Backup as a Service

It is now standard for companies to have access to various types of IT services. These are provided by appropriate IT environments consisting of, among other things, servers, arrays, workstations or applications, which can be managed either internally or entrusted to the care of specialized entities. Regardless of where the environment is physically located, there is always a risk of data loss. Physical hardware failure, software malfunction (installing a faulty update) or simple administrator or system user error do happen, and should be taken into account. Hacking attacks involving viruses encrypting data disks and demanding a ransom for the uncertain promise of regaining access are also increasingly common.

Making cyclic backups, from which lost information can be recovered using backup systems, is an essential measure for securing continuity of access to data.

The backup system consists of two components. The first is the backup server along with the backup medium. This is the place where backups will be stored, which is most often special disk arrays based on deduplication technology. Deduplicators allow (in a nutshell) to achieve huge compression ratios relative to the source data, by eliminating duplication of the same information when making copies. Nowadays, the LTO – magnetic tape – standard is being abandoned. The second component of a backup system is its administrator, who is responsible for all issues related to security copies.

Companies that are considering implementing backup must choose a model for maintaining the backup system. The first option is to invest in solutions and competence, so as to ensure that the backup system is adequate and maintained at the required level. The second option is to transfer one of the elements or the whole process to a specialized company with the appropriate competence.

Co-location or hosting

An important security requirement is that the copies be physically stored in a different location than the source data. One approach offered by SNP Poland (currently All for One Poland) is to take custody of the customer’s equipment (colocation service) so that it operates in specially prepared premises of All for One Data Centers. Not only does the client gain a guarantee of optimal operating conditions for the equipment due to controlled temperature, humidity and air quality, but also the risk caused by power outages or fire is minimized.

If the customer does not have its own hardware infrastructure, the solution may be tobuy access to a properly segregated and configured infrastructure of All for One Data Centers, depending on the demand (hosting).

In addition to the hardware layer, SNP Poland also provides IT administration services, including backup systems, thanks to its possession of qualified specialists.

Backup challenges

Companies choosing to provide security copies in-house face a number of challenges. Consideration must be given to which services and applications are most important, and how backups and other high-availability solutions will maintain business continuity. The Recovery Point Objective (RPO) and Recovery Time Objective (RTO) are important parameters that determine the SLA (Service Level Agreement), or level of service availability.

The RPO describes the moment in the past to which the system can be restored (the time of the last backup). The RPO also defines the amount of data whose loss is acceptable (e.g., data from the last 15 minutes). The RTO determines the time it takes to restore systems from the moment of failure (e.g., one hour).

It is also important to determine how many copies of an item we want to store or how long these copies are to be stored. The larger the budget we have at our disposal, the more stringent the parameters can be.

Expanding infrastructure and data growth often bring one to a limit where the assumed limiting parameters can no longer be met. Problems with system flexibility, size and performance become apparent. Expansion of an existing backup system is usually not cost-effective or technically difficult to implement, and a new system often requires new competence and time to develop and implement. When maintaining your own backup system, you need to take care of proper licensing and paying for support, i.e., vendor support.

Keep in mind that IT departments in companies are often overburdened with other tasks and it is problematic to pay extra attention to copy management. Time-consumption increases as the complexity of the environment increases: new virtual machines, databases, applications with which to integrate. A diverse environment requires knowing what to configure and how to configure it. You also have to reckon with problems occurring on an ongoing basis, which cannot always be solved without consulting specialists. Our experience shows that there is rarely a need to perform data restores, and if there is a major failure, the customer often contacts an outside company anyway for support during more difficult and global problems.

Finally, one of the most important issues that is often forgotten is periodic testing of backups. Even if you can successfully perform a backup on the object of interest, it does not mean that the restoration will always be trouble-free. In the case of databases, for example, often a problem with a backup or restore signals a problem in the database itself that has not yet been detected by administrators. Performing restoration tests from time to time involves employee time and the resources necessary to perform such a test.

Replication

Another way to secure the infrastructure is replication. For entire virtual machines, the process is as follows: Disk resources are set aside on traditional arrays, where, with a fixed RPO time, the data of entire source machines is copied. This differs from traditional backups in that only the last copy is available, and if necessary, the machine so prepared can be run to replace the original machine. This approach reduces the RTO time it takes to restore an object to practically zero.

Replication reveals its greatest advantage at the time of physical hardware failure. In an instant, the entire environment can be run on other resources, which works perfectly with the DRC (disaster recovery) service. Unfortunately, this does not protect against faults at the level of the operating system, which copies itself along with the entire virtual machine on an ongoing basis. Hence the advantage of regular backup in having multiple copies from which to perform recovery.

Backups in Ramirent

One company that makes extensive use of SNP Poland’s systems maintenance services is nationwide construction equipment rental chain Ramirent. The company has been maintaining its systems in colocation at All for One Data Centers for ten years. Originally, it was an environment backup data processing center (Disaster Recovery Center, DRC). It is now an entire production environment, and the DRC is implemented as part of the All for One Managed Cloud infrastructure .

After defining the requirements, Ramirent additionally opted for a traditional backup service (Backup as a Service), ideally complementing the replication service. In this case, an additional advantage is that the client’s source resources are available locally and the speed of copies and restores is not limited by Internet bandwidth.

Kamil Lisiecki, IT Manager, Ramirent

We are responsible for the full administration of the backup system covering the Ramirent environment, including active monitoring of anomalies and troubleshooting of any problems. All for One Data Centers also has the resources to perform disaster recovery testing of the entire environment. The backup system is fully scalable, which means that an increase in the customer’s demand for backup space or a change in a certain range of RPO and RTO parameters is not a problem.

Backup as a service allows Ramirent to fully protect its data while maintaining high availability of the backup system. The efficiency of the system ensures that the backup window in which backups occur is minimized, and the scalability means that the company does not have to worry about providing space for copies as the environment grows. A high degree of automation of backup tasks and care from a qualified team of SNP Poland (currently All for One Poland) administrators allow optimization of the entire process.

The risk of a systems outage is kept to a minimum. If necessary, efficient measures are taken immediately to restore the availability of systems and services. The above aspects translate into relieving the burden on Ramirent’s IT staff and savings from minimizing business downtime.