Click here to Skip to main content
65,938 articles
CodeProject is changing. Read more.
Articles
(untagged)

How to Reduce Your Data Center Footprint

14 Jun 2017 1  
Whether it's live video streaming, live and video on demand transcoding solutions, or virtual set-top boxes, you must position your business to transition to ubiquitous video demands (video for everyone, anywhere, on any device.)

This article is in the Product Showcase section for our sponsors at CodeProject. These articles are intended to provide you with information on products and services that we consider useful and of value to developers.

Click here for a free trial software download

The explosion of video streaming

Annual global IP traffic will surpass the zettabyte threshold this year and double that yearly rate by 20191. Having already increased more than fivefold just in the past 5 years, traffic is predicted to increase nearly threefold over the next 5 years. By 2019, there will be, on average, three networked devices per person.

If you are in the business of processing and streaming media in the cloud, dealing with this amount of content requires significant planning and new approaches to scale up hardware capabilities in order to meet demand. Whether it's live video streaming, live and video on demand transcoding solutions, or virtual set-top boxes, you must position your business to transition to ubiquitous video demands (video for everyone, anywhere, on any device.)

Huge streaming requirements drive Datacenter footprints to expand

Broadcasting companies and content aggregators are facing real challenges in keeping up with ever-increasing volume of digital media. To put this into perspective, let's take a look at some YouTube Statistics for streaming videos:

YouTube StatsData
Total number of people who use YouTube1,300,000,000
Hours of video uploaded to YouTube every minute300
Number of Videos viewed on YouTube everyday4,950,000,000
Total number of hours of video watched on YouTube each month3.25 billion

This data needs not only to be stored, but transcoded and streamed. According to Google, streaming just one minute of a YouTube video eats up 0.00002 kilowatt hours and generates 0.1 gram of carbon dioxide. Take that and multiply it by the 180 billion minutes just for YouTube in a month.

Datacenters, especially video transcoders and streaming media providers, cannot keep adding traditional servers with CPU-only processing and expect to stay competitive. The physical space required plus the power, and other operating costs beg for a better solution that can allow scaling within the current footprint.

This blog will discuss how 3 companies in the digital media industry addressed this overwhelming amount of video data and ended up with scalable, fast solutions while at the same time significantly reducing their datacenter footprints.

Company

Business

Vantrix Media PlatformSoftware defined solutions for live and on-demand video transcoding
iStreamPlanetLive Video Streaming of premium HD content via SaaS to practically any device
ActiveVideoDeveloper of CloudTV*, Pay-TV and Virtual Set-Top Boxes

Vantrix was looking for a solution that would not only reduce their datacenter footprint but also increase their processing capabilities. Something faster, more flexible and able to have larger addressable memory capacity that would be ideal for processing compute intensive functions including transcoding.

IStreamPlanet was running it's own Aventus private cloud on traditional blades but they needed to explore alternative infrastructure to improve the transcoding performance and efficiency in the cloud.

ActiveVideo came to the realization that CPU-only processing was resulting in poor performance that could not scale.They recognized that the cloud was becoming the hub for delivery of advanced user interfaces and online content.

Solutions

CompanyComponents
Vantrix2
  • HP Moonshot* 1500 Chassis loaded with HP ProLiant* m710 Servers
  • Intel® Media Studio Server
iStreamPlanet3
  • HP Moonshot* 1500 Chassis loaded with 5 HP ProLiant* m710 Servers
  • Intel Media Studio Server
ActiveVideo4
  • Kontron Symkloud* servers with Intel® Iris™ Pro Graphics
  • Intel Media Studio Server

The HP ProLiant* m710 Server is built on the Intel® Xeon® Processor E3-1284Lv3 with integrated Intel Iris Pro Graphics P5200 to accelerate high-quality video transcoding. Intel Iris Pro Graphics includes Intel Quick Sync Video which provides hardware acceleration as exposed by the Intel Media Studio Server.

The Kontron Symkloud* is a high density COTS 2RU-sized platform that uniquely leverages Intel® Core™ i7 processors equipped with Intel Iris Pro Graphics to massively scale video transcoding workloads and UI sessions across multiple clustered platforms.

Intel Iris Pro Graphics are the integrated graphics processor series introduced in 2013 with some models of Haswell processors as the high-performance versions of HD Graphics. Intel Iris Pro Graphics was the first in the series to incorporate embedded DRAM.

Intel Quick Sync Video is a hardware video encoding and decoding technology. It is integrated into some Intel CPUs. The name "Quick Sync" refers to the use case of quickly transcoding ("syncing") a video from, for example, a DVD or Blu-ray Disc to a format appropriate to, for example, a smartphone. Intel Quick Sync Video was introduced with the Sandy Bridge CPU microarchitecture on 9 January 2011.

All the above companies chose Intel processors with integrated Iris Pro Graphics P5200 equipped with Intel Quick Sync Video. They also chose to use the Intel Media Studio Server5 that bundles many tools to take advantage of hardware acceleration and optimization for the processing and streaming of digital media.

How Vantrix, iStreamPlanet and Active Video reduced their footprint and capital expenditures

All 3 companies experienced phenomenal results:

CompanyResults
Vantrix

Footprint and capital expenditure: Able to deliver more streams per rack unit while using less power and space.

Vantrix's experience using Intel Quick Sync Video:

  • Increased from transcoding 5 simultaneous 1080i streams to 14 streams per system
  • Increased what each rack can handle from 52 to 147 streams

This optimization resulted in greatly increased application performance, streamlined economics and accelerated business innovation.

iStreamPlanet
  • Bitrate conversions for multiple entire channels that used to require multiple larger processors are now processed on a single ProLiant* m710.
  • Reduced footprint as 9 HP Moonshot* 1500's fit in less space than a typical 42u rack
  • Reduced power consumption to a fraction of the energy required by blade servers
  • Increased number of HD channels while lowering the costs per channel
ActiveVideo
  • Replaced 15-20 racks with just 1 rack delivering TV guides to 1 million subscribers
  • Saw 10X increase in real-time video transcoding sessions
  • Up to 40% improvement in per session price performance due to the following:
    • Greater than 4X increase in number of CloudTV* sessions density, requiring just one datacenter rack to deliver cloud based pay-TV guides to 1 million subscribers (compared to 15-20 racks currently required) and
    • Greater than 10X increase in real-time video transcoding sessions to adapt premium online video content to any pay-TV subscribers.
  • Decreased the capital expenditure profile from $40-60 per subscriber to as low as $1 per subscriber

Summary

Companies are successfully finding innovative ways to address the rapidly increasing growth of IP Video while reducing footprint and costs. All the companies from the case-studies referenced above used the Intel Media Studio Server toolset in order to get the best possible performance running on their new, scalable servers, all of which were equipped with Intel Iris Pro Graphics.

Vantrix saw a 29x improvement beating the industry average7 of 5.54 streams per rack unit and now delivers up to 147 HD streams per rack unit. Their solution is based on the Intel GPU accelerated H.264 codec, taking advantage of Intel Quick Sync Video.

iStreamPlanet saved space and provided room for scaling using the HP Moonshot* 1500. They can now respond quickly to growing customer demands. adding channels or content in a matter of minutes.

For pay-TV operators, scalability in the cloud is a principal concern. ActiveVideo's solution was to use Kontron Symkloud* servers with Intel Iris Pro Graphics to reduce their capital and operating expenses while enhancing their user experience on virtual set-top boxes.

Resources

  1. Cisco Visual Networking Index: Forecast and Methodology, 2014-2019
  2. Vantrix Media Platform* on HP Moonshot* with HP ProLiant* m710p Server Cartridges
  3. Case Study: iStreamPlanet transforms live video streaming with HP Moonshot*
  4. ActiveVideo, Intel technology breakthroughs reduce Virtual Set-Top Box datacenter capex to $1/sub
  5. Intel® Media Server Studio 2016
  6. Get Fast Media Performance through Hardware Acceleration
  7. Global Media and Entertainment Video Transcoding Market

About the Author

Gael Hofemeier is an Senior Software Engineer enabling Business Client and Consumer applications and is also Technical Content writer

*Other names and brands may be claimed as the property of others.

License

This article has no explicit license attached to it but may contain usage terms in the article text or the download files themselves. If in doubt please contact the author via the discussion board below.

A list of licenses authors might use can be found here