ClearML, the leading solution for unleashing AI in the enterprise, announced the launch of its new multi-tenant GPU-as-a-Service for Enterprise. Now, enterprises, research labs, and public sector agencies with large on-premise or cloud computing clusters can offer and manage shared computing on fully isolated networks for multiple teams, business units, or subsidiaries.
For organizations struggling with efficiently provisioning and tracking compute consumption by stakeholder and issuing invoices or internal chargebacks, ClearML provides real-time tenant-based reporting on computing hours, data storage, API calls, and other chargeable metrics that can be integrated into a customer’s financial system.
For business and AI leaders, ClearML serves as a single pane of glass for monitoring and managing AI resources, offering full visibility and granular control over who and how resources are accessed and used. With this vital transparency, IT teams and AI builders can better control budgets and performance as well as ensure compliance with AI governance and data security requirements.
Also Read: Core42 Enhances AI Cloud with NVIDIA Computing
Previously, enterprises, research labs, and public sector agencies juggling multiple stakeholder access to shared compute could either manually manage provisioning for a single user at a time or through a specialized point solution that needed to be added on top of their already cumbersome tech stack.
However, the existing solutions in the market that aim to tackle this issue fall short on offering a robust, unified, enterprise-grade offering. They lack centralized dashboards and controls, support for complex infrastructures such as multi-cloud setups, legacy HPC environments, or environments with a mix of Kubernetes and bare metal clusters; as well as fully secure multi-tenancy. IT and DevOps teams managing infrastructure allocation manually, or even relying on tools that only solve part of the problem, must still address a significant amount of maintenance and overhead that fails to solve the problem of underutilized resources, resulting in a lost opportunity to improve the AI team’s time-to-production.
ClearML’s hardware-agnostic approach provides a streamlined way for infrastructure leaders to view their clusters on a single network for complete visibility and seamless management while growing their computing power and future-proofing their investments with GPUs from different chip manufacturers. Our broad support for AI orchestration also enables organizations to combine their existing GPU / HPC computing clusters of CPUs.
Modern infrastructure teams are also finding themselves managing multiple different types of computing clusters (on-premise, cloud, hybrid, Kubernetes, bare metal, etc.). By using ClearML’s flexible, hardware- and silicon-agnostic architecture to unify and make accessible all of their different computing clusters together on a single infrastructure network, IT and AI infrastructure leaders can extract the most value from their existing compute investments. They gain visibility and the ability to closely monitor and control which resources are used, ensuring the most expensive computing clusters are always fully utilized.
ClearML supports hybrid infrastructures for any deployment type, enabling both the control plane and compute to be either in the cloud or on-prem, (including air-gapped systems) without compromising on any features or functionality. This seamlessly provides end users with an efficient and secure cloud experience whether accessing compute on-premise or in the cloud.
KEY BENEFITS FOR ENTERPRISES INCLUDE:
- Improved ROI and cost management as organizations get more mileage from their compute investments.
- Optimized availability for AI builders to access computing resources for AI development and deployment along with optimized usage via fractional GPUs.
- Frictionless scalability with the freedom for customers to build additional clusters using any hardware or cloud of their choice. Our support for hybrid infrastructure helps optimize costs and preserve the cloud experience.
- Enterprise-grade security with SSO authentication and LDAP integration that ensures access to data, models, and compute resources are restricted to approved users and teams.
- Full operational oversight with a single pane of glass on an infrastructure that supports diverse workloads, including AI and HPC, by monitoring and controlling job queues.
“Our flexible, scalable, multi-tenant GPUaaS solution is designed to make shared computing a working reality for any large organization,” said Moses Guttmann, Co-founder and CEO of ClearML. “By increasing AI throughput for AI model and large language model (LLM) deployment, we’re helping our clients achieve a frictionless, unified experience and faster time-to-value on their investments in a completely secure manner.”
SOURCE: EIN Presswire
