OST Founders CEO Jason Goldberg & CBS Ben Bollen present OpenST Mosaic Protocol at the Upbit Developer Conference 2018 -- 13 September, 2018, Jeju South Korea. Mosaic is presented in context of Ethereum, Ethereum 2.0, Casper, Sharding, and Plasma. OpenST Mosaic is a consensus protocol to run meta-blockchains on top of Ethereum.
 Each meta-blockchain runs in parallel and injects an additional state space into Ethereum. Each meta-blockchain is secured by Ethereum with an open, staked validator set. All transactions on meta-blockchains are asynchronously finalised and
committed onto Ethereum.

1. OpenST Mosaic
@benjaminbollen
Scaling Blockchain Economies to
Billions of Users
@betashop ostmosaic

2. Scaling Blockchain Economies
1. Blockchain economies in historical context
2. Barriers to mass adoption
3. OpenST Mosaic Protocol for scaling blockchain economies
on-chain to billions of users
ostmosaic

3. 1439

4. 1993

5. Blockchain is a Continuation of the Internet Revolution ostmosaic
INTERNET
ANALOG
▼
DIGITAL
ONE-TO-MANY
▼
MANY-TO-MANY
KNOWLEDGE
World Wide Web (1989)
RELATIONSHIPS
Social Media (2004)
Blockchain (2009)
Smart Contracts (2015)
MONEY
Cryptocurrency (2009)
Tokens (ERC-20: 2017)
INSTITUTIONS
DApps
DAOs
Credit: Mark Bonchek, Shift Thinking, OST Advisor

6. Get everyone
and everything
online
Photo by Rawpixel on Unsplash

7. Put a Super-Computer
in Everyone’s Pockets
Photo by rawpixel on Unsplash

8. Redefine Relationships
Photo by Tim Bennett on Unsplash

9. Redefine Money
Photo by Yohann Lc on Unsplash

10. Redefine Institutions
Photo by Yohann Lc on Unsplash

11. Create New Economies
Photo by Anete Lusina on Unsplash

12. Still Early Days of the Revolution
Users in Millions
Less Than 0.6% of Internet Users Have A Crypto Wallet

13. 3 Factors Limiting Adoption:
Developers, Scalability, UsabilityPhoto by Alex Radelich on Unsplash

14. Bridging the Scalability Gap
For Apps & DApps
Photo by JonasVerstuyft on Unsplash

15. Why Ethereum?
Programmable money.
It (proof of) works, today.
Strongest developer community.

16. Problem We Are Trying to Solve
Existing Mainstream Apps
with Millions of Users
Blockchains don’t scale
DApps Same problem
• Provide an on-ramp for existing apps with tens of millions of users to transform their
businesses with blockchain economies, now.
• Enable scalability for emerging DApps, now.

17. September 2017
OpenST Protocol
Staking value tokens on Ethereum for
minting utility tokens on auxiliary chains
Photo by Joshua Earle on Unsplash

18. September 2017
OpenST Protocol
Staking value tokens on Ethereum for
minting utility tokens on auxiliary chains
Move bulk transactions off Ethereum
(but keep them on-chain)
Photo by Joshua Earle on Unsplash

19. September 2017
OpenST Protocol
Staking value tokens on Ethereum for
minting utility tokens on auxiliary chains
Retain ETH smart contract framework
Photo by Joshua Earle on Unsplash

20. September 2017
OpenST Protocol
Staking value tokens on Ethereum for
minting utility tokens on auxiliary chains
Blockchain Economies for
Non-Blockchain Companies
Photo by Joshua Earle on Unsplash

21. September 2017
OpenST Protocol
Staking value tokens on Ethereum for
minting utility tokens on auxiliary chains
Scale DApps Now
Photo by Joshua Earle on Unsplash

22. September 2018
OpenST Mosaic Protocol
Running meta-blockchains on
Ethereum to scale DApps to
billions of users
Photo by Joshuan Earle on Unsplash

23. September 2018
OpenST Mosaic Protocol
Running meta-blockchains on
Ethereum to scale DApps to
billions of users
Photo by Joshuan Earle on Unsplash
Finalize transactions on aux chains
Commit transactions asynchronously on ETH

24. Limitation to
Scaling Blockchains
Photo by Andre Mouton on Unsplash
Full Replication

25. Three Ingredients for Sharding ostmosaic
1. Divide work into shards
2.Communicate between shards
3.Economic finality on each shard

26. OpenST Mosaic in Industry Context
•Plasma = scale each app off-chain; data-availability problem
•Payment channels don’t have smart contracts
•Casper + Sharding = Ethereum v2.0
•Mosaic = Sharding at Layer 2, on-chain, finalised on ETHv1.0
ostmosaic

27. OpenST Mosaic ostmosaic
•OpenST Mosaic is a consensus protocol to run meta-blockchains on top
of Ethereum.
•Each meta-blockchain runs in parallel and injects an additional state
space into Ethereum.
•Each meta-blockchain is secured by Ethereum with an open, staked
validator set.
•All transactions on meta-blockchains are asynchronously finalised and
committed onto Ethereum.
•Mosaic has a message-passing protocol between Ethereum and meta-
blockchains (later directly between meta-blockchains).

28. Simple Calculation (only took me 4 years)
tx
s
∼ 𝒪
(
N
Δt)
tx
s
∼ 𝒪(αN)
Ethereum @ 10 tx/s
simplest toy model:
100x (AUX @ 100 tx/s) = 10.000 tx/s
100 nodes / AUX * 100 AUX * 1 tx / node => 10.000 tx to commit on Ethereum
@10 tx/s takes minimally 1000s on Ethereum at full capacity
but 10.000.000 tx processed on aux; or 99.90% efficient
but we can do even better:
for N nodes, and constant #nodes/AUX
if there is no time-constraint to commit to Ethereum
cost: benefit:
e.g. commit once daily, to finalise 864 millions transactions, for 1.2% of
Ethereum’s daily capacity!

29. Move Bulk of Transactions Off Ethereum & Onto Auxiliary Chains!
Ethereum
Mainnet
Auxiliary
Chains

30. Move Bulk of Transactions Off Ethereum & Onto Auxiliary Chains!
Ethereum
Mainnet
Hold on, sidechains have been around for
years! What’s new here?
Auxiliary
Chains

31. Move Bulk of Transactions Off Ethereum & Onto Auxiliary Chains!
Ethereum
Mainnet
Auxiliary
Chains
A meta-blockchain is fully defined with a consensus engine in a core contract on
Ethereum.
Each meta-blockchain has an open, staked validator set.
Mosaic’s BFT consensus rules allow the meta-blockchain to offload
block formation onto an auxiliary chain.
Gateways allow ERC20 tokens to move in/out of meta-blockchains,
all value kept on Ethereum.

32. ostmosaic
•Tokens are the functional unit of Ethereum at Layer 2.
•Tokens define applications and interaction between
applications (ERC20, non-fungible tokens, …).
•Tokens naturally allow to repartition the state space.
Sharding at Layer 2: “Token Sharding”

33. Sharding at Layer 2: “Token Sharding”
OST GAS
full native EVM interface
scalable, on-chain smart contracts
today on mainnet (v0.9.2)
BUT today we oraclise between the chains
=> Mosaic decentralises state root transfer
Ethereum
Mainnet
Auxiliary
Chains
State Root
Gateway
State Root

34. State Root
Gateway
Ethereum
Mainnet
Auxiliary
Chains
Declare
Confirm
Progress
Progress
Merkle Proof
STAKE
MINT
Hashlock
Hashlock
Message passing is open process with bounties and fees for nodes to complete the flow for all messages.
State Root
Interblockchain Communication

35. Three Ingredients for Sharding ostmosaic
1. Divide work into shards
2.Communicate between shards
3.Economic finality on each shard

36. Intermezzo: Casper the Friendly Finality Gadget
Casper the Friendly Finality Gadget,
V. Buterin, V. Griffith, nov 2017
validators can send vote messages:
checkpoint justified iff exists supermajority link
from previously justified checkpoint
checkpoint finalised iff direct child is justified
slashing conditions, a validator must not publish
accountable safety and plausible liveliness
⟨s, t, h(s), h(t)⟩v
h(t1) = h(t2) ∨ h(s1) < h(s2) < h(t2) < h(t1)

37. Bringing Three Ingredients Together
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chains
(PoS)
B(header, txList, nonce) B(header, txList, majorityVote)
BM : Σ → Σ′

38. Proposing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chains
(PoS)
B(header, txList, nonce) B(header, txList, majorityVote)
BM : Σ → Σ′
BM(kernel K, transition T, seal S)
Meta-Block
K = {height, parent, ΔvalidatorWeights, gasPrice}
T = {dynasty, txRoot, gasUsed, coreId}
S = {
+2
3
weightedVotes}

39. Proposing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
in Core contract meta-block is committed:
T = {dynasty, txRoot, gasUsed, coreId}

40. Proposing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
finalise checkpoints:
in blockstore contract calculate T:
gas and txRoot for justified chain
in Core contract meta-block is committed:
T = {dynasty, txRoot, gasUsed, coreId}

41. Proposing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
finalise checkpoints:
in blockstore contract calculate T:
gas and txRoot for justified chain
any finalised checkpoint’s T
can be used to propose meta-block
Bh(Kh, Td, ⋅ )
in Core contract meta-block is committed:
T = {dynasty, txRoot, gasUsed, coreId}

42. Proposing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
in Core contract meta-block is committed: finalise checkpoints:
in blockstore contract calculate T:
gas and txRoot for justified chain
any finalised checkpoint’s T
can be used to propose meta-block
Bh(Kh, Td, ⋅ )
Bh(Kh, Td, ⋅ ) Bh(Kh, T′d′, ⋅ )
propose new meta-block:
T = {dynasty, txRoot, gasUsed, coreId}

43. Committing Meta-Blocks on Ethereum
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
in Core contract meta-block is committed: finalise checkpoints:
in blockstore contract calculate T:
gas and txRoot for justified chain
any finalised checkpoint’s T
can be used to propose meta-block
Bh(Kh, Td, ⋅ )
Bh(Kh, Tdh
, Sh)
commit new meta-block:
Bh(Kh, T′d′, ⋅ )
T = {dynasty, txRoot, gasUsed, coreId}
with Sh =
+2
3
Majv {⟨Tdh
, s, t, h(s), h(t)⟩v}

44. Holding Validators Accountable
Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
K = {height, parent, ΔvalidatorWeights, gasPrice}
Bh−1(Kh−1, Tdh−1
, Sh−1)
Bh(Kh, ⋅ , ⋅ )
in Core contract meta-block is committed: finalise checkpoints:
in blockstore contract calculate T:
gas and txRoot for justified chain
any finalised checkpoint’s T
can be used to propose meta-block
Bh(Kh, Td, ⋅ )
Bh(Kh, Tdh
, Sh)
commit new meta-block:
Bh(Kh, T′d′, ⋅ )
T = {dynasty, txRoot, gasUsed, coreId}
with Sh =
+2
3
Majv {⟨Tdh
, s, t, h(s), h(t)⟩v}
⟨T1, s1, t1, h(s1), h(t1)⟩v and ⟨T2, s2, t2, h(s2), h(t2)⟩v
1. h(t1) = h(t2)
2. h(s2) < h(s1) < h(t1) < h(t2)
3. s1 = s2 ∧ T1 ≠ T2
a validator MUST NOT publish two votes for which 1, 2 or 3 holds
slashing conditions are immediately enforceable on both chains.

45. Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
proposers
… but Validators Need Carrots too.
GAS
users pay gas fee to proposers for transactions
proposers deposit gas fee
to prefund meta-blocks
VOTES
validators are rewarded for votes ~ gas consumed
validators
validators must have voted on auxiliary
and committed on Ethereum to get rewarded.
OST
OST

46. Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
proposers
… but Validators Need Carrots too.
GAS
users pay gas fee to proposers for transactions
proposers deposit gas fee
to prefund meta-blocks
VOTES
validators are rewarded for votes ~ gas consumed
validators
validators must have voted on auxiliary
and committed on Ethereum to get rewarded.
A GasTarget for a meta-block forces
validators to commit on Ethereum.
Validators can join or logout
in the second following meta-block.
OST
OST

47. Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chains
(PoS)
last finalised checkpoint can always
be committed by anyone.
validator set can be challenged
that meta-blockchain has halted.
for halted core all value held in gateways
on Ethereum can be recovered with ownership proofs
against latest committed (equiv. finalized) meta-block.
No mass-exit required.
finalise checkpoints:
When it all goes wrong: Halting
GAS

48. Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chain
(PoS)
proposers
So what can we use this for?
users send EVM transactions
proposers propose meta-blocks
validators verify and finalise
anyone can commit at any time
validators
• Gateway consumes the state root finalised by validators
Users and contracts can pass messages (without a bridge); currently ERC20-typed
• For ERC20-tokens, once minted on the auxiliary chain, atomic swap is fast to move across.
Off-chain load-balancing logic.
• DApps (users) can opt-in to (and out of) auxiliary chains.

49. Mosaic (PoS)
Ethereum
Mainnet
(PoW)
Auxiliary
Chains
(PoS)
proposers
Summary
validators
•Security is asynchronously derived from Ethereum.
•# cores supported ~ total validator stake ~ driven by utility (rewarded in gas)
•Tens of thousands of transactions per second are asynchronously finalised on
Ethereum — at lower gas price.

50. Current Status - September 2018
• OST KIT & OpenST 0.9.2 used by 120+ external developer teams on Testnet
• OST KIT & OpenST 0.9.2 used by 10+ external partners on Mainnet — 3 months ahead of
schedule
• OpenST 0.9.3 (cross-chain messaging) — completed
• OpenST 0.9.4 (decentralized keys engine & token rules engine) — in progress
• Introducing Mosaic at several events like ETH Berlin & Upbit to solicit community feedback
• Huge thanks to the 65 person OST Team and the fantastic OST Community!

51. What’s Coming Next
Q3 2018:
• Publish OpenST Mosaic Paper for Community Feedback
Q4 2018:
• Continue Mainnet testing of OST, OpenST Protocol
2019:
• Mosaic alpha
• Mosaic beta - warm up validator pool
• Widespread Deployments of OST, OpenST, Mosaic to mainstream businesses and DApps
• Smart contract calls across auxiliary chains
• 2nd-order meta-blocks (aux-on-aux-on-ETHv1.0 or v2.0)
2020:
• Mosaic fully released to decentralised validator pool

52. OpenST Mosaic
@benjaminbollen
Questions & Feedback Welcome
@betashop ostmosaic
github.com/openstfoundation