Ross Schibler, CEOOver the last decade, with the transition to cloud, the process of application development and IT operations have merged into DevOps, giving teams control to rapidly deliver new application features and increase customer satisfaction. Today, DevOps engineers use the continuous integration/continuous deployment (CI/CD) pipeline to automate the process of code testing, delivery and, deployment. Yet, once this code is in production, the process of performance optimization remains predominantly manual. As application optimization requires extensive knowledge, spread across the whole application stack, it is complex, cumbersome and error-prone. This slows down the DevOps teams and yields relatively small improvements in operations. “With organizations globally focusing on the data stream flowing through their applications, the trade-off between performance and cost turns victim,” says Schibler, CEO of Opsani.
Taking matters into hands, Ross Schibler, a serial entrepreneur, went on to create a tool to address the growing issue of automated operations. This marked the birth of Opsani, a company on a mission to empower DevOps teams with automation tools and services supported by AI and machine learning.
“Just like the technological advancement in artificial intelligence (AI) is enabling cars to function without human interference, using AI to tune applications and automate operations will reduce human dependency and bring remarkable gains in performance and efficiency, freeing engineers to create new application features,” Schibler adds.
“Such use of AI benefits engineers coming into the DevOps arena since several parameters are very resource specific and take a long time to study and understand,” explains Bert Armijo, VP of product at Opsani. “These parameters include resources like CPU and memory, OS kernel settings like page sizes, application parameters like cache timeouts or write delays; middleware configuration variables like JVM garbage collection type and pool sizes, database settings—the list goes on. Out of trillions of parameter combinations, that can affect the efficiency of a simple 3-tier application stack, AI finds the best settings to deliver top performance at an optimal cost range,” notes Armijo.
Using AI to tune applications and automate operations will reduce human dependency and bring remarkable gains in performance and efficiency
Opsani operates as a SaaS provider that integrates with the customer’s CI/CD pipeline. The company brings the idea of continuous optimization (CO), powered by AI, as an extension to the current DevOps toolchain into a CI/CD/CO pipeline. The outcomes offered by Opsani are highlighted best through its association with an organization that was running a large application. Plagued with performance and latency issues, the organization spent over six million dollars each year on the application. After engaging Opsani and deploying their application optimization tool, the cost was cut to one-third the initial cost. Similarly, Opsani achieved 260 percent performance optimization for another organization, by reducing the number of replicas and cloud instances in their application.
During the inception of Opsani, Schibler and team created their application optimization tool and tested it on an in-house application. While expecting around 20 to 40 percent optimization, the tool produced a 170 percent improvement in just a day. With this demonstrated success and after witnessing its benefits, approximately three-fourths of the potential clients adopt Opsani’s solution and achieve positive ROI within the same quarter. Recently, team Opsani has introduced a neural network method to find the best results quicker. Continuing the same momentum into the year ahead, Schibler hints that 2019’s focus will be on providing turnkey solutions to customers and multiplying the deployment speed: “We are finding that we can routinely double the efficiency to bring better performance of cloud apps or reduce cost up to two-thirds while maintaining performance.”