Achieving a resilient, high-performance hybrid cloud at the highest efficiency

Within one year, Hyland Software saved over 15% of existing spend on computing costs across AWS and its own datacenters, while maintaining service quality. How? It applied IBM® Turbonomic® to gain control over its highly complex hybrid cloud environment. And it might only be scratching the surface in terms of savings.

As a provider of SaaS and PaaS content management and process automation solutions to more than 14,000 customers worldwide, Hyland’s resource and processing needs are massive. The company deploys its hosted solutions in a hybrid cloud environment, using its own data centers as well as Amazon Web Services (AWS) public cloud. Moreover, Hyland supports over 2,000 customers that trust the company with petabytes of data across all of its products, and the public cloud operate at a huge scale with thousands of Elastic Compute Cloud (EC2) instances and Elastic Block Store (EBS) clusters.

It’s a major challenge to optimize all these computing resources, reducing waste while assuring service performance for customers. That is the mission of Joseph Quinto, Cloud Business Operations Manager at Hyland, and his team. “We have to manage resources across very different sets of parameters,” says Quinto. “Each product has its own resource needs, and within the product sets there are customers across industries with widely varying needs of their own. There are many variables, and each one makes maximizing your resources a different challenge.”

In the past, Hyland tried to manage the challenge by upscaling its teams. But gaining visibility across so many resources in such a large and complex environment—let alone balancing their usage—was beyond human scale. So, Hyland turned to the Turbonomic platform. “We still had a lot of things that were beyond where we wanted them to be,” says Quinto. “But we hit a point where, on our own, we couldn’t do more. That’s where the power of what Turbonomic can do from an AI and machine learning perspective really worked for us.”

The Hyland team applied Turbonomic to their datacenter resources and to a growing portion of their AWS environment. It had immediate impact. “In the first month using Turbonomic on our AWS environment, we took over 2,000 resourcing actions and saved over 8% of our average monthly spend,” says Quinto.

Such fast savings were possible, Quinto says, because Turbonomic comes ready to use: “As soon as we turn it on, it starts to ingest data points and it’s ready to take action.” Turbonomic provides a view of resource usage across the whole technology stack, and uses AI to provide sizing and allocation recommendations. The team can then set policies and rules and automate appropriate resourcing actions at a scale far beyond what was previously possible.

“The amount of balancing we’re able to do now would be impossible to do manually with the number of customers, the number of instances and environments that we’re managing,” says Quinto. “Now, instead of people looking into systems and trying to interpret what capacity they need, they provision a minimum and let Turbonomic manage the resource levels as needed—either scaling up or down. It reduces all the tickets and all the noise that was created before.”

Hyland began using Turbonomic in 2021, applying it almost exclusively to datacenter resources for about two years before extending it to the AWS environment. On the datacenter side, it has become a critical platform for maximizing existing resources and limiting the need to buy more hardware. “It’s no longer a nice-to-have on the datacenter side, it’s really a must-have,” says Quinto. “And it’s getting closer to must-have on the AWS side.”

The team is taking a phased approach with AWS. The initial steps target three core products and use Turbonomic to gain a clear understanding of the minimum resourcing levels they need and set rules to automatically maintain those baselines. “The challenge is the complexity of our environment and data,” says Quinto. “We haven’t built policies or rules for every different permutation of our products. We’ve been strategic about having Turbonomic take actions on the areas where we know we’ve got the right data to support the right decisions.”   

Quinto adds that Turbonomic supplements AWS scaling capabilities in a valuable way. “Some of our products weren’t built on AWS, and they can’t take full advantage of AWS scaling. Turbonomic gives us what we need there.”

Protecting the customer experience

So, how has tighter control over resource usage affected service performance for customers?

“Supporting and improving the experience for customers is the foundation for everything we’re doing,” says Quinto. “We’re using Turbonomic to make sure that customers are given the capacity they need. When they need it. How they need it. We’ve added more customers and we’ve been able to support them with our existing resources, within our established targets.”

In less than a year since applying Turbonomic to the first three products in the AWS environment, Hyland has saved approximately 15% on existing spend. And it has achieved nearly the same 15% in savings related to the datacenter, through resource optimization and cost avoidance.

But there’s much more to come. “The savings so far are significant, but I think we’re still sort of scratching the surface in terms of what we can do,” says Quinto. In fact, as they connect Turbonomic to more of the AWS environment and Hyland products, Quinto and his team have committed to an impressive target moving forward: USD 300 thousand in savings, every month.

One factor that will help hit that target is the combining of Turbonomic with the IBM Cloudability® platform—solutions that work as two complementary levers. The team uses Cloudability to gain cloud cost insights and to surface detailed information about overspend and anomalies, helping them understand where Turbonomic can and should be applied to achieve the greatest savings.    

Quinto also aims to continue refining the systems data fed into Turbonomic to achieve smarter, more nuanced and continuous optimization across all the variables of product type, customer profile and usage patterns. “A product may need more capacity at the front-end layer, another may need more capacity at the back-end layer, or at the search indexes. I don’t know that we’re that smart with Turbonomic yet, but we want to get there. The real goal is to keep improving the sustainability and profitability of our cloud environments while continuing to support and enhance our customers’ experience.”

Hyland Software, Inc. develops and delivers intelligent content solutions for capturing, processing and managing high volumes of diverse content. These solutions help more than 14,000 customers worldwide improve, accelerate and automate operational decisions and workflows. Hyland was founded in 1991 and is based in Westlake, Ohio.