bare metal with gpu compute

Customer Story: Cloud Performance to Stream Across the Globe

Team Performive
by Team Performive on

CUSTOMER OBJECTIVE

The customer’s goal was to improve data processing, delivery, and performance of video content across the globe. To improve end-user experience, the requested content needed to be optimized in real time for each global end-user’s device and the local connection (bandwidth).

THE PERFORMIVE SOLUTION

In this often multi-region scenario, the team devised a two step process to improve speed and performance. First, the video is transcoded to the optimal delivery format using Bare Metal Servers with GPU Compute at the origin point. The video is then transferred to the customer environment closest to the end-user leveraging Performive’s network performance over inefficient public networks. The customer then streams the video content directly to the end-user device – fully optimized for speed and quality. This process occurs several thousand times a day, turning the Performive Managed Bare Metal Cloud with GPU Compute into a workhorse for improved customer experience.

CUSTOMER OUTCOME & ARCHITECTURE

The customer chose multiple Performive locations across the globe (having the choice over 20) to house their content processing and delivery function. Each location is made up of a group of 5 servers. Three servers process front-end workloads and two GPU servers transcode video and images based on the end-user device and connectivity specifications.

CASE STUDY THE PERFORMIVE SECRET SAUCE

Performive’s global reach combined with GPU technology is an affordable solution to increase performance and delivery speed. Large public cloud providers often don’t have the number of regions available and may not allow small environment deployments. Plus, Performive customers get access to our Devoted Engineering Services, to scale, optimize performance, or adjust cost efficiency at anytime. Adding this team of certified cloud and networking architects to your IT team on-demand improves performance not just during onboarding but throughout the life of your service, giving your IT team time back to focus on priority objectives.

GPU COMPUTE FAQ

Why do GPUs deliver better performance to this customer over a CPU?

The GPUs have thousands of small compute units versus 4-16 in a CPU. Therefore, the GPU can render many high-resolution videos or images concurrently.

What is the rendering speed difference between a GPU and a CPU?

Using a CPU a 4k, 20-minute video may take 40 minutes to convert to a 1080p video. That same video, using a GPU for conversion may take only 2 minutes.

What are other use cases for choosing a GPU over a CPU?

  • Live Stream Content
  • VDI
  • Gaming
  • Data analysis / Big Data / AI