Aviz and TensorWave are collaborating to improve GPU services with innovative RoCE-based AI fabrics

Aviz

Today marks an important milestone: Aviz Networks, a leading provider of AI networking solutions, announces its collaboration with TensorWave, a pioneering provider of GPU-as-a- Service, known. The focus of the cooperation is on intelligent networks to support AI. This is achieved by implementing RoCE-based (RDMA over Converged Ethernet) AI Fabrics to optimize GPU-as-a-Service offerings. The implementation is carried out using Aviz’s Open Network Enterprise Suite (ONES), a cross-vendor SONiC solution.

Leveraging Aviz Networks’ technology, TensorWave will expand its GPU-as-a-Service offering by leveraging innovative AMD-based MI300 accelerators to meet the evolving needs of a market where enterprises are seeking generative To use AI, Large Language Models (LLM) and machine learning for their business activities. Aviz Networks is known for its Networking 3.0 product line, which is at the forefront of AI-powered network operations and network management solutions.

TensorWave has quickly become the leader in GPU-as-a-Service. This cooperation will further strengthen the company’s competencies. The integration of Aviz’s technology will enable the management and operation of RoCE-based AI fabrics from multiple vendors, critical for handling diverse GPUs, DPUs and high-radix switches. ONES’ unique capabilities include RoCE orchestration, real-time visibility, and threshold-based anomaly detection. This makes the solution the only manufacturer-independent AI Fabric Controller on the market.

Also Read: LogicMonitor adds Cost-Optimization Solution for CloudOps and Engineering Teams to its platform

Advanced RoCE orchestration and real-time anomaly detection

The ONES stands out for its innovative ROCE orchestration features that can precisely manage buffer settings, Priority Flow Control (PFC) and Explicit Congestion Notification (ECN). This precise orchestration ensures that data is passed through the network in an efficient manner, which is critical for AI and machine learning applications that require real-time processing.

In addition, thanks to ONES, network reliability is improved through visibility and real-time anomaly detection. The suite continuously monitors the network and detects anomalies such as packet loss. Their response prevents potential disruptions before they can impact AI/ML workloads.

Future-proof functionality and integration of Network Copilot

ONES’ features, including support for high-density platform configurations, are particularly beneficial for environments managed using TensorWave. Network Copilot™’s integration with ONES expands these capabilities with intelligent control and automated management capabilities, simplifying complex network configuration and maintenance.

Vishal Shukla, CEO of Aviz Networks, said: “Our technology is instrumental in enabling TensorWave to offer state-of-the-art GPU services. This represents a major step forward in our mission to simplify and optimize networks for AI and AI for networks.”

Darrick Horton, CEO of TensorWave, commented on the collaboration: “The collaboration with Aviz Networks will significantly enhance our ability to offer best-in-class GPU-as-a-Service to our customers. Aviz’s expertise in supporting RoCE and developing AI Fabrics will be instrumental in expanding our service offering to meet growing demand.”

This collaboration underscores Aviz Networks‘ commitment to providing innovative solutions. It also strengthens the market position of TensorWave, which will be equipped with the most modern GPU resources. With these technologies, Aviz and TensorWave are setting new standards in the industry, ensuring their customers have access to cutting-edge GPU resources optimized for the most demanding applications.

SOURCE: BusinessWire