Gorodenkoff - stock.adobe.com

DRaaS decisions: Key choices in disaster recovery as a service

We examine the key decisions when considering DRaaS. Whether to go full self-service, assisted or managed will depend on what you need to protect and your in-house resources

This article can also be found in the Premium Editorial Download: Computer Weekly: IT must play its part in tackling the climate crisis

The way businesses buy disaster recovery is changing – and rapidly. Over the last few years, the growth of cloud-based data storage – led by Amazon’s S3 but closely followed by Microsoft Azure and Google Cloud Storage – has created a vast resource that businesses can use for backup.

But it is also changing the economics of disaster recovery services, and to the benefit of customers.

IDC put the size of the disaster-recovery-as-a-service (DRaaS) market at $3bn in 2018, but expects it to grow to $5.4bn by 2022. Much of this is being driven by cloud-based services.

Businesses are moving towards DRaaS for its performance – gaining greater uptime and shorter recovery times – and for cost reasons.

Replacing dedicated backup datacentres with cloud options taps into the economies of scale that cloud can bring. It removes the need for customers to hold expensive IT infrastructure and real estate, and allows IT departments to move to a consumption-based model for disaster recovery.

But, analysts warn, the market is fragmented. IDC estimates that more than 2,000 cloud service companies offer disaster recovery, in addition to IT and software providers that operate in the conventional disaster recovery market. Specialist DRaaS suppliers include Arcsite, Acronis, Sungard and Zerto, while Amazon offers AWS CloudEndure and Microsoft has Azure Site Recovery.

DR options

Cutting through all the offerings in such a large and diverse market is a challenge. But we can break down services into three main areas – managed, assisted and self-service DR.

Looking at the type of DR service a business wants, and the data and applications it needs to protect, will help to narrow down the options. IT directors should note, however, that DRaaS is primarily a form of backup; it is not a live, fully replicated operating environment.

Self-service DRaaS

Self-service DRaaS involves the customer planning, buying, configuring, maintaining and testing disaster recovery services. And, although options for automation are improving, the IT team will typically need to be available to invoke the DR plan and run the recovery process.

The benefits are flexibility and often cost. The business can choose exactly which mix of recovery services, backup and recovery software, and even the raw storage, it needs. A self-service model can lend itself to mixed environments, with multiple cloud data stores and application-based availability and DR tools.

This allows IT teams to best match the business’s requirements to technology and budget. The heavy lifting comes in assembling a bespoke solution and ensuring there are enough skilled staff to manage it. In that way, self-service DRaaS suits organisations with larger IT teams that already manage DR strategy and run DR assets.

Assisted DRaaS

Opting for assisted DRaaS removes some of the management burden of DR planning and operations. DRaaS companies include consultants who help design systems, and engineers who build – and sometimes operate – the DR capability.

Assisted DRaaS should be cheaper than a fully managed service, as it leverages existing in-house assets and skills. It allows the business to bring in additional support when needed, but still keep in-house the customised or bespoke systems that a service provider might struggle to support.

Equally, if part of the organisation has a robust DR system – conventional or service-based – assisted DRaaS can avoid duplication. The business simply pays for additional capabilities, such as support for a new application, or where the IT team feels that existing arrangements are insufficient.

With assisted DRaaS, the organisation removes some of the administrative burden, and can operate with a smaller in-house DR team, but still keeps overall control.

Managed DRaaS

Managed DRaaS is the most comprehensive, but also the most expensive, option.

The main benefit is that in-house IT teams can hand off DR operations entirely to the third party. This reduces the burden on skilled staff. And, although a managed service is typically more expensive than other DR options, it can be money well spent for a comprehensive service and peace of mind.

The cornerstone of any managed DRaaS option is the service-level agreement (SLA). This will determine the recovery point objective (RPO) and recovery time objective (RTO), and drive the cost. The provider is responsible for deployment, testing and maintenance.

The provider also chooses the platform. A further benefit of managed DRaaS is that systems are largely technology agnostic. The provider decides which cloud services to use for raw storage, or even whether to use the cloud, its own datacentres, or the customer’s own on-premise resources.

IT teams cannot hand over all responsibility, however. IT still needs to understand the business’s operational needs, and communicate that to the managed DRaaS provider. If IT systems – or business requirements – change, the provider needs to know.

Pick and mix – which DR service and when?

Choosing the right form of disaster recovery service depends less on the business’s applications than on its internal capabilities, and its attitude to risk.

A risk-averse organisation that requires high levels of uptime is most likely to opt for a fully managed service. Although costs can be high, it is part of the price of doing business. One trend, identified by backup and business continuity supplier Arcserve, is for lines of business to buy DR services directly. These managers are more likely to opt for a managed service.

Organisations with greater in-house capabilities might choose to trade ease of management for flexibility and lower costs. Cloud-based managed DR can provide very high levels of resilience, but the business might not require those levels across all applications. As a result, an assisted or even self-service approach could be more viable.

Read more on disaster recovery

An in-house approach to DR might also be the most appropriate option for organisations with large in-house datacentres and supporting infrastructure, or where there are legal, security or privacy concerns about handing over data to a third party.

Organisations with a mix of platforms, on the other hand, might opt for an assisted approach. One question for CIOs is how to ensure the business operates if a cloud-based software-as-a-service (SaaS) application were to go offline. The SaaS provider’s own in-house SLAs might not be sufficient. An assisted approach allows the IT team to bolster its own resources with experts familiar with the platform.

But, according to IDC’s Phil Goodwin, IT teams should also be looking forward. In the future, high availability and application availability services are likely to take over from infrastructure-centric DR.

Analysts believe the trend is towards near-instant recovery and “availability as a service”. So, any solution needs to be both robust enough to protect the business today, and flexible enough to exploit emerging technologies.

Read more on Disaster recovery

Data Center
Data Management