Background
A substantial number of organizations are relocating their business systems to the cloud to harness its scalability. Nevertheless, ensuring business continuity protection within the cloud is of paramount importance as disasters, viruses, human errors, and the like still pose significant threats to enterprise data security.
The widespread adoption of cloud platforms as the preferred Disaster Recovery (DR) platform by enterprises of all sizes is a significant market driver for DRaaS. With the assistance of DRaaS solutions offered by leading cloud providers, enterprises can carry out DR seamlessly without making substantial infrastructure investments. This accessibility and compatibility with cloud platforms have lowered the entry barriers, making it increasingly easy for enterprises to implement DRaaS solutions. DR service providers should also make full use of the elasticity of cloud platforms to help customers minimize costs to the greatest extent possible while safeguarding data security.

However, cloud computing is not a once and for all solution. We have recently witnessed the failure of multiple availability zones in Azure. It is evident that Single-cloud DR is not the most comprehensive DR solution, and we need to consider it more thoroughly.

What is the best method for DR in the cloud? Single-cloud DR, cross-region DR within the cloud, or cross-cloud DR? Today, we will discuss the pros and cons of these solutions, and how HyperBDR can help users simplify the deployment of DR and quickly establish DR on the cloud and across clouds.

Ⅰ. Typical In-Cloud DR Methods

First, let's look at the classic cloud business recovery method. Let's take AWS as an example. It pioneered the modern cloud DR model.

The DR approach in AWS is centered around the ability to restore applications promptly and efficiently. If an AWS customer encounters a situation demanding the recovery of applications and data, AWS can initiate a recovery instance within minutes. This guarantees minimal downtime and service interruption. Additionally, these instances can be restored based on the most recent backup point-in-time data, or if the circumstances require, can also be recovered based on a previous point-in-time selected by the user.

A. AWS in-cloud DR based on storage layer
Recovery Objective (RTO) and Recovery Point Objective (RPO) are key indicators that define the acceptable recovery process time and data loss time. In AWS local DR solutions, these indicators hinge on where the data snapshots are stored. The calculation of RTO and RPO varies depending on the storage tier. For instance:
•    Standard EBS Block Storage:
Designed for workloads that require exceptional efficiency and low latency. This storage tier boasts faster RTO and RPO times and is ideal for mission-critical applications.
•    Amazon S3:
Highly scalable, durable, and secure object storage that strikes a balance between performance and cost. RTO and RPO metrics may differ slightly compared to EBS block storage.
•    Glacier:
It is an archival storage solution intended for long-term data archiving with retrieval times ranging from minutes to hours. Therefore, RTO and RPO are typically longer for data in Glacier.

B. AWS Cross-Region DR
In terms of the DR strategy, AWS also recognizes that geographic redundancy is essential for true DR. Hence, they also offer cross-region DR for enterprises. This implies that if the primary region faces a disaster or service disruption, it can fail over to another completely independent region. This ensures the continuous availability of data and the rapid recovery of application runtime.

Ⅱ. Why In-Cloud DR Has Limitations

Although many enterprises rely on cross-region DR within a single cloud provider as their preferred strategy, this approach has its inherent limitations. When problems occur simultaneously in both the region where the business system is located and the DR region, DR fails. In the Azure outage incident in July, we witnessed the simultaneous failure of three availability zones. Why relying solely on cross-region DR within a single cloud provider might not be the most effective approach?

A. Risk of Service Interruption
Although the data centers in different regions are physically separate, they still rely on the infrastructure and management of the same cloud service provider. In the event of a major failure or policy change in the cloud service provider, it could potentially affect all regions.

B. High Cost
Cross-region DR depends on the technology and services of a single supplier, which limits flexibility and choice. If there are issues with the supplier's technology or services, the user's DR plan could be impacted.

C. Lock-in to a Single Supplier
Cross-region DR relies on the technology and services of a single supplier, which limits flexibility and choice. If there are problems with the supplier's technology or services, the user's DR plan could be compromised.

D. Latency Issues
The data transmission latency between different regions might be significant, especially when the regions are far apart. This could influence the efficiency of data synchronization and the recovery time objective(RTO).

Ⅲ. Advantages of Cross-cloud DR

Consequently, similar to the traditional DR concept, we implement cross-cloud DR to a certain extent to mitigate the risk of reliance on a single cloud provider. Of course, cloud vendors are more focused on optimizing their own architecture and services and do not concentrate on researching other heterogeneous platforms. Therefore, we can observe that cross-cloud DR solutions are typically provided by neutral DR tool vendors. By leveraging the advantages offered by multiple cloud services, enterprises can utilize the cross-cloud DR solutions of third parties to enhance business continuity protection plans. Let's initially explore the advantages of cross-cloud DR.

01｜Diversity and Resilience
Cross-cloud DR enables the distribution of data and applications among multiple cloud service providers, reducing dependence on a single supplier and enhancing the system's resilience and fault tolerance.

02｜Cost Optimization
You can choose the most cost-effective solution based on the prices and service features of different cloud service providers to achieve cost optimization. For example, one cloud provider may offer cheaper storage, while another may have more advantages in computing resources.

03｜High Availability
Cross-cloud DR improves the availability of data and applications as even if one cloud service provider fails, another can still take over the business and ensure business continuity.

04｜Reduce the Risk of Service Interruption
By spreading across multiple cloud service providers, cross-cloud DR effectively reduces the risk of service interruption caused by the failure of a single supplier.

05｜Performance Optimization
You can make use of the geographic distribution of different cloud service providers to optimize application performance. For instance, you can choose a cloud provider that is geographically closer to the user to reduce latency and enhance response speed.

06｜Data Sovereignty and Compliance
Cross-cloud DR can more flexibly meet the data sovereignty and compliance requirements of different regions. For example, data can be stored on cloud providers in specific countries or regions in accordance with laws and regulations.

Ⅳ. HyperBDR's Cross-Cloud Data Protection Solution

Of course, most traditional vendors can offer cross-cloud DR solutions, but they still adhere to the traditional 1:1 host standby model, which consumes a considerable amount of resources and makes cross-cloud DR solutions prohibitively expensive, contrary to the original intention of enterprises to use the cloud to reduce costs and enhance flexibility. Regarding the utilization of cloud native capabilities, traditional DR vendors need to broaden their thinking.

HyperBDR is a cloud-native tool that helps users achieve cross-platform recovery of applications. It fully utilizes the cloud-native capabilities provided by the cloud platform, focuses on resolving the seamless flow and rapid recovery of data, automates the driver adaption and resource orchestration of heterogeneous platforms, avoids human intervention, and reduces dependence on computing resources. Through extensive project experience, we have gradually optimized an automated DR solution suitable for most cloud platforms.

•    Storage-Based Recovery
Backup data is stored into the cloud-side object storage or block storage, and HyperBDR utilize the DR cloud platform capabilities to restore it only when a disaster occurs. The backup process does not require computing resources, significantly reducing the DR cost.
•    Minute-Level RPO and RTO
The block storage mode offers higher recovery efficiency, while the object storage cost is lower. Customers can strike a balance based on their own needs. Of course, the speed of recovery from cloud storage depends on the cloud vendor. For example, in Huawei Cloud, the recovery time is within a few minutes. If you need more DR practice data of cloud platforms, you can contact us at enquiry@oneprocloud.com to obtain it.
•    Automatic Resource Orchestration
Employ the API interface provided by the cloud to automatically orchestrate resources. After completing the DR configuration, when disaster takeover and drills are required, the stored data can be restored to the application with a single click. This storage and recovery process fully utilizes the cloud native capabilities and can be completed with a single click on HyperBDR without switching to the cloud platform.
•    Automatic Driver Adaptation
Drivers are a significant issue for business recovery on heterogeneous platforms. With the driver library accumulated through numerous projects, driver adaptation is fully automated during the DR process.
•    Agentless DR
Thanks to the interface provided by AWS, we have reduced a large number of cumbersome actions in obtaining source data. There is no need to install agents on each host, which significantly reduces the workload and minimizes the impact on the source-side business.
By resolving the cross-cloud compatibility issue, the deployment and implementation of DR become simpler. After completing the DR strategy configuration, the only action an enterprise needs to take when recovering its business is to initiate this process. Deploying cross-cloud DR is not only the best solution for business continuity protection in the cloud computing era but also a means for enterprises to further optimize DR costs and gain more flexibility.

Ⅴ. HyperBDR's Compatibility

In the practice of cross-cloud DR, we encounter numerous challenges. The difference between geography  and architectures complicate cross-cloud scenarios. Due to the vast differences between the Asia-Pacific and global markets, we have to undertake extensive adaptations during projects and develop DR scenarios between most cloud platforms. This has became a product advantage and successfully constructed a DR bridge between these cloud vendors. Currently, HyperBDR can support DR between major clouds like AWS, Huawei Cloud, Azure, Alibaba, as well as some international private cloud vendors, covering approximately 40 cloud platform versions.
Deploying HyperBDR is simple: all you need is a host with 8 cores and 16GB of memory, which can, of course, be deployed on your disaster recovery cloud platform. Feel free to contact us @enquiry@oneprocloud.com to obtain HyperBDR's cross-cloud DR solution.