Downtime is the Enemy of Presence
As businesses change and adapt, one thing is for certain: their online presence (website, ecommerce, service portal, public web services, etc.) continues to grow… and is so does the importance of that presence. Expanded presence offers tremendous opportunity, but also brand, business, and operational risk. Outages become much more noticeable and are no longer a “back-of-the-house” issue that can be hidden/minimized from customer view. Thus, downtime has an immediate and negative impact on the business: reputation suffers, automated income streams suffer, and ecosystems and collaboration quickly grind to a halt.
- Take lettuce as an example. The tasty leafy green vegetable we buy from our grocery store now has its own website. This site allows you to see the organic process they use and why you should buy that product. With a smartphone you can even make online purchases. If that web site is down, and you’re in a hurry for ruffage, you might make the run to your local store. Who would have imagined such a thing just a few years ago?
- Want or need to get a Covid test or vaccination – to be part of a public health solution? Schedule an appointment online! We all understand how important that will be in the coming months.
When the services that manage customer facing systems are unreachable the impact of downtime is immediate. No test/vaccine or lettuce for you, only frustration, brand damage, and customer churn as they look for alternatives.
To counter the impact for downtime (and the follow-on effects), data protection products, especially those hosted/managed by third parties, have enhanced their Service Level Agreements’ (SLAs) downtime tolerances, modernizing those tolerances away from hours – to minutes. One such metric often found in one of these SLAs is the Recovery Time Objective (RTO) – i.e., how quickly you can recover from failure to full operations. The bravest providers will only adhere to a stringent RTO when they have fully-orchestrated recovery and ready resources– when as part of a Disaster Recovery as a Service (DRaaS) offering. There are tradeoffs to technology selection, however. Let’s examine what it takes to counter downtime and how two different data protection technologies – Bare Metal Recovery (BMR) and DRaaS – compare.
Introductions to BMR and DRaaS
Bare Metal Recovery (BMR): Bare Metal Recovery was developed to help protect physical servers (and has been extended to protect virtual servers). When doing a BMR-style backup, you need to have an agent installed on the host server (physical or virtual) to enable the backups. The agent allows you to backup an entire server by capturing a disk image, literally every bit on the disk(s). This, in turn, enables you to restore a complete system, including the operating system and its settings, applications (including their configurations and updates), files, folders, and volumes. This saves you time because you don’t have to re-install everything from scratch. With a BMR backup, you have to completely restore the server to bring it back online. You also have to provide the site (hardware or virtualized hardware) to recover from in case of a disaster.
Disaster Recovery as a Service (DRaaS): DRaaS, as its name implies, is Disaster Recovery (i.e., capturing system backups to secondary location so that they may be booted in the event of emergency) – and a Service, meaning – the software to capture, recover, and boot plus the service infrastructure for which to recover/boot on – is made available as one complete offer – Disaster Recover… as a Service. DRaaS is much faster than bare metal recovery, with the ability to start and run systems in the cloud (secondary location) in the event of a man-made or natural catastrophe. DRaaS lets you “instantly” boot a protected server or your entire site and immediately restore your presence and business operations.
With our definitions out of the way it is time to look deeper into the use cases for both solutions.
Comparison of BMR and DRaaS
BMR is ideally suited for tier 3 and 4 systems (by definition, less critical) such as workstations, desktops, laptops, development servers, and print servers. These systems are important, but not critical, to the business and can afford to be down for extended periods of time. That is to say, these are not services that will directly impact brand, commerce, and critical business operations. That said, BMR is extraordinarily flexible, when it comes time to recovery. Administrators can recover full systems individual items such as files and folders. Regardless, what BMR doesn’t have – is a destination. An administrator must supply and define a recovery destination for the items or servers to restore back to – as to get the value of the recovery.
Disaster Recovery as a Service includes builds on these features and includes all the functionality of BMR to provide RTO in minutes. With DRaaS you skip the restore part (1’s and 0’s only move so quick) and go directly to booting them to quickly get services back online. The limiting factor is no longer the hydration of data, but the boot time of the servers within the environment. In addition to greatly reducing downtime and recovery times the complexity of recovery is also greatly decreased thru predefined orchestration…. oh…and you do not need a secondary site to manage or maintain…that is included in the service.
Why are these distinctions important?
Recovery is what matters… not backups
Speed / RTO: BMR is slower and more complex. To recovery you must:
- Set up recovery destination hardware (or virtualization)
- Restore/rehydrate the servers from the backup
- Boot the system into a “recovery mode” and configure networking/connectivity and maybe drivers
- Repeat for each system
Whereas for DRaaS – with orchestration – you click a button and all the systems come to life!
Testing: The only way to truly ensure recovery is available when you need it is to regularly test the environment to find out how it would behave in a failover scenario.
With BMR you need to manually complete all of the tasks above, a huge time and resource investment. It is no wonder that many organizations don’t test often enough; the resources required are too extensive. The opposite is true with DRaaS. At the core of DRaaS is automation, and automation is a huge time and resource saver.
With DRaaS the systems are already prepped and ready. They can be booted offline in an isolated network sandbox, be automatically powered-on and verified, and automation can ensure integrity by delivering verification of a booted/operating system (e.g., in the form of a screenshot or pingable services).
Price: The cheapest solution often end up being the most expensive solution; just ask anyone who got a “great deal” on a used boat. You get what you pay for.
Even though the cost of BMR software is very attractive, the costs of recovery quickly adds up in both direct and indirect costs. In testing management and recovery scenarios, there are a huge number of additional costs from preparation, labor, and reporting perspectives that don’t factor into the licensing cost.
Conversely, with the additional “insurance” paid for with DRaaS – you get the destination compute capabilities, the orchestration, and the automation – making both testing and actual recovery guarantees for IT to deliver on its promise to the business. You can quickly build a case that not only is DRaaS price competitive, it’s necessary and perhaps even cheaper to recover.
Infrascale is here to help! If you want to learn more on how Infrascale Disaster Recovery solutions can keep your data safe and your business operational, visit the Infrascale website or sign up for a product demo. MSPs welcome!