[DAO: bafkrei] Self-Service Audio/Video Streaming

by 0x7ab021aca91ceb3c45086894b03dcfac38031529 (Jusdem#1529)

Should the following Tier 5: up to $120,000 USD, 6 months vesting (1 month cliff) grant in the Platform Contributor category be approved?

Abstract

Over the last year our team at Serraform has been fully committed to supporting the Decentraland ecosystem with a fully managed, enterprise grade, adaptive video streaming service. We’ve successfully worked through all the challenges of streaming into the Decentraland for events of all shapes and sizes, and are now ready to take it to the next level.

We believe that streaming audio and video into Decentraland should be as easy as posting on Facebook, and as flexible as Lego blocks. With your support, we can provide the “remote” you need to drive content at scale with no limits, no buffering, and no jitter.

We need a Decentraland-first streaming platform that enables any streaming need, and one that can grow with our community.

Grant size

120,000 USD

Beneficiary address

0xBb626c562D5b89961Eb2a2519e49d97D132189F7

Email address

justen@serraform.com

Description

Streaming limitless events with the best user experience requires global CDN networks with a massive amount of bandwidth that can scale in real-time; something most of us just don’t have access to. The answer is simply to harness the power of the cloud, efficiently and securely.

With a combined 30+ years of experience in cloud consulting with some of the largest fortune 500 companies through infrastructure, development, automation, and cybersecurity. Not to mention we’re a strong player within the community – Serraform is uniquely positioned to deliver a self-service video streaming platform for Decentraland that is fully-automated and simple to use.

With your support, we plan to further invest in the development of a streaming platform hosted in the cloud and secured with state-of-the-art cybersecurity and compliant solutions. To do this we need to take what we’ve learned and built so far and automate its ability to service you at a moment’s notice. We need the service to scale beyond what our team can manage and let anybody use it themselves when they want, while reducing overall running costs for the community.

We need to develop an interface and an API that can be used for any parcel owner to leverage the power of our streaming platform, allowing the ability for users to upload any audio/video file for streaming, or copy/paste in a Twitch/YouTube stream URL for those who prefer that route. We also know the community wants to broadcast Zoom podcasts or OBS streams directly into Decentraland but lack the ability to do it at scale and without jitter – we’ve been servicing all these use-cases, and now we need to package it up in a simple to use service for anybody to use on their own time. Simply upload your video, paste your link, or request a RTMP URL/key and let the fully automated service handle the rest.

If you’re a developer you might build an interface inside Decentraland or within your own tool if you prefer not to use ours, then leverage our API to deliver on the streaming magic for you. Rent out your venue and allow guests to control their content natively inside your own parcel. Have a club night where your VIP guests control the stream via Twitch or a jukebox collection of content over a stream deck you built! The possibilities will be endless, and this is just the beginning.

Feature highlights PowerPoint: DCL_Grant-Streaming.pptx - Google Slides

Specification

We plan to develop this solution on Microsoft Azure cloud using fully automated code to provision isolated streaming “tenants” using modern DevOps practices. We believe the solution should be cloud-native and have no single points of failure or need for manual intervention, while being able to deliver on your security and compliance needs with maximum horsepower and optimum user experience across the globe.

In addition to our proposed website UI, we plan to develop an API ‘under the hood’ to allow 3rd party app integration, allowing anybody to integrate their own parcel and control preferences with our streaming solution, either in-game or through 3rd party tooling or websites. Why not build your own jukebox collection and let party-goers control your screens within your parcel? Imagine building a stream deck inside your venue!

The content will be encoded with H.264 studio-grade quality up to 1080p, optimized for Decentraland events, and securely stored using Microsoft’s award-winning Defender security toolset. The streams will be provided as adaptive M3U8 streams, meaning various bitrates and resolutions will be available, accommodating any bandwidth limitations – this basically means you get a buffer-free experience on every stream at the best quality possible anywhere and on any device.

The service will support ingestion of any external stream (via OBS or Zoom for example) using an RTMP URL/key pair, but also be able to broadcast streams directly from Twitch and/or YouTube if you already have a URL.

We plan to leverage Verizon’s global premium CDN network allowing for dedicated content delivery to anybody around the world at very low latency and zero buffering. We want to be able to ingest any video source, and re-broadcast it directly into Decentraland, providing a great viewing experience for over 1700+ user events. No more skipping OBS streams for large audiences!

We currently offer this as a service, and we manage it ourselves on behalf of our customers. The goal now is to take your needs and our solution and develop a fully automated backend streaming platform that is self-serviceable by you via our website/API and doesn’t require our team to manage any aspect of it other than growing it based on your continued feedback. We want to scale this working solution and reduce the costs overall for the entire Decentraland community, while making it far more versatile for anybody to leverage.

Personnel

In addition to the below personnel, we plan to further engage with the Decentraland and blockchain talent pool for certain development aspects around our proposed solution.

  • Jusdem – Tech lead and product visionary
  • Paul K. – 12x Microsoft cloud MVP and cybersecurity guru
  • Prabh T. – Cloud services and operations expert
  • Aaron H. – Business and community development expert
  • Patmacs – Decentraland dev ninja and DCL Connect founder
  • ??? – UI/UX developer
  • ??? – Blockchain developer
  • ??? – Cloud DevOps and automation engineer

Roadmap and milestones

Communication Plan: Decentraland forum with monthly status updates.

Estimated Timeline: 3-6 months

  • Month 1 - Initiate
    • Finalized production design
    • Personnel identified and hired
    • Dev work started
  • Month 2 - Automated VOD
    • Automated provisioning of streaming “tenants”
    • Pre-recorded (eg. MP4) file uploads
    • Audio and video encoding (H.264)
      • Makes playback suitable for variety of devices
    • Up to 1 year stream retention/storage
      • Playback any stream for up to a year
    • Unlimited number of audio/video assets
      • Build a library of your favorite streams
  • Month 3 - Automated Live
    • Live (RTMP) stream ingestion
      • Broadcast your own stream from anywhere
      • Works with OBS and Zoom!
    • Live stream recording
      • Record and playback anytime
    • 24/7 stream availability with bandwidth scaling
      • No more limits on audience sizes
    • Audio/video output in M3U8 format
      • Supported by Decentraland
      • Adaptive multi-bitrate
  • Month 4/5 - API Controller
    • Twitch and YouTube streaming capabilities
      • Auto-provisioned upon request
    • API backend
    • Secure data at rest and in transit
  • Month 5/6 - Frontend Controller
    • Website front-end UI
      • Metamask integration
      • Fully self-service and automated
    • Pricing and Payment Methods

Vote on this proposal on the Decentraland DAO

View this proposal on Snapshot

admittedly i don’t know much about streaming video - can you say how your proposal is different/better than this recent proposal which is half the cost?

3 Likes

Great question, and I’m happy to see this other proposal in the pipe!

In short, the other proposal focuses on making the “presentation” of video faster with less latency. Our proposal focuses on making the “processing and delivery” of the stream powerful, effective, easy, and super cheap to do.

In more detail, our proposal is different from the one you mentioned. The mentioned one is actually improving the Decentraland components themselves by adding WebRTC ultra low latency components to the DCL SDK, which also includes modifications to the Decentraland Unity Renderer and Kernel. This is all core to Decentraland itself, and actually compliments ours!

Our proposal differs in that we’re providing the streaming “server” hosted in the cloud, and making it fully self-service and fully automated for anybody to use from a website (or our API) and pay with Metamask. Our service plugs into Decentraland, and the other proposal makes that “plug” even better - everybody wins!

End of the day our streaming “server” will host a ton of super advanced, enterprise grade components, with a global CDN network that spans the globe and make streaming not only powerful and flexible, but super cheap and easy for anybody to use.

This is really needed and sounds like a lot of great features that will really simplify streaming for land owners.

However, there are existing services that already do this and the main barrier of entry is cost. Utilizing verizon’s premium network doesn’t sound cheap. Can you provide how much you expect your service to cost to fully utilize all CDN and VOD capabilities?

2 Likes

Hi @dax , not from the team but can clarify here. The other proposal is about updating the protocol used for the “last mile” of delivery of the video data. Currently, the video sdk component uses an http, H.264 protocol (which would be used in this proposal). This basically is a way to get chunks of the video data every 5-10 seconds instead of with a constant stream. The other proposal intends to create an sdk component that is capable of handling video feed data in RTMP format, which is a continuous stream of data so there is less latency.

This proposal would provide a managed service for land owners to not have to configure their own live streaming server or use an existing managed streaming service, along with a lot of enhancements like the CDN and VOD features.

3 Likes

Is it going to be Open Source? Can I self-host?

Thanks.

4 Likes

Looks like my reply was temporarily removed, so I’ll keep it brief assuming it will come back :slight_smile: plus @mattimus had a great response.

In short, the other proposal is making changes to the Decentraland components in order to make the “presentation” layer of streams more efficient.

Our proposal is focusing on the “delivery” of streams, making it super simple and efficient to stream to DCL in a variety of ways.

1 Like

It will not be open sourced per say. What we’re building couldn’t be self-hosted either, as we need to use the cloud to host it; in our case Microsoft Azure has the horsepower we need to do this at scale.

With that said we’d be happy to provide any templates we build to deploy it to Azure, but the logic and structure of auto-provisioning all of it would still have to be built out using DevOps. Not ideal to do this on your own, for most people anyways.

1 Like

Thanks for the question!

Essentially by premium we mean not Microsoft or Akamai network, but Verizon’s network seems to work the best for global Decentraland events. This to most people is considered “premium”.

As for cost, we’re building within Microsoft Azure and therefore are able to get very close to cost pricing for that network and we intend to share those savings with the community that leverages our service. We only intend to charge a small service fee for our maintenance and support of the system.

More details on Verizon (S1) CDN pricing here: Pricing - Content Delivery Network (CDN) | Microsoft Azure

Additionally, if all goes as planned and more people start to leverage the platform, costs should go down more based on volume of consumption. The more we use, the less Microsoft charges us, and we pass those savings down to the community.

As for VOD pricing it’s a bit tricky to nail down as there’s so many other fees involved and depending on our final design that can change/go lower. However, our intention is to build a user friendly self-service streaming platform using Azure and pass all savings down to the community. The more we build for and use at scale, the cheaper this becomes for all.

1 Like

People can use livepeer.studio, it doesn’t need 120k to work, it’s easy to use and it works with DCL

2 Likes

Great point, however we’ve worked with Livepeer on a few occasions and although it’s great and so are the folks behind it, the network struggles to deliver at scale for large events and it doesn’t include all the same features we’re hoping to build. We had many 404 and 503 errors during some events, and to use some of its features you need to interact with its API so it’s not very intuitive for most when it come to simple VOD playback, for example. Not to mention, our service has features and functionality that are unique and tailored to Decentraland, and integrated with DCL Connect for parcel management; LivePeer doesn’t have this.

We plan to continue working with Livepeer to iron out these issues, but in the meantime to deliver successful events, our customers prefer stability, reliability, and scale. In the future when it stabilizes, we would love to port the service over to Livepeer! We have a slew of things in the pipe for this service, and we’re just getting started :slight_smile:

A cursory search shows me that the name “Serraform” and Decentraland returns only one result:

Visiting your website I get this:

ICANN Lookup shows me the domain name for Serraform.com, the one listed as the email address for this proposal, was purchased just four months ago:

I recognize that video streaming in Decentraland has room for improvement. I don’t know all of the technical ins and outs of how to achieve this. What I do know is that the claims being fully committed to DCL’s ecosystem for the last year don’t check out.

I see no budget details. How will this $120,000 be spent? How much of it goes to pay Full Time Equivalents for programming? How much for servers? Aside from first names where can I see the CVs for the people listed by first name only. Does your team have the skills and experience needed? Currently there’s no way for me to find out.

What I can see is an ask of $120,000 for something that, as @HPrivakos has pointed out, already has an existing solution and won’t require us to spend this additional money. A metric I like to use lately is the Vroomway proposal. That was for $60,000 of money from this community and already it has proved its worth while still in Alpha. Will this proposal return two Vroomway’s worth of value? Doesn’t look like it to me.

Show me examples of your team successfully working through streaming challenges. How big were these events? Where were they held in Decentraland?

Lastly, this is the third proposal I’ve seen from MetaArchs. I don’t put an at before their name because they aren’t in the forum to even discuss this proposal. With 3.8M VP they’ve effectively been able to propose and vote for $240,000 worth of funding originating from two separate grant proposals. Their votes for passage were soon after the proposals posted, before any public debate. They’ve done the same a pending POI proposal.

As an active member of this community for over eight months, who is connected to many of Decentraland’s daily active users, I will be voting no for this proposal.

1 Like

How convient having MetaArchs vote on all these grants. This just looks like we pay you to build a platform.Then everyone using this will have to pay weekly or monthly i assume.

1 Like

I had chosen to not vote right away on this proposal as I really didn’t understand it or why, and wanted to see how the community would respond to it before I voted. After careful consideration and reading through all the community thoughts, on the forums and around various discords, I have VOTED NO ON THIS PROPOSAL.

From what I am seeing over the last week or two, there is a gross abuse of power. An abuse of power against the founding purpose and ideals of the DAO and decentralized platform, and what we are all working towards.

Do not think this is going unnoticed.

2 Likes