Skip to content
Successfully Optimizing High-Volume VOD Transcoding - NETINT

Successfully Optimizing High-Volume Video-on-Demand Transcoding

In this article, based on Emilien Gobillot‘s presentation during the NETINT symposium ‘ASIC-based transcoding for high-volume use cases,’ we explore the unique requirements of high-volume video-on-demand (VOD) transcoding and share optimization tips. Whether dealing with a highly variable number of incoming live streams, bursts of video uploads or the need to make videos playable quickly, these strategies will help you reduce costs and improve your platform’s performance.

Understanding the Challenges of High-Volume VOD Transcoding

When it comes to high-volume VOD transcoding, there are several key challenges to address:

Managing Bursts of Video Uploads

One of the primary challenges in high-volume VOD transcoding is handling sudden influxes of video uploads. When a large number of videos are received simultaneously, it can put a significant strain on your infrastructure. Your system must scale dynamically to accommodate these bursts without compromising performance or reliability. This requires careful planning, resource allocation, and scalable technologies that effectively handle the increased workload.

Quick Video Playback

In today’s fast-paced digital landscape, users have high expectations regarding video playback. They expect videos to be available and playable within seconds of uploading. This adds pressure to your transcoding process, as you need to ensure that videos are processed and ready for playback in near real-time. Achieving quick video playback requires optimized transcoding workflows, efficient encoding algorithms, and caching and content delivery networks (CDNs) to deliver videos quickly to end-users.

Cost-Effectiveness

Building and maintaining a video infrastructure can be expensive, especially when dealing with high volumes of video content. Finding a balance between performance and cost is crucial for the long-term sustainability of your video operations. You must consider hardware costs, storage expenses, bandwidth usage, and the overall operational overhead. Optimizing your infrastructure for cost-effectiveness involves making intelligent technology choices, leveraging cloud-based solutions when appropriate, and implementing efficient resource utilization strategies.

Carbon Footprint

As businesses become more environmentally conscious, reducing the carbon footprint of video transcoding operations has become an important consideration. Video transcoding can be resource-intensive, consuming significant amounts of energy and contributing to greenhouse gas emissions. Exploring energy-efficient technologies, such as hardware-based encoding solutions or cloud-based services that optimize resource utilization, is essential to address this. Additionally, implementing sustainable practices, such as using renewable energy sources and minimizing unnecessary computations, can help reduce the environmental impact of your video infrastructure.

To address these challenges, exploring efficient and scalable solutions that can handle the demands of high-volume VOD transcoding is essential.

Exploring CPU-based Transcoding: Benefits and Limitations

One common approach to video transcoding is using CPU-based solutions. Let’s take a closer look at the pros and cons of this method:

Benefits of CPU-based Transcoding - Versatility and Flexibility

One of the key advantages of CPU-based transcoding is its versatility. CPUs can handle various transcoding tasks, making them suitable for various video processing scenarios. Whether you need to convert videos between different formats, resize resolutions, adjust bitrates, or apply filters and effects, CPUs have the computational power to handle these tasks effectively. Additionally, CPUs are widely available and can be found in most computing systems, making them a readily accessible option for video transcoding.

With CPUs, you can customize and fine-tune your transcoding pipeline according to your requirements. You can choose from various open-source and commercial transcoding software, which provide extensive options for configuring encoding parameters, applying filters, and optimizing video quality. This flexibility allows you to tailor your transcoding workflow to meet the unique needs of your video content and target platforms, ensuring optimal results.

Limitations of CPU-based Transcoding - Cost Inefficiency, Power Consumption, Scalability Challenges

One of the primary drawbacks of CPU-based transcoding is its cost inefficiency, particularly in high-volume scenarios. Transcoding videos using CPUs can be computationally intensive, requiring significant processing power and time. As the volume of videos increases, the cost of running and maintaining CPU-based transcoding infrastructure can quickly escalate. This includes hardware acquisition, energy consumption, and data center operations expenses. In large-scale video processing environments, the cost of CPU-based transcoding can significantly burden budgets and resources.

CPUs are known for their high power consumption, which can have financial and environmental implications. When transcoding videos using CPUs, the processors require substantial energy to perform the necessary computations. This increased power consumption leads to higher electricity costs and contributes to a larger carbon footprint. In an era where sustainability and eco-friendliness are increasingly important, the power-hungry nature of CPU-based transcoding becomes a concern for organizations looking to minimize their environmental impact.

Scaling CPU-based transcoding infrastructure to meet the demands of high-volume video processing can be challenging and resource-intensive. The need for additional CPU resources becomes evident as the number of videos and transcoding tasks grows. However, adding more CPUs to the infrastructure can be complicated. It requires careful planning, hardware provisioning, and system configuration for optimal performance and load balancing. Moreover, the linear scaling of CPUs may not always provide the desired performance gains, as there can be limitations regarding memory bandwidth, storage I/O, and network throughput. This scalability challenge can hinder the ability to handle sudden spikes in video processing workloads efficiently and lead to bottlenecks in the transcoding pipeline.

While CPU-based transcoding has merits, such as versatility and flexibility, there may be better options for high-volume VOD scenarios. The cost inefficiency, high power consumption, and scalability challenges associated with CPUs can hinder the efficiency and sustainability of large-scale video processing operations. To overcome these limitations, exploring alternative technologies, such as hardware-based transcoding solutions or cloud-based services, can provide more cost-effective, energy-efficient, and scalable options for handling high-volume video transcoding workloads.

While CPU-based transcoding has merits, it may not be the optimal solution for high-volume VOD scenarios. Exploring alternative technologies can be a game-changer in overcoming these limitations.

Embracing Hardware-Based Transcoding: The Power of VPUs

A promising solution for optimizing high-volume VOD transcoding comes from Video Processing Units (VPUs). Here is why:

Increased Transcoding Capacity

One of the key advantages of VPUs is their energy efficiency compared to CPUs. Video transcoding is a computationally intensive process requiring substantial power consumption on general-purpose processors. However, VPUs are designed to optimize power usage specifically for video processing workloads. They incorporate advanced power management techniques, such as dynamic voltage and frequency scaling, to minimize energy consumption without compromising performance. By leveraging VPUs, you can significantly reduce the power footprint of your video transcoding infrastructure, leading to lower electricity costs and a smaller carbon footprint. This eco-friendly aspect of VPUs aligns with the growing global emphasis on sustainability and responsible computing practices.

Cost Reduction

Implementing VPUs in your video transcoding pipeline can lead to substantial cost savings. The increased transcoding capacity and energy efficiency of VPUs translate into reduced hardware requirements and operational expenses. With VPUs, you can process more video content using fewer physical servers, resulting in lower capital expenditures for hardware acquisition and maintenance. Additionally, the lower power consumption of VPUs contributes to reduced electricity costs, which can be a significant factor in large-scale video processing environments. 

Moreover, VPUs often offer better performance per dollar compared to CPU-based solutions, making them a cost-effective choice for high-volume VOD transcoding. By optimizing your transcoding costs, you can allocate resources more effectively and invest in other areas of your video infrastructure to enhance the overall user experience.

When considering VPUs, you must conduct thorough benchmarks to assess performance and video quality. Not all VPUs are created equal, and evaluating different vendor offerings is crucial to ensure they meet your specific requirements. 

Benchmarking involves testing the VPUs with a representative set of video content measuring metrics such as transcoding speed, output quality, and resource utilization. This process helps you determine your transcoding pipeline’s optimal configuration and settings, ensuring you can achieve the desired balance between performance and quality. Additionally, benchmarking allows you to compare the results against your existing CPU-based solution to quantify the potential benefits of transitioning to VPUs.

Once you’ve validated the suitability of VPUs for your specific use case, integrating them into your transcoding pipeline can yield substantial benefits. The integration process typically involves modifying your existing video processing workflow to leverage the capabilities of VPUs. This may require adapting your transcoding software or frameworks to support hardware acceleration and optimizing your pipeline to exploit the VPU’s features fully. By seamlessly integrating VPUs into your infrastructure, you can unlock the potential for faster, more efficient, and cost-effective video transcoding, ultimately enhancing the quality of service you provide to your users.

Maximizing the Potential of VPUs: A Real-World Example

To illustrate the impact of VPUs on high-volume VOD transcoding, let’s consider a real-world example. Suppose you have a server with four VPUs and want to optimize your transcoding costs. Here are some assumptions:

  • Hardware amortization period: 3 years
  • Electricity cost: 30 cents per kilowatt
  • Video format: Full HD at 30 frames per second
  • Codec: H.264
  • Renditions: 5 (at the same frame rate and codec)

With these assumptions in mind, let’s analyze the cost structure of transcoding one minute of video:

  • If the server is utilized at 25% on average over the amortization period, transcoding 1 million minutes of video would cost approximately $446.
  • If the server is fully loaded (which is optimistic but provides a point of reference), transcoding 1 million minutes would cost around $126.

Breaking down the cost components, we see that electricity accounts for a small portion (12% at 25% server utilization) thanks to the power efficiency of VPUs. The server represents a significant portion (40%), while the VPUs contribute half of the cost.

This analysis highlights the cost-effectiveness of VPUs in high-volume VOD transcoding scenarios. Allocating resources to VPUs can optimize transcoding costs and achieve a favorable gross margin.

Unlocking New Opportunities: The Impact of Free Encoding (api.video)

One powerful strategy to drive adoption and enable new use cases is offering free encoding (as seen at api.video). 

By providing free encoding, api.video significantly lowered the entry barrier for product builders looking to integrate video features into their applications, websites, or services. Traditionally, encoding costs have been a significant hurdle for developers, especially those working on early-stage projects or startups with limited budgets. This democratization of video technology opens up new possibilities for a broader range of developers, empowering them to incorporate video functionality into their products seamlessly. 

Another key advantage of api.video’s free encoding is the scalability it provides to users. As businesses grow and their video usage increases, they no longer worry about escalating encoding bills that can hinder their expansion plans. Users can scale their video consumption without the financial burden of encoding costs. This scalability is particularly crucial for platforms that experience rapid growth or sudden spikes in video traffic. 

Many creative ideas and use video cases have been held back due to the associated encoding expenses. However, with eliminating encoding costs, developers and businesses can now explore new frontiers in video-based solutions. For example, educational platforms can offer more interactive and immersive learning experiences by incorporating video content without worrying about the financial impact of encoding large educational videos. Similarly, social causes and non-profit organizations can leverage video to raise awareness, engage supporters, and deliver impactful messages without being burdened by encoding costs. Free encoding empowers individuals and organizations to think outside the box and develop groundbreaking video applications that were once considered economically unfeasible.

Optimizing Your Video Infrastructure for Success

Optimizing your video infrastructure is key to success in the world of high-volume VOD transcoding. By understanding the challenges, exploring efficient technologies like VPUs, and considering strategies such as free encoding, you can create a cost-effective, scalable, and user-friendly video ecosystem.

Remember, the goal is to make it easy for product builders to integrate video into their applications and services. By providing a seamless and affordable video infrastructure, you empower them to focus on creating innovative and engaging video experiences for their users.

So, take action today:

  1. Assess your current video infrastructure and identify areas for optimization.
  2. Explore hardware-based solutions like VPUs to boost transcoding capacity and reduce costs.
  3. Consider offering free encoding to remove barriers and enable new use cases.
  4. Monitor and optimize your video infrastructure to ensure peak performance and cost-effectiveness.

Change is unavoidable. By following these steps and staying ahead of the curve, you’ll be well-equipped to tackle the challenges of high-volume VOD transcoding and deliver exceptional video experiences to your customers.

Curious about how VPUs can seamlessly blend with your infrastructure, talk to NETINT engineers about scheduling consultation and start testing!

Picture of Anita Flejter

Anita Flejter

Is a Director of Digital Marketing at NETINT and a producer of Voices of Video bi-monthly event and podcast. She leads in digital marketing with a focus on innovative video tech, driving industry change through strategic campaigns and community engagement.

WATCH ON DEMAND: ASIC TRANSCODING FOR HIGH-VOLUME PLATFORMS

ACCESS NOW:  ASIC-Based Transcoding
for High-volume Use Cases
Including social media, broadcast, interactive platforms, and service providers