[Disaster Recovery] Data Copy Management Technology (CDM)

[Disaster Recovery] Data Copy Management Technology (CDM)

Operation and Maintenance YouthO&M Youth

CDM data protection issues

---

In the current IT environment, a data protection system is an indispensable system. Data protection generally uses backup software to realize automatic data backup, data preservation cycle management, and data recovery when needed. The main purpose of data protection is to be able to recover when there is a problem with the production data. However, nowadays, the amount of data is increasing and the business is becoming more and more Internet-based. When a system goes wrong, it will take a long time to restore from the backup, which will cause a long-term business interruption. Therefore, but for a long time, backup software is just "backup software", it is rarely called "backup recovery software". In the data protection environment that many users have established, whether data can be restored is only at the moment when data is lost. I really understand that this is actually a helpless move for many users.
The storage medium used by the backup system has gradually changed from tape to disk, which has improved data recovery to a certain extent. However, as the amount of data in the business system continues to increase, the backup time is still longer and longer, and the recovery from the backup is still too slow. The role of the backup software in the daily recovery work of the user is getting smaller and smaller. The trouble is getting bigger and bigger. Therefore, many backup software vendors call the backup system "the last line of defense." If there is no revolutionary breakthrough in the technology of backup and recovery, in the future, this kind of software can only be called archive software, and its role in IT systems will become smaller and smaller.

CDM data copy problem

---

Entering the Internet era, the explosive growth of enterprise applications, multiple different applications (query, test, audit, emergency, etc.) will use the same data (customer information, transaction information, etc.). According to IDC statistics, duplicate data consumes 60% of the enterprise Storage capacity, and 65% of storage software purchase costs. In addition, enterprise application environments are increasingly distributed in hybrid environments (data centers, private clouds, public clouds, branches). How to achieve agile flow of data between multiple applications and hybrid environments (data management and data Service), how to reduce the cost of data storage and flow has become one of the main issues of CIO's concern.

Benefits of CDM

---

CDM can obtain the following benefits:

1. Get data faster. Obtain the changing data copy in a permanent incremental and full synthesis manner, and the impact of the copy process on the data source is minimized and consistency is ensured.
2. It is more convenient to use data. CDM can quickly generate multiple copies of data and provide them to multiple different applications in different logical versions, and mount them directly without the need to restore data copies.
3. Manage data more efficiently. Before using CDM, companies need to prepare data for each application to do point-to-point replication and copy; after using CDM, companies obtain data through a CDM platform, virtualize the data into multiple copies and provide data services to multiple applications at the same time. The complex and scattered processes of acquiring data, preparing storage, mounting data, and protecting data are unified and automated on the CDM platform, and the IT infrastructure is changed to DaaS-data as a service.
4. Store data more streamlined. CDM's data virtualization service provides services by virtualizing multiple copies of logical data from one physical data, thus greatly simplifying storage space. Save a lot of hardware investment for "secondary data storage" for enterprises.
5. Data flow is smarter. CDM's data service integrates technologies such as data desensitization, data format conversion (ETL), and cloud interaction (Cloud-out) to support application scenarios such as development and testing, big data analysis, and data circulation in a hybrid cloud environment.
   

CDM data protection scenario

---

Traditional data protection technology realizes OS, database, and backup. Because the tape library has a high price-to-capacity ratio in the past, it is one of the important technologies to provide tape library management, such as tape labeling and automatic tape replacement. As the price-to-capacity ratio of disks becomes higher and higher, disk media is also supported as a backup target, but the backup data format is still tape format, such as tar, cpio, etc.
CDM can put aside the traditional tape concept, make full use of the disk data format, and add more functions, thus greatly expanding the concept of backup.
Taking Oracle backup as an example, the general CDM steps are as follows:

1. Mount the backup and store it as a temporary disk to the database server
2. Generate data copy through database API (RMAN for Oralce)
3. Copy and synthesize incremental data blocks
4. Uninstall the temporary disk of the database server
5. The CDM software generates a snapshot internally and retains the full data copy at that point in time. It is worth mentioning that when the CDM backs up the database, it is done through the database authentication interface, which fully guarantees the consistency of the backup data, which is very similar to the traditional CDP backup technology. Big difference.

CDM has another major advantage in data protection, namely the synthetic full backup function. The smaller the amount of data to be backed up, the faster the backup speed, and the smaller the bandwidth usage and business impact. So you should do as many incremental backups as possible and reduce full backups.
However, the paradox is that during recovery, the recovery speed of a full backup is much greater than that of an incremental backup, because when traditional backup methods are used, full backups must be performed regularly to ensure the recovery speed. The function of synthesizing incremental backup data into full backup data solves this contradiction. After the first backup, it can be incrementally backed up forever in the future, and then a new full amount of data will be synthesized to save, and the backup window will only follow the daily increment. The data is related, not the full amount of data. Therefore, regardless of the amount of data stored from the backup window or the backup, CDM has great advantages.

CDM data recovery scenario

---

If you restore from a backup medium to a business disk, the larger the amount of data, the longer the restoration time. In the traditional backup solution technology, the backed up data needs to be restored before it can be used. The restoration time depends on factors such as data volume, bandwidth, etc., and space for restoration needs to be prepared.
Nowadays, the volume of business data is increasing. Users often encounter database recovery times of several terabytes in days or even more than a week, which greatly affects the RTO of business data and the feasibility of recovery drills.
Most CDMs support the function of providing backup data directly to business machines. In case of emergency, there is no need to prepare space and restore data. Mount the backup media directly to the business server. Recovery can be achieved within a few minutes, whether it is 3TB, 10TB, It's still a 20TB database. Because it no longer needs to import data from the backup back to the system to be restored, the restoration time has nothing to do with the amount of data.
Therefore, CDM has great advantages in terms of backup efficiency and recovery efficiency in environments such as databases with large data volumes and virtualization, and it is a revolutionary technological change for the backup system.