3 Keys to Successfully Achieving High Availability Across Your Cloud Computing Environment
by Craig Richardson, Digital Transformation Practice Leader, cStor
If there’s one thing about cloud computing that could be considered the ‘holy grail,’ it’s high-availability. It’s the idea of anytime, anywhere access to services, tools and data, around the clock… literally non-stop.
High-availability is only one side of the coin, however. Reliability is on the other, and it had better have several ‘nines’ on the end (99.999[…]). Combine these mission-critical factors for your cloud environment and you have the foundation of your business continuity (BC) and disaster recovery (DR) plan.
But building and maintaining systems with that kind of availability and reliability comes at a price… often a very high price. In reality, it’s a price that historically only the largest of large multinational firms could afford. And it’s not enough to simply have a failover cluster of servers sitting in a data center. You need multiple, redundant energy sources for the data center, replication between multiple geographical locations, and maybe most importantly, a true ‘multi-cloud’ approach.
However, with the advent of infrastructure-as-a-service and platform-as-a-service providers, the costs of building such a service have decreased dramatically. So on to the good news: it’s now possible for most cloud-based service providers, especially for software-based services, to offer very aggressive service level agreements and pricing.
The other good news is that there are a growing number of cloud providers that truly understand the best way to achieve the perfect balance of high availability and reliability at a digestible cost… it’s just important to know what questions to ask before you dive into signing contracts.
So you are not caught in a litany of high-pressure sales tactics and can more effectively evaluate and compare providers in a fair and balanced way, we’ve offered some key questions to ask, as well as some important takeaways.
Key Questions to Ask When Evaluating a Multi-Cloud Strategy
As you consider how your company can efficiently and cost-effectively architect the optimal levels of availability and reliability across the cloud, be sure you carefully consider the following issues in your evaluation:
- How can I avoid unnecessary expenses and reduce costs? Although your data is growing and will continue to grow, it’s easy to “over buy” and spend more than necessary at any given point.
- How can I be sure my data is secure in someone else’s environment? Yes, they have better security because it’s their business and their subject matter experts can help, but don’t expect all of the security responsibility to be solely on your cloud provider(s).
- Will I be able to easily spin up systems in the event a mission-critical application goes down?
- What happens to my data if my cloud provider(s) go offline? Be sure you understand the true ramifications and full costs to the business if if one or more of your providers goes offline, including the fine print of your SLA and potential losses if business-critical apps go down.
Remember when Amazon’s EC2 service suffered an outage earlier this year? That outage, according to Amazon itself, lasted nearly 11 hours. Several high profile technology firms had serious issues result from the outage, and the 10-day credit Amazon offered was not likely near enough to cover their true cost to the business. It’s best to have a backup and DR plan in place that insulates your business to the highest degree possible.
Lesson learned. We hope.
Easy to say, but what are the actions you can and should take starting tomorrow to be sure you have high availability, reliability and are as insulated as possible from potential outages? I think there are really three keys to getting this right no matter what type of business you run.
First, perform a thorough analysis of your entire environment, from technologies and applications to compliance and workloads and everything in between, so you know what services make the most sense to extend to the cloud and which should stay on premise. Be sure in that process you apply the same policies and procedures for what moves to the cloud versus what you keep on premise. Common sense should still rule the day.
Second, don’t forget IT serves the business, so outcomes, business services and the customers you serve – both internal and external – should be the top priority while you’re in planning mode. Application blueprinting is an ideal way to objectively assess which application candidates are prime to move to the cloud, and which are not, with the business and outcomes staying top of mind.
Third, identify the top 5 mission critical applications and measure availability and reliability with each provider across cloud environments. This should give you a realistic perspective of the business impact of an outage so you and your provider(s) can collaborate on the best recovery plans.
Looking at your environment in this kind of holistic way helps classify applications into four categories:
- Easy to move
- Will move but with some work
- Business critical and should be treated separately
- Legacy that stay on premise
Believe it or not, technology is the easy part of orchestrating your multi-cloud strategy and migrating to the cloud. The hard part is uncovering all of the hidden land mines that could severely impact the business if your strategy isn’t executed properly. Conducting a cost / benefit analysis will also help you more objectively identify if certain applications make sense to move or not, and more precisely evaluate the cloud readiness of your entire environment.
If You Remember Anything…
- Make sure your providers are using best practices concerning redundancy and security. If you’re not sure what that entails, enlist an expert partner that can help you through discussions and negotiations with the providers.
- Don’t forego the holistic assessment of your entire environment and be sure every component from people to process to applications and users are included.
- Remember your multi-cloud strategy isn’t actually about IT, it’s about best serving the business.
Constructing high-availability in the cloud can certainly seem like a daunting process. The most important aspect is to assume that essentially every component of a system will fail at some point, and to prepare as much as possible for that inevitability. With that approach, you can build a strategy that helps you mitigate risk and, as much as possible, automate the processes to handle it when it occurs.