How to handle cloud outages: multi-cloud and hybrid cloud strategies

Nov 29, 2017 11:42:00 AM / by Matt Ledger

  Matt Ledger

There’s one big problem with moving your business to the Cloud: What if the Cloud goes down?

Just like on-site hosting and storage, Cloud hosting can and does fail. Earlier this year, Amazon’s Cloud service, Amazon Web Services, went down for just four hours. Even that relatively brief period of downtime cost companies in the S&P 500 index an estimated $150 million. Losing access to your data means losing productivity, losing sales, and losing face. And for some businesses, a four hour outage can be crippling.

However, the advantages of moving to the Cloud (which we’ve detailed in our Hybrid Cloud series) are still very real, and the threat of an outage shouldn’t deter you from taking your business there. While you can’t control when a Cloud provider’s systems will go down, you can control your own outage preparation. You can create a redundant system that allows you to access mission-critical systems and data regardless of your connection to any one Cloud.

In this post, we’ll teach you how to preemptively prepare for Cloud outages using Multi-Cloud and Hybrid Cloud architecture, giving you the means to access your data even if the Cloud goes down.

First Things First: Identify What Data’s Important

Creating redundant data access costs money. And while the details differ, both the outage mitigation strategies we’ll outline below amount to the same thing: investing in multiple data centers to create redundancy.

We know you can’t spend a fortune on data storage. That’s part of why you wanted to move to the Cloud in the first place, right? To ensure you back up only essential data and applications, we recommend splitting your systems into three tiers, Mission-Critical, Nice to Have, and Not Needed, using the following guidelines:

  • For each system, ask yourself, is this truly critical? If it went down for a few hours, could our company make do? If it went down for a day, could we make do? If you answer “No” to either of these questions, you’ve found a Mission-Critical system you need to back up.

  • When you identify a non-critical system, ask yourself whether it’s important enough to merit backing up anyways. Would your company gain more by backing up the system than you’d lose by having the system fail during an outage? If so, you’ve identified a Nice to Have system.

  • If a system doesn’t meet either the first or second guidelines, classify it as Not Needed. Backing this system up in case of an outage would likely be a waste of money.

After you’ve identified what you need to back up and what you’d like to back up, start looking at how much you can spend on redundancy. Hopefully, you’ll have enough funds to back up your Mission-Critical systems at least once. If you have more money to work with, you can look at backing up your Nice to Haves or creating extra layers of redundancy for your Mission-Critical systems.

The good news is, data storage is cheaper than ever, especially in the Cloud. You can use the ubiquity of Cloud-based storage to your advantage, creating a Multi-Cloud infrastructure that protects you in the event one Cloud service provider goes down.

More Clouds, Less Downtime: Multi-Cloud

The most efficient way to protect yourself from Cloud outages is to store your data in more than one provider’s Cloud. This strategy, called Multi-Cloud, assumes it’s unlikely multiple Cloud providers will fail at once. So when one provider goes down, you can switch your system over to another Cloud containing the same data, reducing or even eliminating downtime.

There are a couple kinks in the Multi-Cloud strategy. The first is keeping your multiple Clouds in sync, so data in one is accurately reflected in the other(s). Thankfully, cloud control platforms and onramps, such as OpenStack, can help you manage this synchronization. If you decide to manage Cloud synchronization in-house, make sure your IT team has a plan for driving data to all of your Clouds at once, so you don’t lose anything in an outage. This might require revamping some of your business processes and functions to ensure they play nicely with each Cloud provider’s storage methods. In this case, it might pay to update your system by converting pieces of it into easily accessible APIs.

The second kink is implementing a failover method that kicks your system from one Cloud to another when an outage strikes. Ideally, failover should be automated, so the switch from one Cloud to another is seamless. Platforms such as Chef allow you to easily design, implement, and test your failover plan, so you’ll know your system’s availability before an outage hits.

After you’ve worked out these details, and also determined which providers to store your data with, you can create your Multi-Cloud. In addition to protecting you from Cloud outages, Multi-Cloud architecture protects you if any one Cloud provider goes out of business completely. You might also find that different Clouds are better suited for different processes, allowing you to optimize access to your system.

However, for some critically important data, extra Cloud storage isn’t enough. After all, what happens if your Internet connection itself goes down? We recommend storing this data on-site but connecting it to the Cloud, using Hybrid Cloud architecture.

Store Data Locally, Access It Globally: Hybrid Cloud

Local data storage is still good for something. It allows you to keep your own copy of truly mission-critical data and systems, which you can rely on no matter whether one Cloud or several go down. However, in our increasingly connected world, you’ll want to make sure your local data is accessible over the Internet and through the Cloud as well. This strategy, known as the Hybrid Cloud, protects you from Cloud outages by allowing you to access and update your data locally during an outage, and then push those updates out to the Cloud after service resumes.

The Hybrid Cloud strategy requires that you make your local data easily accessible, and that you retain enough local storage to back up your mission-critical systems and data. What you get in return is a system that’s completely Cloud-outage proof. The entire Internet could go down, but you’ll still have access to your essential data and systems, so you can keep working in-house while you wait for your Cloud providers to come back up.


Regardless of how you do it, it’s extremely important to prepare for Cloud outages by creating a redundant data access system that won’t go down if any one Cloud provider does. We’ve outlined two methods of building redundancy, using Multi-Cloud and Hybrid Cloud strategies, but we know those methods aren’t the only ones out there. We’d love to hear how you’re preparing for potential Cloud outages, and whether your strategies differ from our own, so feel free to tell us your story in the comments below!

Interested in learning more about Hybrid Cloud?
Download our free whitepaper to get started.

New call-to-action

Topics: Cloud and hybrid cloud, Thriftly

Matt Ledger

Written by Matt Ledger

Matt is a former journalist and current blogger who loves technology and learning new things. He’s been writing in the software industry for nearly five years, having recently landed at Mertech.