Large Grant: Fabstir Web3 Media Player #3

Project Name: Fabstir Web3 Media Player #3

Organisation: Fabstir

Primary Contact: Jules Lai (Founder and CTO of Fabstir)

Summary:

The one-year grant period focuses on bringing real-world adoption to Fabstir Media Player whilst helping to ensure S5 scales effectively with production use cases. The primary objectives are to increase adoption of Sia.tech’s decentralised storage and develop additional software integrations for Sia within Web3.
The Fabstir Media Player will undergo a comprehensive security audit by an approved auditor for listing in the MetaMask Snaps Directory, potentially exposing the player and ecosystem to MetaMask’s 30+ million monthly active users.

Previous grant proposals #1 and #2

Objectives:

OrbitDB Storage Adapter

Fabstir Media Player uses FabstirDB; a general purpose decentralised graph database. This uses OrbitDB and ipfs. The OrbitDB storage adapter will allow Fabstir’s database to be instead stored on Sia.

  • Develop S5 storage adapter for OrbitDB that:

    • Maps user keys to S5 CIDs via S5 streams
    • Enables peer-to-peer sync of key-value mappings
    • Implements core OrbitDB operations:
      • put, get, del operations
      • iterator functionality
      • merge operations
      • clear operations
    • Handles conflict resolution using S5 streams’ revision numbering
    • Maintains data consistency across peers
    • Implements radix trees for efficient:
      • Prefix path lookups
      • Range queries
      • Key-value store operations

S5 Network Scalability Testing & Integration

The initial phase begins with the Non-Multiplex Cinema short film festival. I am the founder and will be organising and hosting the events in London’s Soho district. Filmmakers will utilise Fabstir Media Player to upload their submissions, whilst our selected panel of judges will curate the final festival programme. This event provides an excellent opportunity to test both the robustness of the media player and S5’s CDN network under real-world conditions.
In the latter half of the year, we will expand to support full-length feature films up to 4K resolution, with several US distribution companies already committed to using Fabstir once launched. A big test will come when Fabstir Media Player gains approval to be listed on MetaMask’s Snaps directory with the potential of a flood of new users.

  • Implement comprehensive testing framework for S5 network that:

    • Develops load testing infrastructure:
      • Simulates concurrent video uploads
      • Tests multiple resolution streaming
      • Measures network latency under load
      • Monitors CID retrieval performance
    • Implements analytics module into Fabstir Media Player:
      • Tracks node performance utilisation
      • Monitors media formats streamed for size and bitrate
      • Analyses storage sizes and distribution
      • Logging system that includes network health statistics
    • Host S5 node:
      • Learn from S5’s isolated node for Fabstir
      • Deploy Fabstir’s hosted S5 node
      • Measure network latency under load
      • Host Sia node to enhance reliability for efficient backup and restoration

AI Search Engine Backup/Sync

AI search engine has been developed and is operational. It can use different open-source LLM models from Facebook’s Llama to Mistral AI. The system enables users to selectively authorise which of their assets stored in Fabstir’s decentralised graph database can be indexed for search capabilities - ranging from media content to various forms of structured data. This granular control ensures user privacy whilst maximising the utility of semantic search across Fabstir’s network. The task here is to decentralise the vector embeddings and store them on Sia, using S5 to sync the data across Kubernetes pods.

  • Develop S5-based distributed system that:

    • Maintains vector embeddings across Kubernetes pods:
      • Enables pod failover
      • Provides load balancing
      • Syncs via S5 streams
    • Implements incremental backups to Sia:
      • LLaMA/Mistral AI embeddings
      • Elasticsearch indices
      • Preserves semantic search capability
    • Handles pod synchronisation:
      • Uses S5’s distributed architecture
      • Maintains vector similarity calculations
      • Preserves index optimisations
    • Implements versioning system:
      • Uses S5 streams for manifest tracking
      • Enables differential updates
      • Allows selective data fetching by pods

Security Audits & Implementation

Sayfer, a leading Web3 cybersecurity company, brings substantial expertise to auditing both smart contracts and MetaMask Snaps. Their team includes senior security researchers with proven track records in blockchain security, including notable achievements like discovering the Badreveal exploit that affected 10% of NFT projects. With over 100 successful audits across various chains and ecosystems, their comprehensive audit package will evaluate Fabstir’s smart contracts and MetaMask Snaps against the Smart Contract Security Verification Standard (SCSVSv2). This standard is accepted by major regulations and embraced by industry leaders including the Ethereum Foundation. Their expertise in MetaMask Snaps security assessment is particularly valuable for meeting Consensys’s marketplace requirements, ensuring Fabstir Media Player can be safely distributed to MetaMask’s 30+ million users.

  • Execute comprehensive security enhancement that:

    • Implements smart contract improvements:
      • Strengthens NFT and ABT token factory contracts
      • Secures marketplace operations
      • Enhances micropayment systems
      • Fortifies RMRK nestable implementations
    • Enhances MetaMask Snap security:
      • Implements Consensys marketplace requirements
      • Strengthens state management and storage
      • Improves key handling protocols
      • Updates permission boundaries
    • Develops monitoring and maintenance:
      • Creates security monitoring dashboard
      • Implements automated testing framework
      • Establishes update protocols
      • Maintains compliance documentation
    • Implements post-audit system:
      • Tracks contract interactions
      • Monitors transaction patterns
      • Updates security guidelines
      • Provides deployment checklists

Budget:

The grant proposal is for $127,940 to support the project’s open-source development for 12 months. The budget allocation is as follows:

Item Cost ($)
Jules Lai to be full-stack Web3 developer 85,000
UI developer and designer 4,000
Devops engineer and support 5,000
AI developer 2,000
Security Audits & Implementation 14,000
GPU & CPU cloud 11,940*
Hardware/software 3,000*
Legal/accounting 1,500
Contingency and Miscellaneous 1,500
Total 127,940

Breakdown of GPU & CPU cloud cost

*Preliminary quotes have been gathered from Vultr for GPU cloud, Cloudflare for S5 node host and Hetzner for Sia node hosting.

Item Calculation Total
GPU cloud NVIDIA A16 GPU $350/mth * 6 $2100
GPU cloud NVIDIA L40S GPU $1220/mth * 6 $7320
CPU cloud Cloudflare Business $250/mth * 7 $1750
CPU cloud Hetzner EX101 $100/mth * 7 $770

Hosted NVIDIA GPU servers have to use professional range of GPUs. AV1 video encoding that has better compression will be used with Ada Lovelace architecture for last six months to facilitate scaling to up to full length movies at 4K resolution.

Breakdown of Hardware/Software Budget

Item Purpose
Developer laptop (4K capable) Mobile development/testing
32" 4K monitor Enhanced development workspace
NAS with RAID configuration Secure backup storage
2TB NVMe SSD Development environment
Software licences:
- Cloud backup service Offsite backup redundancy
- Development IDE subscriptions Enhanced coding tools
- Security/antivirus suite System protection
- Monitoring tools System performance tracking

Milestones and Timeline

Q1 (Months 1-3): Foundation Development

  • Month 1:

    • Complete OrbitDB adapter architecture design
    • Implement basic S5 streams integration
    • Set up initial testing framework for S5 network
    • Implement core OrbitDB operations with S5
  • Month 2:

    • Complete OrbitDB-S5 conflict resolution
    • Begin smart contract security improvements
    • Complete radix tree implementation
  • Month 3:

    • Finish initial security audit phase
    • Begin MetaMask Snap security enhancements
    • Optimise OrbitDB performance

Q2 (Months 4-6): Integration & Testing

  • Month 4:

    • Deploy load testing infrastructure
    • Implement analytics monitoring
    • Scale testing with concurrent users
    • Improve backup/restore mechanisms
    • Implement feedback from security audits
  • Month 5:

    • Complete MetaMask Snap marketplace requirements
    • Switch from Base Sepolia testnet to Base Mainnet
    • Implement advanced analytics features
    • Complete integration testing phase
    • Implement feedback and fixes from real-world short film festival
  • Month 6:

    • Set up hosted Sia node
    • Set up hosted S5 node
    • Finalise security monitoring dashboard
    • Real-world deployment testing for movie-length 4K

Q3 (Months 7-9): Optimisation & Scaling

  • Month 7:

    • Complete production deployment for L40S cloud infrastructure
    • Complete security post-audit implementation
    • Fine-tune pod load balancing
    • Optimise streaming performance
    • Deploy production monitoring systems
  • Month 8:

    • Deploy AI search engine backup system
    • Implement AI vector embeddings sync
    • Integrate versioning system for AI backups
    • Enhance AI search pod synchronisation
  • Month 9:

    • Help and support for film industry participants
    • Test full system under load
    • Implement feedback and fix any issues

Q4 (Months 10-12): Production & Documentation

  • Month 10:

    • Start comprehensive documentation
    • Finalise all security implementations
  • Month 11:

    • Complete all documentation
    • Address any remaining issues
    • Final system optimisation
  • Month 12:

    • Overall maintenance, scaling and security tasks

Future Benefits to Sia

  • Enhances Sia’s Web3 ecosystem integration:

    • Implements production OrbitDB adapter for Sia storage
    • Provides enterprise-grade AI search capabilities
    • Demonstrates real-world scalability through film festival platform
    • Positions Sia as a viable alternative to IPFS
  • Expands Sia’s market presence:

    • Strengthens position in the media streaming sector
    • Opens new use cases for professional content creators
    • Provides benchmarks for enterprise-level storage solutions
    • Attracts Web3 developers through documented integrations
  • Increases adoption through:

    • Integration with MetaMask’s 30M+ monthly active users
    • Professional filmmaker community engagement
    • Enterprise-ready distributed database solutions
    • Open-source tools for Sia integration
  • Benefits from Fabstir ecosystem (not part of grant):

    • Fabstir’s Non-Multiplex Cinema film platform brings offchain users to onchain
    • Fabstir’s decentralised subscription models bring the entertainment industry to the Media Player
    • Fabstir’s customisable media streaming platforms enable third parties to bring further adoption to Sia

Risks

  • Technical risks:

    • S5 network performance under concurrent load is untested at scale
    • Integration complexity between OrbitDB and S5 streams
    • MetaMask Snaps marketplace approval timeline uncertainties
    • AI model integration complexity with distributed storage
  • Operational risks:

    • GPU cloud costs may fluctuate based on market demand
    • Resource allocation for concurrent transcoding loads
    • Timeline dependencies on third-party security audits
    • Initial user adoption rate uncertainties
  • Mitigation strategies:

    • Comprehensive testing framework implementation
    • Staged rollout of features with monitoring
    • Regular communication with the S5 author Redsolver
    • Flexible architecture allowing for rapid adjustments

Hi Jules, please make sure your proposal confirms to the formatting found here.

One thing we specifically need are your answers to the residency and bank account questions, but also try to make sure you hit all the questions and sections as presented in the template at the above link.

In addition to the above, the Committee usually avoids funding grants that put hardware-related expenses in their budget. Better remove these items, please.

Okay, will do that tonight.

Okay, will do that then.

Large Grant Proposal

Introduction

Project Name: Fabstir Web3 Media Player #3

Name of the organisation submitting the proposal: Fabstir

Describe your project:

Fabstir Media Player is entering a new phase focused on bringing real-world adoption whilst helping to ensure S5 scales effectively with production use cases. The primary objectives are to increase adoption of Sia.tech’s decentralised storage and develop additional software integrations for Sia within Web3.

The project will:

  • Develop an OrbitDB storage adapter for Sia to enable decentralised graph database capabilities
  • Implement comprehensive S5 network scalability testing and integration
  • Create backup/sync functionality using S5 for transcoder and AI
  • Undergo security audits and implementation for MetaMask Snaps marketplace listing

Previous grant proposals: #1 and #2

Who benefits from your project?

  1. Sia ecosystem:

    • Expands Sia’s presence in Web3 through production-ready OrbitDB integration
    • Provides proven scalability testing of S5 network with real-world video streaming use cases
    • Demonstrates Sia’s capability to handle enterprise-grade workloads through film festival platform
    • Positions Sia as viable alternative to IPFS, particularly for media streaming where IPFS struggles
    • Introduces Sia to MetaMask’s 30+ million monthly active users through Snaps marketplace
    • Brings professional filmmakers and content creators to Sia for transcoding and archiving content
    • Create reusable open-source tools for Sia integration that other developers can leverage
    • Provides documented performance benchmarks for enterprise-level storage solutions
    • Expands Sia’s utility through AI search capabilities built on its storage
    • Attract NFT marketplace users through support for movie-length video NFTs
    • Helps establish Sia in the entertainment industry through support of multiple video formats
    • Creates a bridge between traditional content platforms and Web3 through Sia storage
    • Increases Sia network usage through peer-to-peer video streaming
    • Demonstrate Sia’s cost-effectiveness compared to traditional cloud storage for media
  2. Content creators and distributors:

    • Professional filmmakers and music creators gain decentralised distribution
    • Automated transcoding to multiple formats/resolutions
    • Content distribution companies can leverage Web3 capabilities
    • Content creators get enhanced ownership controls
    • Integration with existing NFT marketplaces
    • Access to decentralised subscription marketplaces for permissionless access
    • End-to-end encryption and cost effective redundancy backup options
  3. Web3 developers:

    • Access to documented Sia integrations
    • Open-source tools for building on Sia
    • Enterprise-ready distributed database solutions
    • MetaMask Snaps integration examples
  4. End users:

    • Enhanced media streaming capabilities
    • Improved content discovery through AI search
    • Better data ownership and control
    • Simple onboarding to Web3 through social login (Gmail, Git, Discord, etc.)
    • Cross-device access to content library
    • Transparent content ownership through NFTs without technical barrier

How does the project serve the Foundation’s mission of user-owned data?

The project advances user-owned data by:

  1. Decentralised Storage:

    • Stores media content on Sia network
    • Enterprise-grade graph database storage of data to Sia storage
    • Enables user control of encryption keys
    • Provides redundant storage across nodes
    • Allows direct content retrieval via advanced indexing and S5
  2. Data Sovereignty:

    • Selective data sharing permissions
    • Private key-based access control
    • Control of permissionless and permissioned access to markets
    • Self-hosted node options
    • Users maintain control of their AI search indices
  3. Open Standards:

    • Implements NFT standards for ownership
    • Uses open-source storage protocols
    • Provides transparent data formats
    • Enables data portability

Are you a resident of any jurisdiction on that list? No

Will your payment bank account be located in any jurisdiction on that list? No

Grant Specifics

Amount of money requested and justification with a comprehensive breakdown of expenses:

Total requested: $124,940 for 12 months

Detailed breakdown:

Item Cost ($) Justification
Full-stack Web3 developer 85,000 Primary development of all components
UI developer and designer 4,000 Interface design and implementation
Devops engineer and support 5,000 Cloud infrastructure management
AI developer 2,000 AI integration specialist
Security Audits & Implementation 14,000 Professional security auditing
GPU & CPU cloud 11,940 Infrastructure costs detailed below
Software licences & tools 1,000 Development tools and services
Legal/accounting 1,000 Compliance and financial management
Contingency 1,000 Buffer for unexpected costs

GPU & CPU Cloud Cost Breakdown:

  • GPU cloud (NVIDIA A16): $350/month × 6 = $2,100
  • GPU cloud (NVIDIA L40S): $1,220/month × 6 = $7,320
  • Cloudflare Business: $250/month × 7 = $1,750
  • Hetzner EX101: $100/month × 7 = $770

Timeline with measurable objectives and goals:

Q1 (Months 1-3): Foundation Development

Month 1:

  • Complete OrbitDB adapter architecture design
  • Implement basic S5 streams integration
  • Set up initial testing framework for S5 network
  • Implement core OrbitDB operations with S5

Month 2:

  • Complete OrbitDB-S5 conflict resolution
  • Begin smart contract security improvements
  • Complete radix tree implementation

Month 3:

  • Finish initial security audit phase
  • Begin MetaMask Snap security enhancements
  • Optimise OrbitDB performance

Q2 (Months 4-6): Integration & Testing

Month 4:

  • Deploy load testing infrastructure
  • Implement analytics monitoring
  • Scale testing with concurrent users
  • Improve backup/restore mechanisms
  • Implement feedback from security audits

Month 5:

  • Complete MetaMask Snap marketplace requirements
  • Switch from Base Sepolia testnet to Base Mainnet
  • Implement advanced analytics features
  • Complete integration testing phase
  • Implement feedback and fixes from real-world short film festival

Month 6:

  • Set up hosted Sia node
  • Set up hosted S5 node
  • Finalise security monitoring dashboard
  • Real-world deployment testing for movie-length 4K

Q3 (Months 7-9): Optimisation & Scaling

Month 7:

  • Complete production deployment for L40S cloud infrastructure
  • Complete security post-audit implementation
  • Fine-tune pod load balancing
  • Optimise streaming performance
  • Deploy production monitoring systems

Month 8:

  • Deploy AI search engine backup system
  • Implement AI vector embeddings sync
  • Integrate versioning system for AI backups
  • Enhance AI search pod synchronisation

Month 9:

  • Help and support for film industry participants
  • Test full system under load
  • Implement feedback and fix any issues

Q4 (Months 10-12): Production & Documentation

Month 10:

  • Start comprehensive documentation
  • Finalise all security implementations

Month 11:

  • Complete all documentation
  • Address any remaining issues
  • Final system optimisation

Month 12:

  • Overall maintenance, scaling and security tasks

Potential risks that will affect the outcome of the project:

Technical Risks:

  • S5 network performance under concurrent load is untested at scale
  • Integration complexity between OrbitDB and S5 streams
  • MetaMask Snaps marketplace approval timeline uncertainties
  • AI model integration complexity with distributed storage

Operational Risks:

  • GPU cloud costs may fluctuate based on market demand
  • Resource allocation for concurrent transcoding loads
  • Timeline dependencies on third-party security audits
  • Initial user adoption rate uncertainties

Mitigation Strategies:

  • Comprehensive testing framework implementation
  • Staged rollout of features with monitoring
  • Regular communication with S5 author Redsolver
  • Flexible architecture allowing rapid adjustments

Development Information

Will all of your project’s code be open-source?

Yes, all code developed for this grant will be open-source and available on GitHub under MIT licence. The project uses some closed-source components (ffmpeg for transcoding, MetaMask for wallet integration) but all code developed will be open-source.

Leave a link where code will be accessible for review:

Do you agree to submit monthly progress reports?

Yes, I agree to submit monthly progress reports in the forum following the required format.

Do you agree to designate a point of contact for committee questions and concerns?

Yes. Jules Lai (Founder and CTO of Fabstir) will be the primary point of contact.

Provide links to previous work or code from all team members:

Previous Sia grant work:

Monthly progress reports:

Have you developed a proof of concept for this idea already?

Yes. The Fabstir Media Player is already operational and was developed under previous Sia grants. This proposal builds upon that foundation to scale the implementation and add new capabilities. Previous work can be seen in the monthly progress reports and GitHub repositories linked above.

Do you agree to participate in a demo at our monthly community call at significant milestones or after the grant’s completion?

Yes, I agree to participate in a demo at our monthly community call at significant milestones or after the grant’s completion.

Contact Info

Email: jlai[at]fabstir.com

Any other preferred contact methods:
Telegram: Telegram: Contact @JulesL78