Keysight Introduces AI Data Center Builder to Validate and Optimize Network Architecture and Host Design
Keysight Technologies (NYSE: KEYS) has launched the Keysight AI (KAI) Data Center Builder, an advanced software suite designed to evaluate AI infrastructure performance through real-world workload emulation. The solution focuses on validating how new algorithms, components, and protocols impact AI training performance.
The software integrates Large Language Model (LLM) and other AI model training workloads into the validation of AI infrastructure components, including networks, hosts, and accelerators. It features a library of LLM workloads like GPT and Llama, with various model partitioning schemas such as Data Parallel, Fully Sharded Data Parallel, and 3D parallelism.
Key capabilities include experimenting with parallelism parameters, analyzing network utilization, identifying bottlenecks, and understanding job completion time impacts. The solution enables AI operators, GPU cloud providers, and infrastructure vendors to validate AI cluster designs and optimize workload performance without expensive large-scale deployments.
Keysight Technologies (NYSE: KEYS) ha lanciato il Keysight AI (KAI) Data Center Builder, una suite software avanzata progettata per valutare le prestazioni dell'infrastruttura AI attraverso l'emulazione di carichi di lavoro reali. La soluzione si concentra sulla validazione di come nuovi algoritmi, componenti e protocolli influenzano le prestazioni di addestramento dell'AI.
Il software integra carichi di lavoro di Large Language Model (LLM) e altri modelli di addestramento AI nella validazione dei componenti dell'infrastruttura AI, inclusi reti, host e acceleratori. Presenta una libreria di carichi di lavoro LLM come GPT e Llama, con vari schemi di partizionamento dei modelli come Data Parallel, Fully Sharded Data Parallel e parallelismo 3D.
Le principali capacità includono la sperimentazione con parametri di parallelismo, l'analisi dell'utilizzo della rete, l'identificazione dei colli di bottiglia e la comprensione degli impatti sui tempi di completamento dei lavori. La soluzione consente agli operatori AI, ai fornitori di cloud GPU e ai fornitori di infrastrutture di convalidare i progetti di cluster AI e ottimizzare le prestazioni dei carichi di lavoro senza costosi deploy su larga scala.
Keysight Technologies (NYSE: KEYS) ha lanzado el Keysight AI (KAI) Data Center Builder, un conjunto de software avanzado diseñado para evaluar el rendimiento de la infraestructura de IA a través de la emulación de cargas de trabajo del mundo real. La solución se centra en validar cómo los nuevos algoritmos, componentes y protocolos impactan en el rendimiento del entrenamiento de IA.
El software integra cargas de trabajo de Large Language Model (LLM) y otros modelos de entrenamiento de IA en la validación de los componentes de infraestructura de IA, incluyendo redes, hosts y aceleradores. Cuenta con una biblioteca de cargas de trabajo LLM como GPT y Llama, con varios esquemas de particionamiento de modelos como Data Parallel, Fully Sharded Data Parallel y paralelismo 3D.
Las capacidades clave incluyen experimentar con parámetros de paralelismo, analizar la utilización de la red, identificar cuellos de botella y comprender los impactos en los tiempos de finalización de trabajos. La solución permite a los operadores de IA, proveedores de nube GPU y vendedores de infraestructura validar los diseños de clústeres de IA y optimizar el rendimiento de las cargas de trabajo sin implementaciones costosas a gran escala.
Keysight Technologies (NYSE: KEYS)는 Keysight AI (KAI) 데이터 센터 빌더를 출시했습니다. 이 고급 소프트웨어 제품군은 실제 작업 부하 에뮬레이션을 통해 AI 인프라 성능을 평가하도록 설계되었습니다. 이 솔루션은 새로운 알고리즘, 구성 요소 및 프로토콜이 AI 훈련 성능에 미치는 영향을 검증하는 데 중점을 둡니다.
이 소프트웨어는 Large Language Model (LLM) 및 기타 AI 모델 훈련 작업 부하를 AI 인프라 구성 요소의 검증에 통합하며, 여기에는 네트워크, 호스트 및 가속기가 포함됩니다. GPT 및 Llama와 같은 LLM 작업 부하의 라이브러리를 제공하며, Data Parallel, Fully Sharded Data Parallel 및 3D 병렬성과 같은 다양한 모델 파티셔닝 스키마를 포함합니다.
주요 기능으로는 병렬성 매개변수 실험, 네트워크 활용도 분석, 병목 현상 식별 및 작업 완료 시간 영향 이해가 포함됩니다. 이 솔루션은 AI 운영자, GPU 클라우드 공급자 및 인프라 공급자가 AI 클러스터 설계를 검증하고 비용이 많이 드는 대규모 배포 없이 작업 부하 성능을 최적화할 수 있도록 합니다.
Keysight Technologies (NYSE: KEYS) a lancé le Keysight AI (KAI) Data Center Builder, une suite logicielle avancée conçue pour évaluer les performances des infrastructures d'IA à travers l'émulation de charges de travail réelles. La solution se concentre sur la validation de la manière dont les nouveaux algorithmes, composants et protocoles impactent les performances de formation de l'IA.
Le logiciel intègre des charges de travail de Large Language Model (LLM) et d'autres modèles de formation d'IA dans la validation des composants d'infrastructure d'IA, y compris les réseaux, les hôtes et les accélérateurs. Il dispose d'une bibliothèque de charges de travail LLM comme GPT et Llama, avec divers schémas de partitionnement de modèles tels que Data Parallel, Fully Sharded Data Parallel et parallélisme 3D.
Les capacités clés incluent l'expérimentation avec des paramètres de parallélisme, l'analyse de l'utilisation du réseau, l'identification des goulets d'étranglement et la compréhension des impacts sur les temps d'achèvement des tâches. La solution permet aux opérateurs d'IA, aux fournisseurs de cloud GPU et aux fournisseurs d'infrastructure de valider les conceptions de clusters d'IA et d'optimiser les performances des charges de travail sans déploiements coûteux à grande échelle.
Keysight Technologies (NYSE: KEYS) hat den Keysight AI (KAI) Data Center Builder eingeführt, eine fortschrittliche Software-Suite, die entwickelt wurde, um die Leistung von KI-Infrastrukturen durch die Emulation von realen Arbeitslasten zu bewerten. Die Lösung konzentriert sich darauf, zu validieren, wie neue Algorithmen, Komponenten und Protokolle die Leistung des KI-Trainings beeinflussen.
Die Software integriert Arbeitslasten von Large Language Model (LLM) und anderen KI-Modelltrainings in die Validierung von KI-Infrastrukturkomponenten, einschließlich Netzwerken, Hosts und Beschleunigern. Sie verfügt über eine Bibliothek von LLM-Arbeitslasten wie GPT und Llama sowie über verschiedene Modellpartitionierungs-Schemata wie Data Parallel, Fully Sharded Data Parallel und 3D-Parallelen.
Zu den wichtigsten Funktionen gehören das Experimentieren mit Parallelisierungsparametern, die Analyse der Netzwerknutzung, die Identifizierung von Engpässen und das Verständnis der Auswirkungen auf die Abschlusszeiten von Jobs. Die Lösung ermöglicht es KI-Betreibern, GPU-Cloud-Anbietern und Infrastruktur-Anbietern, die Entwürfe von KI-Clustern zu validieren und die Arbeitslastleistung ohne teure großangelegte Bereitstellungen zu optimieren.
- Introduction of new revenue-generating product in the growing AI infrastructure market
- Product enables cost savings for customers by avoiding expensive large-scale deployments
- Addresses critical market need for AI infrastructure validation and optimization
- No specific revenue projections or pricing details provided
- Competitive positioning and market share information not disclosed
Insights
Keysight's new KAI Data Center Builder represents a strategic expansion into the high-growth AI infrastructure validation market. This software suite addresses a critical pain point for AI operators by emulating real-world AI workloads without requiring physical large-scale deployments - traditionally a major capital expense barrier.
The solution's ability to validate network architecture and host design is particularly valuable as AI operators struggle with optimizing expensive GPU clusters. By emulating LLM workloads like GPT and Llama, Keysight enables customers to identify bottlenecks and optimize configurations before deploying actual hardware - potentially saving
The timing is excellent as AI infrastructure spending continues accelerating, with organizations deploying increasingly complex GPU clusters requiring specialized validation tools. The product targets multiple high-value customer segments: AI cloud providers, infrastructure vendors, and enterprise AI operators.
This release strengthens Keysight's position in the AI testing market, complementing their existing test and measurement portfolio while extending their addressable market. By focusing on performance optimization, Keysight is targeting a key concern for AI operators faced with massive capital costs for GPU infrastructure and aiming to maximize return on these investments.
This product addresses a fundamental cost-to-value gap in AI infrastructure deployment. When building AI systems, organizations face a catch-22: they need production-scale environments to properly test workloads, but can't justify the
The KAI Data Center Builder creates a crucial bridge by providing sophisticated workload emulation that simulates network communication patterns of actual AI training jobs. This permits fine-tuning of critical parameters before physical deployment, solving a major risk-management challenge.
The solution targets precisely where AI infrastructure underperforms - in data movement efficiency between GPUs. By allowing experimentation with different parallelism parameters and partition schemas, it helps optimize the extremely expensive GPU resources that typically cost
Most valuable is the ability to identify specific bottlenecks in collective operations, analyze network congestion, and understand impacts on job completion time. Organizations deploying LLMs like GPT typically see
- Validates the performance of AI infrastructure by emulating real-world workloads
- Evaluates how new algorithms, components, and protocols improve the performance of AI training
- Adjusts and optimizes the parameters of both AI workloads and system infrastructure without investing in expensive large-scale deployments
Keysight AI (KAI) Data Center Builder is an advanced software suite that emulates real-world workloads to evaluate how new algorithms, components, and protocols impact the performance of AI training.
AI operators use various parallel processing strategies, also known as model partitioning, to accelerate AI model training. Aligning model partitioning with AI cluster topology and configuration enhances training performance. During the AI cluster design phase, critical questions are best answered through experimentation. Many of the questions focus on data movement efficiency between the graphics processing units (GPUs). Key considerations include:
- Scale-up design of GPU interconnects inside an AI host or rack
- Scale-out network design, including bandwidth per GPU and topology
- Configuration of network load balancing and congestion control
- Tuning of the training framework parameters
The KAI Data Center Builder workload emulation solution reproduces network communication patterns of real-world AI training jobs to accelerate experimentation, reduce the learning curve necessary for proficiency, and provide deeper insights into the cause of performance degradation, which is challenging to achieve through real AI training jobs alone. Keysight customers can access a library of LLM workloads like GPT and Llama, with a selection of popular model partitioning schemas like Data Parallel (DP), Fully Sharded Data Parallel (FSDP), and three-dimensional (3D) parallelism.
Using the workload emulation application in the KAI Data Center Builder enables AI operators to:
- Experiment with parallelism parameters, including partition sizes and their distribution over the available AI infrastructure (scheduling)
- Understand the impact of communications within and among partitions on overall job completion time (JCT)
- Identify low-performing collective operations and drill down to identify bottlenecks
- Analyze network utilization, tail latency, and congestion to understand the impact they have on JCT
The KAI Data Center Builder's new workload emulation capabilities enable AI operators, GPU cloud providers, and infrastructure vendors to bring realistic AI workloads into their lab setups to validate the evolving designs of AI clusters and new components. They can also experiment to fine-tune model partitioning schemas, parameters, and algorithms to optimize the infrastructure and improve AI workload performance.
Ram Periakaruppan, Vice President and General Manager, Network Test & Security Solutions, Keysight, said: "As AI infrastructure grows in scale and complexity, the need for full-stack validation and optimization becomes crucial. To avoid costly delays and rework, it's essential to shift validation to earlier phases of the design and manufacturing cycle. KAI Data Center Builder’s workload emulation brings a new level of realism to AI component and system design, optimizing workloads for peak performance.”
KAI Data Center Builder is the foundation of the Keysight Artificial Intelligence (KAI) architecture, a portfolio of end-to-end solutions designed to help customers scale artificial intelligence processing capacity in data centers by validating AI cluster components using real-world AI workload emulation.
Keysight will showcase KAI Data Center Builder and its workload emulation capabilities in booth #1301 at the OFC 2025 conference, April 1-3, at the Moscone Center,
Resources
- Product Page: KAI Data Center Builder
About Keysight Technologies
At Keysight (NYSE: KEYS), we inspire and empower innovators to bring world-changing technologies to life. As an S&P 500 company, we’re delivering market-leading design, emulation, and test solutions to help engineers develop and deploy faster, with less risk, throughout the entire product lifecycle. We’re a global innovation partner enabling customers in communications, industrial automation, aerospace and defense, automotive, semiconductor, and general electronics markets to accelerate innovation to connect and secure the world. Learn more at Keysight Newsroom and www.keysight.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20250401898311/en/
Keysight Media Contacts
Andrea Mueller
andrea.mueller@keysight.com
Fusako Dohi
+81 42 660-2162
fusako_dohi@keysight.com
Jenny Gallacher
+44 (0) 7800 737 982
jenny.gallacher@keysight.com
Source: Keysight Technologies, Inc.
FAQ
What is the main purpose of Keysight's new KAI Data Center Builder software?
Which LLM workloads are included in the KEYS Data Center Builder library?
What specific performance metrics can be analyzed using the KEYS AI Data Center Builder?
How does the KEYS KAI Data Center Builder benefit AI infrastructure development?
