Arista Introduces Intelligent Innovations for AI Networking
Arista Networks (NYSE: ANET) has unveiled advanced capabilities to enhance AI cluster performance and efficiency. The company introduced Cluster Load Balancing (CLB) in Arista EOS®, designed to maximize AI workload performance with consistent, low-latency network flows.
The new EOS Smart AI Suite features RDMA-aware flow placement for uniform high performance and low tail latency. Additionally, Arista's CloudVision Universal Network Observability (CV UNO) now provides AI job-centric observability with features including AI Job Monitoring, Deep-Dive Analytics, Flow Visualization, and Proactive Resolution.
The company's Etherlink AI Platforms offer 800G/400G systems compatible with Ultra Ethernet Consortium. CLB is currently available on select platforms (7260X3, 7280R3, 7500R3, 7800R3), with support for 7060X6 and 7060X5 platforms scheduled for Q2 2025. CV UNO enhancements for AI are in customer trials, with general availability planned for Q2 2025.
Arista Networks (NYSE: ANET) ha svelato capacità avanzate per migliorare le prestazioni e l'efficienza dei cluster AI. L'azienda ha introdotto Cluster Load Balancing (CLB) in Arista EOS®, progettato per massimizzare le prestazioni dei carichi di lavoro AI con flussi di rete costanti e a bassa latenza.
Il nuovo EOS Smart AI Suite presenta un posizionamento dei flussi consapevole di RDMA per prestazioni elevate uniformi e bassa latenza finale. Inoltre, il CloudVision Universal Network Observability (CV UNO) di Arista ora offre osservabilità centrata sui lavori AI con funzionalità che includono Monitoraggio dei Lavori AI, Analisi Approfondita, Visualizzazione dei Flussi e Risoluzione Proattiva.
Le piattaforme Etherlink AI dell'azienda offrono sistemi da 800G/400G compatibili con l'Ultra Ethernet Consortium. CLB è attualmente disponibile su piattaforme selezionate (7260X3, 7280R3, 7500R3, 7800R3), con supporto per le piattaforme 7060X6 e 7060X5 previsto per il secondo trimestre del 2025. I miglioramenti di CV UNO per l'AI sono in fase di prova presso i clienti, con disponibilità generale pianificata per il secondo trimestre del 2025.
Arista Networks (NYSE: ANET) ha presentado capacidades avanzadas para mejorar el rendimiento y la eficiencia de los clústeres de IA. La compañía introdujo Cluster Load Balancing (CLB) en Arista EOS®, diseñado para maximizar el rendimiento de las cargas de trabajo de IA con flujos de red consistentes y de baja latencia.
El nuevo EOS Smart AI Suite presenta una colocación de flujo consciente de RDMA para un alto rendimiento uniforme y baja latencia final. Además, el CloudVision Universal Network Observability (CV UNO) de Arista ahora proporciona observabilidad centrada en trabajos de IA con características que incluyen Monitoreo de Trabajos de IA, Análisis Profundo, Visualización de Flujos y Resolución Proactiva.
Las plataformas Etherlink AI de la compañía ofrecen sistemas de 800G/400G compatibles con el Ultra Ethernet Consortium. CLB está actualmente disponible en plataformas seleccionadas (7260X3, 7280R3, 7500R3, 7800R3), con soporte para las plataformas 7060X6 y 7060X5 programado para el segundo trimestre de 2025. Las mejoras de CV UNO para IA están en pruebas con clientes, con disponibilidad general prevista para el segundo trimestre de 2025.
아리스타 네트웍스 (NYSE: ANET)는 AI 클러스터 성능과 효율성을 향상시키기 위한 고급 기능을 공개했습니다. 이 회사는 일관된 저지연 네트워크 흐름으로 AI 작업 성능을 극대화하도록 설계된 클러스터 로드 밸런싱 (CLB)을 Arista EOS®에 도입했습니다.
새로운 EOS 스마트 AI 스위트는 균일한 높은 성능과 낮은 꼬리 지연을 위한 RDMA 인식 흐름 배치를 특징으로 합니다. 또한, 아리스타의 클라우드비전 유니버설 네트워크 옵저버빌리티 (CV UNO)는 AI 작업 모니터링, 심층 분석, 흐름 시각화 및 능동적 해결을 포함한 기능으로 AI 작업 중심의 옵저버빌리티를 제공합니다.
회사의 이더링크 AI 플랫폼은 울트라 이더넷 컨소시엄과 호환되는 800G/400G 시스템을 제공합니다. CLB는 현재 선택된 플랫폼(7260X3, 7280R3, 7500R3, 7800R3)에서 사용할 수 있으며, 7060X6 및 7060X5 플랫폼에 대한 지원은 2025년 2분기로 예정되어 있습니다. AI를 위한 CV UNO 개선 사항은 고객 시험 중이며, 일반 가용성은 2025년 2분기로 계획되어 있습니다.
Arista Networks (NYSE: ANET) a dévoilé des capacités avancées pour améliorer les performances et l'efficacité des clusters d'IA. L'entreprise a introduit Cluster Load Balancing (CLB) dans Arista EOS®, conçu pour maximiser les performances des charges de travail d'IA avec des flux réseau constants et à faible latence.
La nouvelle EOS Smart AI Suite propose un placement de flux conscient de RDMA pour des performances élevées uniformes et une faible latence finale. De plus, le CloudVision Universal Network Observability (CV UNO) d'Arista offre désormais une observabilité centrée sur les travaux d'IA avec des fonctionnalités telles que le Suivi des Travaux d'IA, l'Analyse Approfondie, la Visualisation des Flux et la Résolution Proactive.
Les plateformes Etherlink AI de l'entreprise offrent des systèmes 800G/400G compatibles avec l'Ultra Ethernet Consortium. CLB est actuellement disponible sur des plateformes sélectionnées (7260X3, 7280R3, 7500R3, 7800R3), avec un support pour les plateformes 7060X6 et 7060X5 prévu pour le deuxième trimestre 2025. Les améliorations de CV UNO pour l'IA sont en phase d'essai chez les clients, avec une disponibilité générale prévue pour le deuxième trimestre 2025.
Arista Networks (NYSE: ANET) hat fortschrittliche Funktionen vorgestellt, um die Leistung und Effizienz von AI-Clustern zu verbessern. Das Unternehmen führte Cluster Load Balancing (CLB) in Arista EOS® ein, um die Leistung von AI-Workloads mit konsistenten, latenzarmen Netzwerkflüssen zu maximieren.
Die neue EOS Smart AI Suite bietet RDMA-bewusste Flussplatzierung für gleichmäßige hohe Leistung und niedrige Endlatenz. Darüber hinaus bietet Aristas CloudVision Universal Network Observability (CV UNO) jetzt AI-jobzentrierte Beobachtbarkeit mit Funktionen wie AI Job Monitoring, Deep-Dive Analytics, Flow Visualization und Proaktive Lösung.
Die Etherlink AI Plattformen des Unternehmens bieten 800G/400G Systeme, die mit dem Ultra Ethernet Consortium kompatibel sind. CLB ist derzeit auf ausgewählten Plattformen (7260X3, 7280R3, 7500R3, 7800R3) verfügbar, während die Unterstützung für die Plattformen 7060X6 und 7060X5 für das zweite Quartal 2025 geplant ist. Die Verbesserungen von CV UNO für AI befinden sich in Kundenversuchen, mit einer allgemeinen Verfügbarkeit, die für das zweite Quartal 2025 geplant ist.
- Introduction of new AI-optimized networking solutions enhancing performance and efficiency
- Real-time monitoring and analytics capabilities through CV UNO platform
- Support for high-speed 800G/400G connectivity
- Scalability from small clusters to 100,000+ accelerators
- Extended timeline for full platform support, with some features not available until 2H 2025
- AI observability enhancements still in trial phase
Insights
Arista Networks has unveiled key innovations that strengthen its positioning in the strategically important AI networking infrastructure market. The introduction of Cluster Load Balancing (CLB) and enhanced CloudVision Universal Network Observability (CV UNO) capabilities addresses critical technical bottlenecks that have AI workload performance.
The CLB innovation is especially notable as it solves a significant technical challenge - optimizing large bandwidth flows that are characteristic of AI workloads. By implementing RDMA-aware flow placement, Arista is targeting improved throughput and reduced latency, which directly impacts AI training and inference efficiency. The endorsement from Oracle Cloud Infrastructure adds significant validation to the real-world applicability of these solutions.
From a competitive standpoint, these enhancements further differentiate Arista's Etherlink AI Platforms in the high-performance networking space, which is becoming increasingly critical as organizations scale their AI infrastructure. The company's forward compatibility with Ultra Ethernet Consortium standards also demonstrates strategic foresight in a rapidly evolving technical landscape.
While immediate financial impact isn't quantifiable from this announcement alone, the innovations represent Arista's continued execution in capturing market share in high-growth AI infrastructure deployments. The phased availability approach (some features available immediately, others coming in Q2 and 2H 2025) suggests a methodical product roadmap aligned with enterprise adoption cycles.
Arista's new EOS Smart AI Suite represents a significant technical advancement for AI networking infrastructure. The Cluster Load Balancing technology addresses a fundamental challenge in AI clusters - the presence of fewer but larger bandwidth flows that traditional load balancing approaches handle poorly.
What makes this innovation particularly valuable is its RDMA-aware flow placement capabilities. By optimizing traffic flow bi-directionally (leaf-to-spine and spine-to-leaf), CLB ensures more consistent performance with lower tail latency - a critical metric for distributed AI workloads where the slowest node often determines overall performance. The microsecond-level visibility provided by the AI Analyzer allows for unprecedented granularity in performance monitoring.
The enhanced CV UNO capabilities deliver substantial operational benefits through comprehensive job monitoring, deep-dive analytics, and flow visualization. The real-time telemetry streaming via EOS NetDL Streamer provides continuous high-frequency data, eliminating the blind spots inherent in traditional SNMP polling approaches.
Equally important is the integration between network and compute layer telemetry, creating a unified observability framework that spans the entire AI infrastructure stack. This holistic approach enables faster troubleshooting and issue resolution - critical capabilities as organizations scale their AI deployments and face increasingly complex performance optimization challenges.
EOS Smart AI Suite fuels peak AI workload performance
Powering Smart AI Networking
The Arista EOS Smart AI Suite is designed for AI-grade robustness and protection and empowers AI clusters with an innovation called Cluster Load Balancing — a new Ethernet-based AI load balancing solution based on RDMA queue pairs that enables high bandwidth utilization between spines and leaves. AI clusters usually have low quantities of large bandwidth flows. Basic load balancing methods are often inefficient for AI workloads, resulting in uneven traffic distribution and increased tail latency. CLB addresses this by using RDMA-aware flow placement, to ensure uniform high performance for all flows while keeping tail latency low. CLB takes a global approach, optimizing traffic flow in both directions, leaf-to-spine and spine-to-leaf, ensuring balanced utilization and consistent low latency.
"As Oracle continues to grow its AI infrastructure leveraging Arista switches, we see a need for advanced load balancing techniques to help avoid flow contentions and increase throughput in ML networks,” said Jag Brar, vice president and Distinguished Engineer, Oracle Cloud Infrastructure. “Arista’s Cluster Load Balancing feature helps do that.”
Holistic AI Observability
CV UNO, the AI-driven 3600 Network Observability platform powered by Arista AVA™, delivers seamless, end-to-end AI job visibility by unifying network, system, and AI job data within the Arista Network Data Lake (NetDL™). EOS NetDL Streamer, a real-time telemetry framework that continuously streams granular network data from Arista switches into NetDL. Unlike traditional SNMP polling, which relies on periodic queries and can miss critical updates, the EOS NetDL Streamer provides low-latency, high-frequency, event-driven insights into network performance, key to supercharging large-scale AI training and inferencing infrastructure. Designed for AI accelerator clusters, it accelerates impact analysis, pinpoints issues with precision, and enables rapid resolution—ensuring job completion times are minimized. Some of the key benefits include:
- AI Job Monitoring – Unlocks a comprehensive view of AI job health metrics, including job completion times, congestion indicators (ECN-marked packets, PFC pause frames, packet drops), and buffer/link utilization for real-time insights.
- Deep-Dive Analytics – Uncovers critical job-specific insights by analyzing network devices, server NICs (e.g., PFC out-of-sync events, RDMA errors, PCIe fatal errors), and associated flows — pinpointing performance bottlenecks with precision.
- Flow Visualization – Harnesses the power of CV topology mapping to gain real-time, intuitive visibility into AI job flows at microsecond granularity — accelerating issue inference and resolution.
- Proactive Resolution – Detects anomalies early and correlates network and compute performance within NetDL — ensuring uninterrupted, high-efficiency AI workload execution.
Arista AI Centers Driven by AVA
Arista’s Etherlink™ AI Platforms deliver ultra-high-performance, standards-based Ethernet systems for next-gen AI networks. Offering 800G/400G fixed, modular, and distributed platforms that are forward-compatible with Ultra Ethernet Consortium (UEC), Etherlink scales from small AI clusters to massive deployments with 100,000+ accelerators. Arista features the AI Analyzer, powered by Arista AVA, which delivers high-resolution traffic data at 100-microsecond intervals, enabling precise performance optimization and troubleshooting. This allows network administrators to optimize performance, quickly troubleshoot issues, and make informed decisions for AI-driven networks. Arista AVA also powers a remote EOS AI Agent, that streams telemetry from SuperNICs or servers to NetDL, ensuring seamless network monitoring, debugging, and QoS consistency across the entire stack.
Availability
- CLB
- Available today on 7260X3, 7280R3, 7500R3 and 7800R3 platforms.
- Support on 7060X6 and 7060X5 platforms scheduled for Q2 2025
- Support for 7800R4 scheduled for 2H 2025
- CV UNO is available today. The observability enhancements for AI are in active customer trials, with general availability scheduled for Q2 2025
Read industry support for this announcement here.
Register here for a webinar on April 10, 2025 to learn more about the EOS Smart AI suite and read a blog authored by Arista’s Praful Bhaidasna, director Product Management, here.
About Arista Networks
Arista Networks is an industry leader in data-driven, client-to-cloud networking for large AI, data center, campus and routing environments. Its award-winning platforms deliver availability, agility, automation, analytics, and security through an advanced network operating stack. For more information, visit www.arista.com.
ARISTA, AGNI, AVA, CloudVision, CV UNO, EOS NetDL Streamer, AI Analyzer, EOS AI Agent, EOS, Etherlink, MSS, and NetDL are among the registered and unregistered trademarks of Arista Networks in jurisdictions worldwide. Other company names or product names may be trademarks of their respective owners. Additional information and resources can be found at www.arista.com. This press release contains forward-looking statements including, but not limited to, statements regarding the performance and capabilities of Arista’s products and services. All statements other than statements of historical fact are statements that could be deemed forward-looking statements. Forward-looking statements are subject to risks and uncertainties that could cause actual performance or results to differ materially from those expressed in the forward-looking statements, including rapid technological and market change, customer requirements, and industry standards, as well as other risks stated in our filings with the SEC available on Arista's website at www.arista.com and the SEC's website at www.sec.gov. Arista disclaims any obligation to publicly update or revise any forward-looking statement to reflect events that occur or circumstances that exist after the date on which they were made.
View source version on businesswire.com: https://www.businesswire.com/news/home/20250312093935/en/
Media Contact:
Amanda Jaramillo
Corporate Communications
Tel: (408) 547-5798
amanda@arista.com
Investor Contact:
Investor Advocacy
Rudolph Araujo
Rod Hall
Tel: (408) 547-8080
ir@arista.com
Source: Arista Networks
FAQ
What is Arista's new Cluster Load Balancing (CLB) feature and how does it improve AI workloads?
When will ANET's new AI networking features be available across all platforms?
What key benefits does Arista's CV UNO provide for AI workloads?
What specifications does ANET's Etherlink AI Platform support?
