As organizations more often use cloud storage, they’re taking advantage of additional cloud services that can help them reduce costs and improve operational efficiencies. One cloud service type that’s gaining popularity is cloud disaster recovery. The term “cloud disaster recovery” is often used interchangeably with disaster recovery as a service or DRaaS. This article refers to the former term.
Cloud disaster recovery is a backup and restore strategy that applies not only to data, but also entire virtual machines, servers and corporate networks. The operative word is “strategy” because businesses need to decide for themselves how best to use such a service.
Rather than simply subscribing to a service option and hoping for the best, it’s important to understand what the company’s priorities are in advance, so an appropriate disaster recovery plan can be put in place. Otherwise, the business may lack timely access to critical resources when disaster strikes. The results of such as mistake could be extremely costly or even fatal, so it’s important to understand the trade offs among the available cloud disaster recovery options.
Cloud disaster recovery is an attractive option for any size organization, regardless of where they fall along the cloud maturity spectrum. This is because cloud disaster recovery provides speed and cost advantages over traditional disaster recovery approaches. As enterprise infrastructures continue to become more virtual, an increasing amount of data and IT operations are moving into the cloud. Therefore, their disaster recovery strategies need to evolve.
Pros and Cons of Different Cloud DR Approaches
Cloud disaster recovery is more flexible than traditional forms of disaster recovery because subscribers have more DR recovery solutions. Instead of backing data up from a data center to tape, for example, cloud subscribers have many more options including:
- Backing up from a data center.
- Backing up from a private cloud.
- Backing up from the same cloud service used to store data.
- Backing up from a hybrid cloud environment.
- Restoring to the original environment.
- Restoring to the cloud versus on-premise.
Given the greater flexibility and the relatively low cost of cloud disaster recovery, it’s an attractive option, although businesses are wise to consider their priorities first. In fact, part of any good cloud disaster recovery plan prioritizes the recovery of assets. If disaster strikes, what’s critical and what isn’t?
For example, when it comes to data, there’s the concept of hot, warm, and cool storage. “Hot” data is that which needs to be readily available. Of the three types of data, hot data is the data that is accessed most often. “Warm” data is data that’s accessed less often than hot data, such as historical data used for reporting purposes. “Cold” data is data that is rarely used but must be retained. The stratified approach to data informs the associated service level agreements (SLAs) and also the associated costs of storing the data.
A similar temperature-type concept applies to the disaster recovery of entire sites:
- A hot site is a complete copy of the production site. Its purpose is to minimize downtime in the event of a natural or man made disaster.
- A warm site has established connections between a primary site and a secondary remote backup site. Recovery is delayed, but not as much as a cold site.
- A cold site is one that is essentially unprepared for disaster recovery so when disaster strikes, it takes considerable time to get the site back online. Not surprisingly, a cold site is the cheapest option, although when disaster strikes it may prove to be an expensive option from a total cost perspective.
Cloud disaster recovery uses a comparable model that ranges from cold to hot. Specifically, customers can choose backup (the slowest and cheapest option), a minimal version of an environment, a partial version of the environment or full (multi-site) disaster recovery which is a SAN-to-SAN enterprise back up replication method that runs in the cloud and on-site. The benefit of multi-site is that traffic is rerouted to the cloud during recovery.
Cloud Disaster Recovery Planning
Cloud disaster recovery planning begins with the prioritization of many things, including applications, data, and services. Each asset should have an acceptable recovery target associated with it. Understanding the top tips for disaster recovery planning as a service is essential.
There are two reasons why prioritization is necessary: cost and operational prudence. Treating all assets equally is neither cost-effective nor practical because companies would over pay or under pay for services. What is the business impact if a particular application, other IT asset, or certain type of data becomes unavailable? Some assets are more critical than others, so a cloud disaster recovery plan should reflect those priorities.
Two important cloud disaster recovery metrics are recovery time objective (RTO) which is the time it takes to restore a business process to its target level and recovery point objective (RPO) which defines the acceptable level of data loss. The target metrics need to be defined as part of the plan.
It’s also important to understand the scope of threats to business continuity. That is, the kinds of natural or man made disasters might cause business disruption. For example, when Hurricane Harvey struck Houston, many businesses in the area lacked a disaster recovery plan.
Other prudent elements of a plan include who needs to be notified or involved in the event of a disaster and what the budget is for cloud disaster recovery. IDC estimates that the total cost of unplanned application downtime per hour is about $100,000 per hour for a non-critical application for a Fortune 1000 company. A critical application failure can cost between $500,000 and $1 million per hour. Knowing the trade offs between costs and recovery time is essential for enterprise backup and recovery management.
Benefits of Cloud Disaster Recovery
- Flexibility: Organizations are not tied to any particular type of architecture, so regardless of where they are on their cloud journey, they can choose an option that meets their needs.
- Cost: A cloud disaster recovery service is cheaper than physically duplicating an environment. This fact has enabled smaller organizations to take advantage of disaster recovery options that they couldn’t afford otherwise.
- Faster recovery times: Backing up from the cloud is faster than backing up from tape. Hosting both sites or storing data on the same cloud as the cloud disaster recovery service can have additional time advantages. The virtualized nature of the cloud also offers advantages over a physical twin. For example, if a virtualized server fails, another can be spun up in minutes. Since virtualized servers are technology-independent, different applications, operating systems, data, and patches can be store on them, so all of that is automatically restored with the virtual server. In a non-virtualized environment, each element has to be restored individually.
- Elasticity: This general benefit of cloud computing applies to cloud disaster recovery. As data grows and environments become more complex, scalability is not an issue.
- Compliance: Faster recovery times may help avoid fines for missing deadlines.
Risks of Cloud Disaster Recovery
- Security: The biggest concern is the multi-tenant nature of cloud environments on which backup environments or data are hosted.
- Recovery: Can take longer than desired without comprehensive planning. Asset prioritization, SLAs, and bandwidth of the connections between the original environment and the backup environment should be considered.
- Outage: Cloud environments aren’t perfect. It’s wise to consider this risk as part of an overall disaster recovery strategy.
Cloud Disaster Recovery Services
Cloud disaster recovery services are available from cloud providers and managed service providers (MSPs). Both offer cloud services, but the difference between them is that MSPs provide a higher level of IT support. In both cases, it’s important to ensure that the service level agreements (SLAs) are driven by the company’s priorities, RTOs and RPOs.
Managed Service Providers
Cloud Disaster Recovery Options
Whatever is defined as the core is restored first.
Whatever is defined as the highest priority partial environment (larger than the core is restored first.
The original and backup stay in sync.