Winning with Hybrid: How 5 Companies Optimized Performance with Hybrid Cloud Infrastructure

If you’re like 83 percent of CIOs right now, you’re planning to move some workloads away from the public cloud this year.

But there's also a good chance you don’t have the in-house experience to run an on-prem data center or manage a hybrid cloud setup. That can make it hard to know where to start or what to expect.

Not to worry. In this piece, we’ll look at five examples of companies optimizing performance with hybrid cloud infrastructure. As you read, keep in mind: these are intentionally high level. In any real-world scenario, the best course of action requires asking a lot of questions to understand the unique situation of the business in question. 

Example 1: An Overpriced Chatbot Leads to Cloud Repatriation

The situation: A SaaS company introduces an AI chatbot to help customers find support documents. The IT team decides to launch the chatbot with pay-as-you-go GPUs on the public cloud. After all, the chatbot is an experiment and they don’t want to commit too many resources if it’s a flop.

Initially, adoption is low. But then the front-end team updates the website so the CTA to use the chatbot is more prominent. More and more customers start using it, and the company’s cloud bill skyrockets. The CFO sends an angry email to the head of the chatbot team demanding answers.

The diagnosis: This is a classic example of one common driver of cloud repatriation: right tech then, wrong tech now.

The public cloud is an excellent place to experiment for precisely the reasons this team identified: low upfront costs make it inexpensive to run experiments.

But there's another truism about the public cloud: once a workload reaches a steady state, it often makes more sense (financially and operationally) to move it out of the public cloud.

That’s known as cloud repatriation: moving some (or, rarely, all) of your workloads out of the public cloud and into a private cloud, an on-prem data center, or a colocated setting.

In this case, the IT team consulted with an experienced managed services provider to identify an on-premises hosting setup for the chatbot. The result: lower latency and lower, more predictable monthly costs.

The takeaway: If your public cloud bills are significantly higher than they used to be – especially if you’re using GPUs in the public cloud – that’s a sign to look into hybrid cloud solutions like cloud repatriation. A hybrid setup might let you achieve better performance for a similar investment – or maintain your performance for a smaller investment.

Example 2: A Nationwide Outage Leads to More Diversified (and Resilient) Infrastructure

The situation: A cloud architect for a healthcare company deploys resources across multiple availability zones within a single region of the East Coast for redundancy. The platform hums along great – until a few months later, when an unexpected outage brings down the entire system, affecting users nationwide.

The diagnosis: This is an example of what we call right tech, wrong execution.

Tech outages are a reality of life. In architecting solutions, you have to plan for these things, because they happen no matter who’s in charge. The responsibility of the person designing and building cloud architecture is to plan for the most likely scenarios and then deal with the impacts.

In this scenario, the cloud architect did plan for business continuity by deploying across multiple availability zones. But they didn’t think about what would happen in the event of a full-region outage. With what little we know about this scenario, it sounds like a case of the architect not knowing what they didn’t know.

This is why working with a more experienced partner can be so valuable to prevent business outages. They know to ask questions you might not even be thinking about. After an outage, they can help you reassess and regroup with the right questions:

• What caused the outage?
• How can you address that failure point in your design?
• How can you set yourself up for better availability and greater resiliency moving forward?

The takeaway: In this scenario, it’s not necessarily the technology that needs to change. Rather, this organization needs a better plan for potential outages and better mitigation resources.

Example 3: Ghost Instances Lead to Tapping a Managed Services Team

The situation: A global entertainment company hosts a worldwide livestream event on AWS. It was the obvious choice to help the company manage unpredictable spikes in traffic. But after the event ended and audiences dropped off, several instances were left running at full capacity for hours – and the company was still being charged.

The finance team was furious.

The diagnosis: This is another example of right tech, wrong execution. AWS was a solid decision for the point-in-time solution this company needed. But whether someone on the IT team missed key requirements or implemented requirements poorly, the result was expensive. Machines weren’t turned off and they weren’t using auto-scaling to make changes.

The takeaway: The public cloud is an excellent solution for many hosting needs. But it has to be managed properly. Without proper management, costs can quickly spiral, as this company saw.

The good news is that managing a public cloud instance is not something your organization has to do on its own. In fact, most organizations can’t afford to staff enough people with the right skill set to actively and continuously manage their public cloud use.

This is where a trusted partner can prove invaluable, helping ensure that your product or service functions as needed without incurring unexpected costs.

Example 4: Million-Dollar Downloads Leads to a Hybrid Storage Solution

The situation: A video production company stores terabytes of video footage in the public cloud. For a short-term project that requires global collaboration, they need to download large portions of that footage to their laptops for editing.

When they get their cloud bill, it’s 10x what they were expecting.

The diagnosis: This one depends. It might be a case of right tech then (when the use case was archival storage), wrong tech now (when the use case is downloading data to edit). Or maybe it’s wrong tech, right execution – meaning the exact technology ended up being too expensive, but the approach of using the simplest available method to host data made sense.

As I mentioned at the top, every situation is different. Making a “diagnosis” is impossible without knowing the details. Still, thinking about these situations in broad terms can help IT leaders understand the nature of the infrastructure problem they’re trying to solve – and therefore whether a hybrid setup makes sense as a solution.

For example, if the data downloading was a one-time, time-sensitive thing, then maybe the company doesn’t need to make any changes. It’s very possible that the expense would have been justified.

The takeaway: Know what questions to ask before you make a cloud infrastructure decision. There's no one-size-fits-all solution, which means every decision depends on someone considering the specifics of your organization’s situation against the potential solutions available.

Example 5: Forgetting to Hit Delete Leads to Managed Services Support

The situation: A mid-sized SaaS company regularly spins up cloud instances for testing and development. But devs don’t always remember to shut down instances they’re finished with. The result: orphaned databases and other resources that are not being actively used but continue to accumulate costs month over month.

The finance team notices an increasing number of these that aren’t generating business value.

If this one sounds familiar, you’re not alone. Many companies that moved to the public cloud did so without thinking through how their operations would change once they got there.

When devs requested additional servers from an on-prem data center, they submitted a ticket and IT gave them a server. Maybe they followed the same protocol in the cloud, but nobody added a step for shutting down the instances when the work was complete.

The diagnosis: Right tech, wrong execution.

The takeaway: Maybe the most important thing to remember about using the public cloud is that it can do pretty much anything – but that doesn’t mean you should do everything there. In other words: just because you can doesn’t mean you should.

With the support of a managed services team, the SaaS company didn’t have to worry about creating or enforcing rules for spinning up and shutting down instances – or otherwise managing their public cloud usage. They could focus on fixing bugs, building new features, and otherwise helping the business excel.

When You’re Considering a Hybrid Cloud Setup, Follow This 4-Step Process

We’re in a dynamic, constantly evolving world. If it’s been multiple years since you last considered your cloud infrastructure, now is a great time to review it and determine whether a different combination of public cloud, private cloud, on-prem, and colocated hosting might serve you better.

To do that effectively, follow this four-step process:

1. Ask whether you’re in the right place. Whether it’s a massive cloud bill, an unexpected outage, lagging performance, or murmurings you’re hearing about hybrid capabilities, follow your instinct to question your strategy if you’re all-in on the public cloud. Starting the conversation is the first step to ensuring your infrastructure is optimized for growth.
2. Review workload placement. This is a great step to collaborate on with an experienced partner. The goal: assess your current utilization and identify your true infrastructure needs.
3. Get the right plan. While 83 percent of CIOs are considering cloud migration this year, only nine percent expect to go fully on-prem. Consult with an experienced partner to understand the many options available and identify the one that best fits your business needs.
4. Migrate successfully. This might involve timing data migration, refactoring applications, and much more. The goal: end up with a cloud infrastructure that delivers better performance and offers greater financial transparency – and, in many cases, lower bills.

If you're ready to start that process today, get in touch. We’d love to help you understand whether your current situation makes you a good candidate for cloud repatriation.