Your browser is not supported. Please update it.

5 September 2019

Start-up Eluvio changing the very fabric of video delivery

If we said P2P networking is destined to replace the traditional CDN, people would respond with “it’s been tried before”. If we proclaimed that a blockchain-based architecture could throw the transcoding industry into disarray, we would be laughed out of the building. But while disruption manifests itself in many different forms, few have intrigued us more recently than a start-up called Eluvio, emerging from stealth mode only today – with the potential to achieve all of the above and more.

“It would be arrogant to say our technology could replace this or that – native disruption is not the idea here,” said a modest Michelle Munson, co-founder of Eluvio, speaking to Faultline Online Reporter this week.

Eluvio’s vision is difficult to describe in a nutshell, so we can forgive it for putting the promise of CDN-less video delivery in the opening gambit of its debut press release. Ultimately, the Eluvio Content Fabric technology is a multi-layered architecture designed to eliminate the requirement to create additional copies of files used in distribution networks or storage facilities. It functions as an overlay network for video distribution directly from the stream or file source – implemented in a blockchain network to create a new kind of D2C distribution platform.

You might need to squint slightly, but just by scratching at the surface of the company’s Content Fabric software platform, in development for around two years, we can begin to envisage how video delivery networks in the future could look decidedly different. In doing so, it could disrupt more than your major third party CDN supplier – potentially commoditizing various roots of the video ecosystem, from ads and personalization to monetization and audience reporting.

Now let’s put the Content Fabric under the microscope. It comprises four core layers – Data, Code, Contract and Consumable Media. The process begins with raw metadata in the Data layer, where the fabric breaks down file and stream-based media into elements like binary data, metadata, blockchain ledger smart contracts and code. It stores and manages large form content in a decentralized architecture meaning scalability, minimal network traffic and eliminating the need for file copies.

The Code Layer then delivers media in one “ultra-low latency” pipeline, instantly rerouting content over the highest bandwidth paths with lowest latency – which is all figured out by some clever machine learning algorithms. Content is then protected from the infrastructure with encryption in the Contract Layer, controlling content access via a fabric blockchain ledger with software contracts for monetization.

Finally, the Media Platform Layer renders and delivers content on a per-request basis, triggering the composition of consumable media from elements in the Data Layer, as visualized in the attached graphic.

“Unlike a file-based world, the Content Fabric creates a representation upon ingest – creating shards which are spread throughout the nodes in the work. This is designed for scale,” added Munson.

We remained skeptical despite being impressed, pressing that surely for tier 1 content owners with mammoth on-demand catalogs and hundreds of live channels, a major third party CDN remains a more viable option. Munson half agreed, noting that her Content Fabric is dimensionally different from a technological perspective compared to a traditional CDN – but that doesn’t mean there won’t be long-term consequences.

Our curiosity needed more. We asked Munson if she agreed with the idea that the expansion of blockchain, despite its growing influence in the digital media economy, is restricted somewhat by the nature of networks – in the sense that non-cryptocurrency networks offer no financial reward for sealing blocks. “Certainly there has been ill-applied engineering in blockchain,” admitted Munson, again citing scalability along with benefits of decentralized data like flexible monetization and transparent user data sharing.

“Well isn’t this basically peer-to-peer networking?” we queried. Almost relieved we had asked the question, Munson admitted that Content Fabric shares a lot with P2P techniques, both leaning on content-centric routing and taking account of content to produce high-quality video as the network grows.

While we have heard plenty of noise lately about various technologies and protocols reducing the reliance on third party CDNs, taking a shot at the transcoding market is less common. Eluvio is assuring content owners they can reduce reliance on complex and costly transcoding services, as well as cloud storage provides and aggregators, through its A/V pipeline which transcodes, packages and delivers live and on-demand 4K video.

This brought to mind Livepeer, a start-up we came across over the summer talking about decentralized encoding. “I’ve got a lot of respect for those guys but this is not so much anti-CDN technology,” said Munson. Indeed, Livepeer is in a different game to Eluvio, targeting cryptocurrency miners which have acquired substantial numbers of GPUs to power their mining activities, although there is some decentralized overlap. Strangely enough, given Munson’s time at IBM, this concept was identified 50 years ago with the invention of virtualization by IBM sharing CPU resources between multiple workloads, admittedly not then decentralized which came some 15 years later with clusters developed by a now forgotten company called Digital Equipment.

For context regarding the scale of the fundamental problem identified by Eluvio, a typical program for digital distribution equates to 4,800 different versions, according to Eluvio, based on having 10+ viewer platforms, in 12 major languages, with 4 major DRMs, in 10 ABR resolutions. That’s significantly higher than Netflix, which we are told transcodes around 120 versions per title, but accounting for the abundance of legacy devices in the field elevates this number.

Eluvio can currently only talk publicly about one customer today, MGM, although Munson assured us there are more. She could only reveal that the US media giant is using Eluvio across “broad properties” and could not specify whether this included Epix or others.

The 22-person Eluvio was founded after Munson and fellow co-founder Serban Simu identified inefficiencies and costs within video delivery while heading up IBM’s Aspera unit where they invented the FASP (Fast and Secure Protocol) fast file transport protocol.

It’s easy to get overexcited when a technology presents itself using all the right buzzwords without sounding like a broken record. Still, of the vast majority of start-ups that never see the light of day, if Eluvio somehow doesn’t fulfill its potential then at least the team could say they inspired the next generation of genuine innovation in video delivery networks.