Design of Backup System and Selection of Backup Technology

The choice of deduplication technology:


     The benefits of deduplication during backup are obvious, and it has become the mainstream technology in the industry. The current development direction of this kind of technology, the most common way, is divided by the location where the weight division operation occurs:

In the data source side (host-based) backup, do the deduplication before the backup, and then back up. This method can prevent more duplicate data from being transmitted to the backup server through the network for deduplication, reducing network pressure, but the disadvantage is that it will take up the time and resources of the backup client for deduplication.

The backup server is used for deduplication and online processing (Inline or online processing). Data Domain is a representative of this technology. In this method, data is deduplicated after being read in and before being saved to disk. That is, while backing up, removing weight at the same time. The advantage of In-line is that it saves disk space, and at the same time deduplication can be done in one step, which is particularly simple, but the disadvantage is that it consumes a lot of CPU and consumes a lot of CPU resources, resulting in performance degradation.

     So when we choose which data deduplication technology to use, we should first figure out where our duplicate data occurs most, and then decide whether to perform deduplication at that location. If it is within an enterprise, the sender sends an email with attachments to all employees. The data is stored on the host. In this case, host-based data deduplication can be used.

 

 

Scalability options:

     The architecture of an enterprise-level backup system should be extremely flexible and extensible, so that it is an ideal solution for a distributed enterprise environment. The company's backup strategy must be technically realized, implemented, and centrally managed from a central area to all organizations. The backup system should support the connection between the local area network and the wide area network.



     After the implementation of the system, the network pressure caused by the transfer and circulation of data between different physical sites should be reduced as much as possible. The data on the local or remote host is backed up to the server in the central area. As a centralized backup system, it does not need to deploy any hardware or reside any backup administrators at the branch sites to protect the data of these branches.



     For a large enterprise with many branches, we can back up the data of all branches in a centralized backup system and automatically copy the backup data to the central site and special disaster recovery sites. In addition, we can copy the management data and user data stored on the central backup server to a remote disaster recovery site. All backup, recovery and replication activities are managed in the central data center using friendly graphical interface tools.

 

 

Choice of disaster recovery technology:

 

     The most common disaster recovery technology is remote replication, which is to copy the data on the local backup system to the remote backup system through the WAN. In order to reduce network traffic and save time, we should only copy the backup data processed by deduplication technology to the disaster recovery system.

 

      Understanding all hardware and software related to replication technology is very important for designing a backup and recovery system. As a designer, you need to consider all aspects including:

A set of supported storage arrays, such as DELLEMC Symmetrix or CLARiiON, or other storage arrays with appropriate levels of firmware and API

A set of supported snapshot software solutions, such as MS VSS, DELLEMC TimeFinder or SnapView

A set of supported operating systems and a server version with backup software installed, a client version and any required backup agents, such as: DELLEMC Networker, Samentac NetBackup

A series of authorized backup software application modules, suitable for ordinary computer terminal backup, database backup, and large data file system backup, such as: Networker server/client, Networker SnapImage, Networker Module for Oracle

A set of supported physical volume (media) management tools, for example: DELLEMC Unisphere

A set of supported multi-path solutions for data transfer from backup source to target, for example: DELLEMC PowerPath

A well-designed overall infrastructure, for example: SAN based on shared storage, LAN based on the host's own disk