Omnissa, NVIDIA Combine Horizon and vGPU for Workplace AI Use Cases
IT DAILY ·
✦ AI Summary
Omnissa on the 11th unveiled AI workload support cases based on the combination of NVIDIA vGPU and Omnissa Horizon.
Omnissa said that Horizon and NVIDIA vGPU can provide AI developers with GPU-accelerated workspaces on demand.
It also explained that in Thick-client applications without API support and in high-risk tasks, AI agents can recognize screens and work, while administrators can join a Live Session to support handoff between humans and AI.
Omnissa, a digital workplace platform specialist, on the 11th unveiled AI workload support cases based on the combination of NVIDIA vGPU and Omnissa Horizon. The topic of the article is Omnissa and NVIDIA’s announcement of workplace AI use cases based on the combination of Horizon and vGPU. NVIDIA vGPU is NVIDIA’s GPU virtualization technology, and Omnissa Horizon is a VDI and application virtualization platform. The combination in question is NVIDIA vGPU and Omnissa Horizon.
The released image shows the Omnissa Horizon structure, and the photo source is Omnissa. The announcement shows how Omnissa has paired NVIDIA’s GPU virtualization technology with its virtual desktop and application virtualization platform to present AI workload support cases.
Recently, companies have been using containers and GPU clusters to build AI infrastructure. For API pipelines and Stateless computing, the combination of container and GPU clusters is regarded as fast and effective. Stateless computing is characterized by one-off tasks.
Omnissa judged that not all AI tasks are suited to that approach. In particular, it said that tasks requiring a persistent desktop environment have limits if handled only with existing container infrastructure. It also said existing container infrastructure has limitations when API support is unavailable and a dedicated Thick-client must be run directly.
As a way to address this, Omnissa proposed an architecture that combines Horizon and NVIDIA vGPU. Omnissa Horizon is a VDI and application virtualization solution. Horizon enables remote access to virtual PCs in a company server, such as Windows and Linux, through personal computers and smartphones, and supports secure work.
NVIDIA vGPU is NVIDIA’s virtualization software technology that allows a single GPU to be shared by multiple virtual machines (VMs). NVIDIA vGPU functions to split the performance of large server GPUs. This allows the divided GPU performance to be allocated among multiple users and virtual PCs (VDI).
Omnissa emphasized that using Horizon and NVIDIA vGPU allows GPU-accelerated workspaces for AI developers to be provided on demand. Developer environments can be set up within minutes, and performance can be scaled to a higher GPU computing tier when needed during work. After the task is complete, resources can be returned.
In this environment, work can be done in sessions protected by DLP functions, and higher GPU profiles can be used for intensive computing. As a result, the time required to build an AI development environment is reduced from several weeks to less than 1 hour.
Meanwhile, many use cases for Thick-client applications without API support exist in core corporate operations. Representative examples include ERP, core financial systems, SCADA, and HMI, and these applications operate based on screens and user interactions. For this reason, simple container-based automation may make it difficult to apply AI in these areas.
Omnissa Horizon and NVIDIA vGPU support AI agents in directly recognizing and interacting with application screens. This assumes an environment in which AI performs work based on the screen.
Such an environment can also be used for high-risk, high-importance tasks such as contract review, AML investigations, and SOC work. However, in those tasks, humans are responsible for the final judgment even after AI has collected and analyzed the data. For this, an environment is needed in which humans can smoothly take over work that AI was performing.
Horizon allows administrators to directly join a Live Session while an AI agent is running. Through this, administrators can see in real time the screen the agent is viewing and the state of the work. This approach does not require reconstructing a separate work environment or reestablishing previous work context, making transitions between humans and AI smoother.
Based on its cooperation with NVIDIA, Omnissa plans to support the expansion of AI adoption by leveraging companies’ existing work environments and infrastructure, and to support AI application even for high-value tasks that are difficult to address with container-based environments alone.
Source: IT DAILY · Seong Won-young
Original: https://www.itdaily.kr/news/articleView.html?idxno=241548
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
View originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.