up and restorable. SANs were built for redundancy, typically supporting several different RAID architectures. The SAN’s backup ability depends on the technology it supports. Once upon a time, snapshots offered the best chance of getting most of your data back, but nowadays modern storage companies, like Reduxio, offer true “Point in Time recovery”, giving you the ability to dial back your storage system to 1-second before an “unfortunate event” occurred, effectively abolishing data loss. Further, this is accomplished without a physical backup process, so you don’t lose a major chunk of your storage capacity to archived snapshots. Instead the history of the storage array is maintained as space-friendly metadata, which can be used to regenerate the array back to a previous state, right to the second.

A good SAN will give you redundancy in both your drives and your data, but you still need a backup strategy in the event of that doomsday scenario. There are multiple ways to approach this, depending on your budget. The Cadillac plan is to have another SAN unit in another geographic location that mirrors the first one. In the event one array is destroyed, operations could fail over to the redundant SAN unit with fairly little disruption. The other option is to have your SAN mirror its data out to some cloud storage. This requires more recovery time (based on the amount of data and the internet bandwidth), but for some companies it may be sufficient.

Making Sure You Are Fully Covered

Remember that there are two parts to data security: (1) You need a backup of the data that is completely independent of your current storage network, and (2) You need redundant processes in place to make sure that data remains accessible on the network when routine hardware failures occur. Recent innovations in storage technology have made it faster than ever to recover from data-loss events and even minimize the losses. Insomuch as your business relies on it, make sure to shop around and do your homework whenever data security is at stake.

To learn more about how technology is changing the way data is managed, stored, protected, and recovered, contact the storage professionals at Core Evolution Group. If you are planning a storage upgrade or just want to learn how new technologies may change the way you think about what is possible, then please download the eBook titled The Essentials when Upgrading a Storage Solution

learn the basics of upgrading your storage solution