StreamForge Media Cut Video Processing Time by Up to 30%
About the client
StreamForge Media provides infrastructure for processing video content.
Its platform receives source material from customers, performs transcoding and preparation of multiple output formats, and delivers the resulting files to customer platforms. The workload combines sustained CPU utilization with large volumes of incoming and outgoing network traffic.
Challenge
StreamForge Media initially relied on a combination of VPS instances and public cloud infrastructure. As processing volumes increased, the company encountered two major limitations.
The first was inconsistent CPU performance during long-running encoding workloads. The second was the growing cost of transferring large volumes of video data. To scale efficiently, StreamForge Media needed infrastructure combining:
- strong single-core and multi-core CPU performance;
- NVMe storage;
- high network throughput;
- large traffic allowances;
- the ability to add processing nodes as demand increased.
Solution
After testing an initial server, StreamForge Media gradually moved its main processing infrastructure to Unihost dedicated servers.
The company selected the KDS-04, a high-bandwidth Unihost configuration powered by an AMD Ryzen 9 9950X with 16 cores, combined with 64 GB DDR5 RAM, 960 GB NVMe storage and a 10 Gbps network connection with up to 330 TB of traffic. The configuration is currently available in Germany, the Netherlands, and Japan.
For StreamForge Media, the combination was particularly well suited to video processing. High CPU performance allows multiple transcoding jobs to run concurrently, while the 10 Gbps connection provides significantly more headroom for transferring large source files and processed video. Several KDS-04 nodes now form the company’s processing pool, while a separate server handles temporary storage and job distribution.
Tasks are automatically assigned to available processing nodes according to current server load, making it straightforward to increase capacity by adding additional servers.
Results
Following the migration, average processing time for typical workloads decreased by approximately 25 – 30%. At the same time, total compute and bandwidth costs fell by roughly 40% compared with the previous infrastructure at a similar workload level. There was another important benefit: predictable unit economics.
When StreamForge Media takes on a large new project, the team can estimate the required infrastructure and its cost in advance instead of maintaining a large contingency for variable cloud bandwidth charges.
“Our requirement is straightforward: the servers need to process video quickly without turning bandwidth into a separate financial risk. Dedicated infrastructure fits that model very well.”
What’s next
StreamForge Media plans to expand its processing cluster and add dedicated storage and backup capacity. As customer file sizes continue to grow, the company also expects to expand its use of Unihost’s 10 Gbps server infrastructure.
Results at a glance: up to 30% faster processing, approximately 40% lower costs, 10 Gbps processing infrastructure
