01. One-to-one stand-alone hot backup - traditional disaster recovery method
Stand-alone hot backup is a backup solution that uses two servers to ensure high availability and is the most common disaster recovery mode on the market. In stand-alone hot standby, a primary server and a standby server are kept in sync to ensure immediate switching to the standby server when the primary server fails or goes down. This ensures system continuity and reliability, and reduces the risk of business interruption and data loss.
Specifically, the timing of immediate takeover depends on the readiness and degree of automation of the backup system. In some highly automated backup systems, immediate takeover may take only a few seconds, while in some less automated backup systems, it may take minutes to hours. Overall, the goal of immediate takeover is to convert the backup system to the previously failed system as quickly as possible to ensure business continuity and user experience.
02. Many-to-one/hostless hot backup - HyperBDR exclusive technology
Many-to-one host hot backup and non-host hot backup are unique cloud-native automatic disaster recovery technologies of Wanbo Zhiyun. Compared with single-machine hot backup, it no longer needs to configure a backup host on the target platform 1:1 during the data synchronization stage For data synchronization, it only needs to use the interface capabilities of the cloud platform and the characteristics of native resources to transmit the source data to the target cloud side through a point-to-point encrypted network. While meeting the requirements of business continuity, it greatly reduces the consumption of computing resources, making disaster recovery possible.
Specifically, HyperBDR can choose to use block storage or object storage as the backup medium for different business needs to meet different levels of disaster recovery requirements. Of course, this will also make the cost of cloud storage different.
1. Many-to-one host hot backup - cloud hard disk mode
Wanbo Zhiyun HyperBDR disaster recovery software captures block-level full and incremental data by calling the native API interface provided by the cloud platform itself, and by scheduling source-side Agent and Agentless programs, and transmits it to the target cloud side through a point-to-point encrypted network. Storage gateway", no data is retained locally, safe and efficient, and then the cloud storage gateway writes the data to the target cloud disk for backup.
When the data in the production system is damaged, the user selects the historical time to be restored, finds the corresponding copy data in the cloud disk, generates a virtual clone volume, mounts and restores it to the disaster recovery host in minutes, and quickly restores the business.
2. Many-to-one host hot backup-object storage mode
Wanbo Zhiyun HyperBDR disaster recovery product also supports the backup of block-level full and differential data slices to object storage. In this scenario, HyperBDR no longer needs to use any computing resources of the disaster recovery platform. By calling the object storage standard S3 API interface, the full amount and incremental data are stored in the specified bucket (Bucket) of the object storage. At the same time, multiple copies with time point attributes can be generated according to the set backup strategy. Users can Use these snapshot time points to restore the business system as a whole.
When the production system fails, the user selects the time point to be restored, uses the orchestration capability of the cloud platform API to restore the data from the object storage to the block storage, and then mounts and restores it to the disaster recovery host to quickly restore the business.
3. Differences between many-to-one host hot backup and no-host hot backup modes
1) RTO difference:
During the backup and recovery process, data compression and storage processing will also affect the RTO. Since the object storage mode needs to convert the data from block storage to object storage for storage, it is restored to block storage only when the business recovery is started. This restoration process means that It is bound to increase the data processing time, thereby prolonging the recovery time.
2) Cost difference:
| block storage mode |
backup cost |
Subscription fee for cloud storage gateway (ECS host) |
Subscription fees for cloud storage |
| Recovery and drill costs |
On-demand subscription ECS host fees for recovery and rehearsal |
||
| object storage model |
backup cost |
Subscription Fees for Object Storage |
|
| Recovery and drill costs |
On-demand subscription ECS host fees for recovery and rehearsal |
||
It can be seen that the main difference between the two modes lies in the backup cost. Due to the low cost of object storage, the backup (storage) cost of mode 2 is about 90% lower than that of mode 1.
03. Market opportunities for non-host hot backup
Compared with traditional stand-alone equipment, the two disaster recovery modes of Wanbo Zhiyun have greatly reduced the cost of disaster recovery data storage, while the RTO time can still maintain a high level.
| disaster recovery mode |
Cloud disaster recovery data storage cost |
Business recovery RTO time |
| Stand-alone hot standby |
About 150% of the production environment |
Second level ~ minute level |
| Many-to-one host hot standby |
About 30% of the production environment |
Minute level (related to the cloud platform, the minimum is 5 minutes) |
| No host hot standby |
About 3% of the production environment |
Minute level (related to the amount of data, 500G valid data is about 30 minutes) |
The high overall cost of disaster recovery has always been the main pain point of customers. With the cost reduced on a large scale, enterprises can have more business systems to achieve a higher level of disaster recovery to ensure business continuity and meet relevant production compliance requirements . As the proportion of enterprises using the cloud increases, the use of hostless hot backup for disaster recovery on the cloud will become a new opportunity for business growth in the future.
And because of the strong adaptation between HyperBDR and the cloud, the object storage data copy of the disaster recovery in the cloud can be pulled up on the cloud with one click, which also brings more usage scenarios for the disaster recovery data, such as drills on the cloud, Test verification, big data analysis, etc.
