Staying Ahead of Disasters: Improve Business Operations by Modernizing Disaster Recovery

In my last post, I outlined 7 common ways a traditional Disaster Recovery model falls short. Here I’ll review how changing an upstream variable – the type of storage used – can have a dramatic and positive impact on disaster recovery and improve business operations.

Traditional storage platforms have challenges around capacity and durability that constrain your Disaster Recovery capabilities. Object storage was built to address those shortcomings, and it has become widely available thanks to the hyperscalers. Let’s take a look at some of the ways this highly scalable and highly durable storage medium alters Disaster Recovery.

Consolidation

Consolidating data, including metadata, within hyperscaler object storage inherently positions it offsite. Object replication across different regions further increases resiliency. Since silos are eliminated, the loss of a corporate location or even temporary network disruptions no longer impede access to data from other sites.

Continuous Updates

New and changed data are encrypted at the edge, sent to the object store and written immutably as new objects, with all changes tracked as new versions. The golden copy of data now resides in the object store. The result is a fully automated, versioning file system that allows you to revert all or a portion of data to a previous point in time.

Rapid Site Recovery

Since the golden copy resides in object storage, the edge appliances are stateless. In the event of a site loss, a new edge appliance can be deployed and connected to the centralized object store relatively quickly. No longer do you have to fear the sudden catastrophic event or worry about getting replicas completed and offsite in time.

Seamless User Redirection

Users from an affected site can be redirected to a new edge appliance, or an edge appliance located in another facility with access to the same data as the original (failed) one. Load balancers can automate this redirection, routing users to a secondary edge appliance upon primary edge failure.

Durability

Traditional failover and failback paradigms were partially born out of the necessity to address durability concerns with on-prem storage platforms. Object storage was built for durability, and hyperscaler data centers and regions are engineered for robustness, providing a higher level of resilience compared to corporate data centers. Data sent to hyperscaler object storage is safer, more secure, and better protected.

So if changing the underlying storage medium can have a dramatic impact on our approach to Disaster Recovery, how can this ultimately impact the business?

1. Enhanced RPO and RTO

Whether the recovery needs are surgical (single file) or broad (entire volume), it is much faster to manipulate pointers to reference historical objects than it is to move data. Continuous versioning, coupled with all data being online, reduces RPOs from hours to mere minutes, and RTOs from days or weeks to minutes. After an event, these much faster restores minimize lost revenue by accelerating return to normal operations.

2. Cost Efficiency

The new Disaster Recovery model significantly improves cost efficiency by eliminating the need for 3^rd party backup and replication solutions – both in terms of (hard) subscription costs and (soft) admin/IT resource cost. The hub-and-spoke architecture with edge caching reduces inter-site traffic needs to a minimum, which reduces expensive bandwidth requirements critical for replication and off-site backup.

3. Simplified Management

The human brain can only manage so many tasks. Previously, IT was forced to resort to short cuts, such as skipping unfinished replications, because there were just too many systems and solutions to babysit. A modern Disaster Recovery model is far simpler to operate and maintain, allowing IT to provide a higher level of service to the organization.

4. Scalability and Flexibility

Rigid solutions based on fixed RAID groups, pre-allocated LUNs, and customized replication schedules impose constraints that do not match the needs of today’s flexible, dynamic organizations. A hub-and-spoke architecture built around durable object storage is not only far more scalable, but it preserves flexibility for future needs in ways that traditional storage architecture cannot.

Inverting the Ketchup Bottle

In my experience, despite technological advances that enhance virtually everything upstream, Disaster Recovery has gotten more difficult, complex, and touchy over time. Occasionally, when talking with a customer, I compare the new cloud architecture to the advent of the upside-down ketchup bottle – it’s the sort of obvious innovation that makes you wonder why we weren’t doing it that way all along.

The answer is that until relatively recently, object storage hasn’t been as readily available. But today there is no reason to rely on an outdated approach. Traditional Disaster Recovery paradigms are inherently limited because they are built on rigid storage platforms that don’t afford flexibility with growth.

A new way of thinking is needed to keep up with today’s data demands. A modern Disaster Recovery solution takes deliberate and strategic advantage of the benefits of modern technology, fundamentally simplifying and strengthening the overall solution.