Backups: how to back up your data successfully

A nightmare scenario for companies: important commercial data is lost after a hardware defect, a user error or a cyber-attack – and there was no backup. In the worst case, a loss of data may result in the company going out of business. Therefore, you need a sophisticated strategy for backup, recovery and archiving too.

In today’s world, companies save all their data on digital systems and manage all their business processes using software applications. Of course, reliable access to this information is essential. No company can afford to lose (important) data. Potential consequences are downtime, production stops, loss of reputation and financial losses potentially extending to millions of euros. Backups and data recovery are therefore compulsory tasks for any IT department.

However, the effort needed to manage company data is increasing. Reasons for this include large increases in data quantities, compliance requirements, as well as heterogeneous IT architecture with virtualised and physical servers. This means companies must adhere to strict requirements when storing, processing and archiving business-critical data. If companies disregard these obligations, legal difficulties concerning taxes (for example in audits), fines and other criminal proceedings may result.

Linking backups with archiving

However, a backup only partially covers legal requirements such as retention requirements with document authenticity or full-text search. Therefore, companies should always consider data backup and archiving together. Backups primarily concern short-to-medium-term storage and recovery of data, in case it should be lost or stored somewhere inaccessible. Since rapid access is necessary, hard drives or SSDs are usually used (backup to disk).

Long-term data archiving should reduce the load on the live systems and fulfil compliance requirements. Unlike conventional backups, archived data (for which online access is not or only rarely needed) is deleted from the source system. This saves costs since, after archiving, the data is located on cheaper media (generally tapes).

It depends on the type of data

How and how often data should be backed up depends on the applications, data volumes and company-specific requirements. Companies need to clarify in advance how time-critical the data to be protected are. For example, control data have a different timescale to production or CRM data, which need to be accessed every day or even every hour. What is the latest point at which data must be available once more, in order to comply with the service level, for example? Here, the keywords are Recovery Point Objective (RPO) and Recovery Time Objective (RTO).

  • Recovery Point Objective (RPO) – the RPO specifies how much data loss is acceptable to a company. For example, if the administrator performs a daily backup, the RPO will be 24 hours. This means that all data which has been changed or created in the last 24 hours could be lost.
  • Recovery Time Objective (RTO) – for how long does the user have no access to his or her data? This is the time permitted to fully restore the business processes, measured from the moment the incident occurred. Of course, this value depends on how critical the application is to the business.

Backup variants

The most common backup methods are full backup, incremental backup and differential backup.

A full backup is a copy of all the entire data stock. Due to the large quantities of data involved, this is almost impossible to do at frequent intervals since it puts enormous strain on the network and the backup server. Therefore, companies should increasingly move towards using incremental backups. After a full backup, an incremental backup only backs up the data blocks which have changed since the last backup, in order to keep the quantity of data being transferred and backed up as small as possible. This is also the advantage over a differential data backup, where all the changes since the last full backup are backed up.

As a rule of thumb, production data should usually be backed up incrementally every hour; in the accounting department one backup per day or two or three times a week is generally sufficient.

Replication and separate storage locations

Another tool for increased security is replication via snapshot, i.e. an exact copy of the entire data stock or of specific data records. For optimum recovery, replication can be used to write the same data stock to disk systems at two different locations. If one system should then fail, the second system is used. This makes data loss and downtime almost impossible. Therefore, companies should always keep their replication and backup data at two physically separate locations, and perhaps also in the Cloud.

Backup in the Cloud

Cloud backup offers an alternative to storing data at an external location. “Storage as a Service” is becoming increasingly attractive, particularly for small and medium-sized organisations. The advantage ­– no investment is required, no expertise has to be acquired, and you only have to pay for the storage you actually need to use. Cloud storage is easily scalable, enabling it to handle increasing quantities of data in the future. On the other hand, data privacy and security is an issue. Companies should consider carefully whether to take this step, after thoroughly checking the cloud provider’s conditions, security standards and user reviews.

Just in case: IBM storage solutions

As a long-term business partner, RuP GmbH works with IBM solutions covering every possible scenario – from applications with low memory requirements, rapidly-growing Cloud solutions right through to modern virtualised storage environments. IBM storage products can handle every challenge – cost effectively, flexibly and in a customer-friendly way.

We will show you the different system configurations, give you advice on their advantages and disadvantages, and help you work out which solution would be most appropriate for your business.

IBM’s comprehensive storage portfolio includes classic hard drive storageflash storagetape storage or IBM Spectrum storage as a solution for software-defined storage.