Friends, thank you for the comments, we took into account your interest in the backup repository and continue to reveal the urgent theme of backups.
Backup data is often perceived by administrators as a classic IT task, in which there is no room for experiments. It is believed that it is enough to choose one of the proven solutions and implement it. However, in practice, traditional backup schemes are becoming more expensive and less convenient. In addition, an increase in volume naturally leads to a decrease in reliability. Therefore, new backup systems are being developed that are more suitable for modern equipment and specific business tasks.
Backup systems are distinguished not only by the available schemes and speed, but also conceptually - at the level of their architecture. Classic versions are designed for dedicated storage systems with a single-tier structure. These are separate tape libraries or disk arrays, which are used only for storing backups. According to the selected schedule, they are periodically recorded new data from the productive system in which employees directly work.
Today, single-tier storage systems no longer meet the data availability criteria. As the productive system becomes more complex and its volume increases, both the time it takes to create backups and the recovery from them increases. At some point, it becomes unacceptably high, and the business incurs losses.
')
The main problem is that single-tier storages no longer cope with large disk operations (IOPS) or process data too slowly. For example, a queue of thousands of requests may form to a damaged database in a production. The base must be urgently reanimated, and the time to restore it from a full copy will take several hours.
The solution to the problem of low availability of backups has become three-tier storage systems. They allow you to split backups by their current relevance and provide a reduction in their recovery time at times.
At the first level of such storage systems, the fastest low-capacity drives work, for example, enterprise-class SSDs. With an exorbitant price and relatively small capacity, they have orders of magnitude higher IOPS and record read / write speeds. This is ideal for operating data that you may need to restore right now. Typically, such data in the company is less than 5%, so the volume of the first layer is chosen minimally sufficient.
At the second level, hybrid storage systems are usually placed fast (by the standards of HDD) and relatively capacious SAS-drives as an intermediate solution. This is a trade-off between speed, volume and price. They contain much of the data that may need to be restored in the near term.
The final layer, as a rule, becomes an array of Nearline-SAS disks. They have a large capacity and affordable price, but very low speed. Therefore, they are suitable for long-term storage of those copies of data that are unlikely to need to be restored in the near future. For example, for already completed projects.
Each subsequent layer of a three-tier storage system provides cheaper data storage at the cost of increased recovery time. If you can get them from the first layer in just a few minutes, then unfolding the backup from the third layer can take hours and even days. Therefore, it is important to correctly calculate the hybrid storage configuration and apply additional methods to increase data availability.
You can buy a hybrid storage system as a ready-made system from a well-known vendor, or even create it yourself on the basis of suitable components and open source software. However, over time, the proprietary storage system will encounter difficulties with scaling, while a homemade storage system will always have difficulties with maintenance by other specialists.
There is another conceptually new approach that mitigates the disadvantages of hybrid storage while retaining its key advantages. It comes down to transferring the key functions of managing and maintaining backup systems to the software level. The same three-level
backup storage scheme can be implemented using
Acronis Backup Advanced . Acronis storage node included in the distribution kit can automatically transfer backups between different media. For example, if you deploy Acronis Storage Node on a Windows Server with a SSD RAID array, disks, and a tape library, it will move backups between these devices according to a predetermined schedule.
Acronis Storage also allows you to organize modern three-level storage backups without the need to purchase expensive storage. With it, you can quickly deploy storage using any HDD and SSD. You can increase its volume simply by adding drives of any type and capacity. All configuration, management and status monitoring are performed through a simple web interface. The implementation and maintenance of Acronis Storage does not require an administrator's deep knowledge.
In such a system, backups are automatically moved according to a schedule between storage devices with different speed indicators. Even within the same backup server, they can move between disks with a different interface and rotation speed.
Acronis uses this solution to organize its own cloud backup storage and successfully tested it on volumes of several petabytes in three data centers.
Findings:1. Traditional single-level backup storage systems rarely meet modern data availability criteria.
2. Hybrid storage systems use three-level storage of backups, separating them by priority and ensuring high data recovery speed.
3. Using software solutions, it is possible to reduce the cost of three-level backup storage and simplify their maintenance.
4. Acronis Storage eliminates the need to purchase expensive storage and simplifies its scaling up to petabyte volumes.
Useful links:
»
Webinars on backup and data protection»
Demo version for backup of corporate servers Acronis Backup Advanced»
Comparing Acronis and VeeamDo not be lazy, Friends, make backups on time!