IRVINE, CA – January 24, 2018 - appLariat today announced the SmartConfig AI Engine, the first AI-based solution that automatically allocates infrastructure resources to containerized applications to assure that they meet changing loads and SLAs, while minimizing Kubernetes infrastructure resources. Integrated with the appLariat Delivery-as-a-Service (DaaS) platform, the full solution allows new and existing enterprise applications to be automatically converted to cloud native and delivered on-demand and fully optimized, to any cloud, for any purpose.
“One of the most vexing challenges enterprises face when delivering containerized applications to Kubernetes is configuring application resources to meet expected load and SLAs while at the same time minimizing the cost of the underlying Kubernetes infrastructure,” said Mazda Marvasti, CEO, appLariat. “The appLariat SmartConfig AI Engine is the first solution to leverage AI and machine learning techniques to ensure applications meet changing loads within SLAs while optimizing the underlying Kubernetes environment to minimize cloud infrastructure costs.”
To understand the challenges of running apps in containers on Kubernetes, consider that the effective management and placement of hundreds, or even thousands of application workloads into a given Kubernetes cluster must take into account numerous factors. These include:
- Application resource demands
- Prioritization of application workloads
- SLA requirements
- Tolerance for disruptions
- Affinity and anti-affinity concerns
- Cluster capacity
- and horizontal vs vertical scaling
While Kubernetes supports these concepts, it has no way of automatically determining, assigning or managing them and does nothing to optimize the underlying infrastructure in support of application requirements. These decisions are currently left to the development or operations personnel deploying the applications, however the facts are that optimum configuration of cloud native apps has crossed the threshold of human capability.
To that end, the appLariat DaaS platform now includes the SmartConfig AI engine, which leverages a supervised learning mechanism that is then utilized in real time to extract configuration information, which appLariat then applies to the deployed application running on the Kubernetes infrastructure.
The workflow of SmartConfig is as follows:
- The appLariat DaaS platform initiates a deployment of the application to a test cluster
- The SmartConfig Stochastic App Tester then runs a series of tests on the application to determine its behavior
- The results are then used as the training set for multi-level supervised learning algorithms, which represent the non-linear behavior of the inputs (application resource allocations) to the outputs (application load and SLA).
- With initial learning complete, the app is deployed in production and SmartConfig AI is enabled for the expected load. A configuration extractor then uses the learned behavior to extract an optimized resource allocation configuration for the currently running application based on the current or desired load and SLA.
- The optimization mechanism extracts the configuration that produces the smallest node footprint for the Kubernetes infrastructure.
- The changes are then applied to the application via the appLariat DaaS API.
- External monitoring systems are configured to request updated configurations as load and/or SLA changes are identified in real time.
With appLariat’s SmartConfig AI Engine integrated with the customer’s chosen monitoring solution, they are assured that all application deployments will meet varying load within SLAs on optimized Kubernetes infrastructure, minimizing cloud infrastructure costs, while assuring application performance. No other continuous delivery platform provides this capability.
appLariat provides the industry’s only AI-driven application Delivery-As-A-Service (DaaS) platform that makes enterprise applications cloud native and delivers them, on-demand and fully optimized, to any cloud, for any purpose. Enterprise customers select appLariat to automate containerization, deployment, optimization and ongoing management of their applications and Kubernetes, across private, public and hybrid cloud environments. appLariat provides powerful policy control capabilities to ensure enterprise-class governance, compliance and security. With appLariat, customers reduce their cloud infrastructure costs by up to 50% through optimized delivery and management of application deployments throughout the development lifecycle. To learn more, visit www.applariat.com.