The Story of Blockast: How It All Started (Part 2)

The Explosive Growth of Streaming and Open Caching

The pandemic hit just a few months after I embarked on the journey to provide multicast services to the transportation industry, bringing the world to halt. Alongside the chaos and uncertainty, stay-at-home mandates forced more workers and students than ever to become dependent on the internet…and video streaming.

Bandwidth shortages were no longer limited to remote networks. The streaming wars had taken off and even the biggest media companies were not profitable. I joined the SVTA, the largest international technical association dedicated to solving the streaming video industry’s technical and infrastructure challenges. In the SVTA, I found the partners we needed to get multicast-enabled CDNs into communities. There industry leaders in SVTA had already been working on Open Caching, an open standard and architecture where streaming services, CDNs, ISPs, and individuals at home could participate and benefit in running their own content deliver infrastructure.

The Lightbulb Moment

Standard interconnection interfaces are essential to infrastructure, enabling multiple providers to work together as a single network. Banks rely on SWIFT to allow their customers to send money to each other. The 3GPP specification enables telcos to leverage each other’s towers and enable users to roam across them. The CDN interconnection interface used by Open Caching allows content owners to deploy their services into numerous CDNs using the same interfaces. These interfaces are highly restricted and locked down with layers of bureaucracy and firewalls, meaning new deployments and expanding coverage is slow.

Bitcoin enabled institutions and individuals to transact without the need for SWIFT and trust. The work we did with Magma simplified access to 3GPP interfaces so that networks like Helium could enable independent operators to deploy new cell towers on behalf of operators for a rev share in a trustless fashion. The same did not exist for CDNs and there was a lightbulb moment as I started to see how decentralization and an incentive layer were the missing links for CDNs based on Open Caching and multicast to proliferate.

Multicast: The Beating Heart of the Network

Although multicast has been around for decades, there probably hadn’t been a real need for it (which maybe accounts for how little deployment there has been) until online streaming took off. As more people stream today, the ISP middle mile is full of duplicated data streams that can’t be scaled with the current unicast CDN paradigm. Even as we add more edge compute to deduplicate streams, the middle mile itself becomes the bottleneck, particularly for internet providers using wireless backhaul. It’s a problem that is extremely concerning to network operators as demand for higher bitrate content like 4K and VR/XR grows and there is no way to meet it. That’s why in recent years, the telecom and broadcast industries have come together to develop hybrid broadband networks which can leverage the capacity in OTA broadcast (made available through datacasting channels in ATSC 3.0, eMBMS, and even satellite) to create more capacity. While there is little question this is where future networks are headed, progress has been slow as there isn’t the interconnection facilities and incentive mechanisms in place for all these independent network operators to work together on more efficient content delivery.

So I thought about that Helium model again. I loved how they built out a network faster and cheaper involving the community in the process which made me think, “how can I do that for content delivery?” I asked myself, “what if blockcast enabled users to host their own multicast-enabled nodes so that they could receive multicast streams and then stream that content to their neighbors?” And that’s when I knew I was onto something big because approaching content delivery in that way, kind of like peer-to-peer but with so much more control for content owners, would further take the burden off the ISP (so they could deliver even more high-quality content via multicast) and make use of all the unused home internet bandwidth that’s out there.

But the real ah-ha moment came when I connected this thinking with that Helium lightbulb moment I had before — we could save content owners money, by making content delivery less costly through multicast, and allow individual users to generate money by being part of the content delivery chain. The end-result could be a new kind of content delivery network with capacity from commercial CDNs, ISPs, broadcasters and, most importantly, end users.

The Missing Piece

So, I took all the knowledge I gained from working with Facebook’s traffic team, the experience from Magma, the lightbulb moment with Helium, the SVTA Open Caching specifications, and designed the Blockcast architecture.

But, Blockcast is more than just that architecture. It’s a business predicated on people operating parts of the network. Up until now, content delivery, which generates hundreds of millions in annual revenue, has been largely controlled by a few big companies. CDNs often terminate on ISP networks so the ISP, who is on the hook to deliver the content their subscriber is requesting, has to deliver the content through their network with no compensation. And capacity in those networks is an issue. That’s why streaming platform operators often use multiple CDNs — no single network is big enough to meet a global audience. CDNs don’t even peer with rural ISPs, because it’s not cost effective for the CDN to put physical hardware with an ISP that may only serve a few thousand people. Only in the U.S. those users served by rural ISPs represent about 45MM people who have gig access up to the ISP, but the ISP has limited bandwidth from their provider. It’s a mess.

How Lisa Joined This Adventure

At the same time, and you can call this serendipity or fate or kismet or whatever, my friend Lisa started telling me about different kinds of infrastructure networks that were built on crypto rails. When I thought about the positive reception Helium had, being able to build out 10x the number of nodes in one year than we did at Facebook in five, I knew the timing was right to bring zero-trust primitives to content delivery.

Lisa and I had been friends for 10 years. We met at a computer science class my senior year at Berkeley. We connected through our love for infrastructure and building bridges. She was looking to build bridges from the U.S. to China and I was looking to build bridges to the Middle East and the rest of the world. I was interested in telecom infrastructure, she was interested in financial infrastructure. We had very similar missions and the kind of impact we wanted to have in the world. And while she went on to work for Celo, a layer-1 blockchain focused on financial inclusion for the masses, I was banging my head against the wall on solving CDN challenges. Thankfully, Lisa and I, together, see the potential of using crypto primitives, not only incentivizing the supply side of the network, but also allowing individuals, ISPs, and CDNs around the world to come together to build a better, more performant CDN that could not only serve web3 traffic but meet the security requirements and SLAs to serve Hulu videos, Olympic Game live streams, Fortnite game patches, etc.

The Journey Continues

The result of zero-trust primitives integrated into how content is delivered over the internet would truly change content delivery. It could decentralize the business and spread it out across the ecosystem, enabling everyone, from ISP to DePIN operator to homeowner, to participate in the economics. Combined with multicast and caches deep into the network (in the actual homes where the content is being watched), the result would be a CDN massively more scalable and much closer to the end user. That last part has always been a big motivation for streaming operators because the closer the content is to the user, the better the experience. But getting really high-bitrate content there, like 4K, is the problem, especially when it’s all unicast. Giving people an opportunity to help “extend” content delivery, and earn rewards for being part of the delivery chain, would ultimately make a much more scalable “network”. But it’s really not about just building a better network by involving anybody who wants to host a caching/delivery node. It’s about making the act of delivering content better with everyone’s help by enabling content to get closer to the user and using multicast to reduce the unnecessary waste of capacity.