Stoic provides an Intelligent Data Workbench™ to customers who are looking to simplify and scale big data analytics and drive innovation for major organizations. Their data science customers come from well-established industries that have clear requirements for scalability to Terabytes of data, high performance, security, and data governance.
These enterprise-level requirements drove Stoic to provide a self-hosted option for its cloud-native application.
Stoic understood that an on-prem solution would require a robust delivery platform for its Kubernetes-based application. Creating that simple delivery system, ironically, turned out to be a complex endeavor.
Stoic’s Intelligent Data Workbench™ has always included a cloud-native, SaaS offering. Up and coming data scientists can use Stoic’s free Client Edition, which runs client-side in the browser. Meanwhile, Stoic’s enterprise customers choose the Cloud Edition to meet their more sophisticated needs. These well-established enterprise organizations are from highly regulated industries such as banking, hedge fund management, power utilities, and governmental & intelligence agencies. Bringing Cloud Edition to on-prem environments helps meet the performance and governance requirements that these organizations require.
Stoic’s solution is based on a cloud-native foundation and operates in a Kubernetes cluster. Yet, Stoic’s customers typically operate legacy infrastructure, and not all of them have yet developed expertise deploying Kubernetes.
Stoic knew they needed to simplify the adoption of their solution. They wanted to deliver innovation reliably, and with full lifecycle best practices.
They set out to build their own tools and workflow to deliver the Intelligent Data Workbench™ into customers’ on-prem environments. Bundling manifests worked well for their internal CI/CD process, so Stoic assumed it would work for customers. Unfortunately, Stoic quickly began to encounter roadblocks; for instance, they would need to support a variety of target platforms from bare metal to public cloud. They also began to realize that this process would entail walking individual customers through initial installation, configuration, and updates.
The team returned to their requirements, reconsidering their most basic assumption – customers needed an operational Kubernetes cluster before moving forward. However, only a few of their enterprise customers had moved beyond basic testing of containers and Kubernetes.
The team reviewed and updated their key requirements:
Stoic was worried that using existing standards was going to be too complex for customers. They realized that they could no longer assume their customers had familiarity with Kubernetes, Helm Charts, or manifests.
The team began to see they were creating technical debt, and worse – support issues for their valued customers. They wondered, “who ends up ‘owning’ the environment?” Could Stoic accidentally find itself in the cluster management business?
While attending the CNCF (Cloud-Native Computing Foundation) conference, Ismael Ghalimi, Stoic’s CEO, was introduced to the team at Replicated. He quickly realized he may have just found a solution to his team’s problem.
The Replicated software delivery and lifecycle management solution checked off their key requirements.
Stoic’s project team started working with Replicated KOTS (Kubernetes Off-The-Shelf) immediately. They started with the free trial and within 2 days they were successfully building out test Kubernetes cluster environments and delivering application manifests.
It quickly became clear to the Stoic team that they could improve their time to market by partnering with Replicated. With the expertise they had accrued, they now needed to go beyond standard, out-of-the-box functionality.
The Stoic product team embarked on a six-week project that would finalize their production offering. They worked directly with Replicated technical teams to extend native functionality in three key areas; integrating Helm Charts, integrated package delivery, and licensing:
Helm Charts: The Stoic team was invested in utilizing Helm as a framework to define manifests and planned to continue this practice. By partnering with Replicated, Stoic extended support for Helm Charts, speeding up Stoic’s time to market.
Package Delivery & Configuration: Replicated KOTS allows Kubernetes to be installed, using a single command. The Stoic development team wanted a similar capability for the delivery of their Workbench application. They were able to extend KOTS to enable the end-user to deploy the Stoic application through a web interface. Configuration is completed by presenting the end-user with a simple form – no editing YAML files, with their strict syntax requirements.
License Management: An unexpected benefit for Stoic was that Replicated KOTS solved their license management needs. The team integrated a workflow to allow sales team members to manage end-user licensing. Prior to this workflow, Stoic’s engineering team was tasked with creating and managing license keys. Now, Stoic’s customer-facing sales reps can create, download and forward unique license keys with no engineering burden.
Replicated’s KOTS platform had addressed the key requirements for Stoic’s team. They successfully extended the platform for their unique needs. As their work deepened, they saw that the platform had solved issues that Stoic hadn’t yet considered.
Kubernetes Cluster Deployment: Stoic’s customers have a wide variety of back-end environments, and Stoic recognized that they could not define new architectures. So, the ability to support Kubernetes in the customers’ platform of choice was critical. This was one of the biggest hurdles that Stoic faced when considering an in-house solution.
End-user updates: Large organizations move slowly – and for good reasons. For adherents to CI/CD methodologies, it can be surprising that enterprise customers do not immediately roll out application updates. Instead, they plan and schedule even the most incremental changes.
Stoic’s delivery solution defaults to the customer in these cases. The Workbench team creates incremental improvements, and the customer applies updates in their own time frame.
Air-Gapped environments: Stoic customers often need to work in air-gapped environments that have no outbound connections to the internet. To support such use cases, Stoic extended Replicated KOTS air-gap functionality with its own bundling tools that allow them to move installation and update packages into the isolated environment without having to reach out to the internet.
Support Bundle: Stoic utilizes a diagnostic feature within Replicated KOTS to simplify troubleshooting for their customers. The “support bundle” is a utility framework that captures relevant system information, configuration, and log data. It automates troubleshooting by applying logic to spot errors, known issues, and config changes.
Stoic extends this framework beyond cluster management to capture relevant Workbench diagnostics. The Replicated Support Bundle feature allows customers whose governance prohibits data from leaving their datacenter to scrub sensitive data also from the Stoic log files.
The partnership with Replicated gives Stoic more time to deliver on their customers’ needs. Stoic benefits in three important areas:
Once Stoic discovered they had options to buy rather than build a solution, they quickly understood the benefits. They could get to market faster, and with less technical debt for the future. The partnership with Replicated streamlined their customer support and simplified the deployment of updates which deliver new innovation.
The Replicated KOTS platform enables Stoic to better serve its customers. Most importantly, customers get an improved experience, beginning with initial evaluation all the way through deployment and production operations.
By partnering with Replicated, Stoic’s team is able to focus on improving their product without the distraction and technical debt of managing a complex application delivery lifecycle.
As Ismael puts it, “Not only is Replicated’s technology great but the team behind it is made up of really awesome people. Their product works, the pricing model makes sense, and the level of technical support you get is first-class.”
With their on-premises deployment issues solved, Stoic’s engineers can now get back to what’s important – their own unique functionality. To learn more about how Replicated can provide your organization with Kubernetes deployment solutions, please request a demo today.