In a move to improve efficiency for enterprises processing video content at high volumes, Akamai added a media-optimized offering based on Nvidia GPUs to its cloud services portfolio

Although Nvidia GPUs are typically discussed for supporting large language model (LLM)  and artificial intelligence (AI)/machine learning (machine learning) workloads, this new cloud service from Akamai targets businesses currently underserved by existing GPU offerings, which tend to be too expensive and without the exact functionality needed by media and entertainment companies.

“Media companies need low-latency, reliable compute resources that maintain the portability of the  workloads they create,” Akamai VP of Cloud Products Shawn Michels said. “What we’re doing with industry-optimized GPUs is one of many steps we’re taking for our customers to increase instance diversity across the entire continuum of compute to drive and power edge-native applications,” Michels said.

Akamai’s new service is based on the Nvidia RTX 4000 Ada Generation GPU, a single slot graphics card that processes frames per second 25-times faster than traditional CPU-based encoding and transcoding processes, according to internal benchmarking conducted by Akamai.

Nvidia RTX 4000 Ada also offers improved price performance when deployed on Akamai’s edge network, according to the vendor. This means media and entertainment companies can build resilient and scalable architectures for deploying faster, more reliable and portable workloads, Michels said.

Broad use cases for the Nvidia RTX 4000 GPU on Akamai’s global edge network include digital content creation, 3D modeling, rendering, inference and video content and streaming.

In addition to media-specific scenarios like video transcoding, live video streaming, virtual reality (VR) and augmented reality (AR), this cloud service also applies to IT teams building applications tied to other industry use cases like generative AI (genAI), data analysis, scientific computing, high performance computing (HPC), gaming and graphics rendering.

“In order to support a wide range of workloads, you need a wide array of compute instances,” Michels said.

The Nvidia RTX 400 GPU uses the Nvidia Ada Lovelace architecture to address a primary application of GPU cloud computing – genAI. With 20 gigabytes (GB) of GDDR6 memory for large models and datasets, 192 fourth-gen Tensor Cores and a new Fine-Grained Structured Sparsity feature, the GPU provides 4-times faster throughput for tensor matrix operations when compared to the previous generation of technology.

GPU cloud computing is also common in data analysis due to the nature of processing vast amounts of data. GPUs can accelerate these time- and computationally-intensive tasks by processing data in parallel for optimized analysis.

In addition, HPC applications like modeling and simulation are worthy targets of GPU-enabled cloud computing. The tech can speed up simulation building, calculations and other computationally-intensive processes for improved results and performance.