Remote desktops, powered by the DCV protocol, enable organizations to:

• Run graphics-intensive applications remotely without requiring expensive local workstationsdesktop access.

• Streamline workflows by providing browser-based or multi-screen remote desktop access.

• Reduce operational costs by centralizing computing resources and optimizing usage.

Challenges without proper implementation:

• Bandwidth inefficiencies: Poor performance in high-latency or low-bandwidth environments.

• Underutilization of resources: Wasting GPU, CPU, or storage capacity.

• High operational costs: Maintaining costly workstations for each user.

2.1 Performance Optimization

DCV is designed to provide a seamless experience, even in challenging network conditions. To maximize performance:

• Leverage bandwidth-adaptive streaming: DCV dynamically adjusts streaming quality based on network conditions, ensuring near real-time responsiveness with minimal image degradation.

• Deploy GPU-accelerated instances: For graphics-intensive tasks, utilize servers with physical GPUs or vGPUs (virtual GPUs). Optimize GPU memory sharing with 10 – 15 multiple virtual desktops with full 3D performance optimizing GPU, CPU and memory utilization.

• Support high-latency setups: DCV supports high latency of 100-200 ms+ to allow remote users e.g. from different continents to access centrally deployed resources.

​

2.2 Cost Optimization

DCV eliminates the need for expensive local workstations by centralizing compute resources. To optimize costs:

• Share GPU, CPU and memory resources between different virtual sessions ensuring highest efficiency of

• Host multiple full performant 3D virtual Linux sessions for different users on the same physical or virtual Linux server – 10 to 15 sessions are no problem with modern GPUs

​

2.3. Scalability and Flexibility
DCV is inherently scalable and supports a distributed VDI desktop infrastructure as well as dedicated DCV servers for remote users. Best practices include:

• Manage desktop resources: Use the session management strategy matching your needs best leveraging popular schedulers like SLURM, Gridware as well as the DCV Session Manager. Integrated hooks allow for starting servers on-prem as well as in the cloud on-demand.

• Support for multi-monitor setups: DCV can stream up to 4 monitors at large resolutions, making it ideal for engineers working on complex simulations.

• Deploy browser-based access: Use DCV’s HTML5 client to enable secure, device-agnostic access, reducing IT overhead for native client management. For best performance we recommend the native client.

​

2.4. Security and Compliance

DCV prioritizes security to ensure data privacy and compliance:

• TLS encryption: DCV streams pixels instead of geometries, ensuring that sensitive data never leaves the server. All streams are end-to-end encrypted using AES-256.

• Access control policies: Implement strict authentication and access policies to control user permissions and prevent unauthorized access.

• Audit logs: Use audit logs to monitor user activity and ensure compliance with data protection regulations like GDPR.

3.1. Security and Compliance

DCV’s flexibility allows for dynamic resource allocation to balance workloads across visualization nodes:

• Use workload managers: Integrate DCV with workload managers like Slurm or HPC Gridware Cluster Scheduler to allocate resources based on job priority and GPU/CPU availability.

• Application centric scheduling: Automatically assign the best available remote visualization server for the respective engineering application – including DCV application only mode so engineers can focus on their work

​

3.2. Collaboration and Multi-User Support
DCV supports multiple users working collaboratively on the same session:

• Shared sessions: Enable team members to collaborate in real time with full 2D/3D support.

• Role-based access for collaboration: Configure sessions to allow for either collaboration mode (shared editing) or viewing mode, depending on user access policies.

Energy efficiency is critical in large-scale deployments of DCV remote desktops:

• GPU sharing: Use DCV’s GPU-sharing capabilities to support up to 10-15 virtual desktops on a single GPU, reducing energy consumption and hardware costs.

• Sharing resources: Using the resources in a shared approach allows for cost-saving and significantly better resource utilization than assigning dedicated VMs or server to remote desktop sessions.

• Dynamic Voltage and Frequency Scaling (DVFS): For on-premises setups, optimize power consumption by dynamically adjusting GPU and CPU frequencies based on workload demand.

Monitoring and tuning are essential to maintaining optimal performance:

• Use monitoring tools: Tools like the schedulers support by DCV session management or DCV’s built-in monitoring can track session performance, latency, and resource usage in real time.

• Track GPU and network metrics: Analyze GPU usage with nvidia-smi and monitor network bandwidth.

• Regular updates: DCV regular software updates offer users to benefit from new features and performance improvements.

DCV is tailored for a variety of industries:

• Automotive: Volkswagen engineers rely on DCV for 3D remote visualization, accessing high-end Linux workstations from remote locations.

• Electronics: LG Electronics uses DCV to accelerate product design, enabling engineers to access CAE and HPC resources remotely with low latency.