OMG Network, what we know so far, part 2
by pwolf88 and Joel Foster (OMGpool)
This is part 2 of a series of articles on the OMG Network that we started in 2018: ‘What we know so far’. This tech-heavy write-up is compiled based on the OMG Mainnet launch on June 1 and all the community AMA sessions organized by Messari that were supporting the launch.
A Layer-2 scaling solution for Ethereum
Scaling the Ethereum root chain (Layer 1) using Plasma
The OMG Network uses Plasma to increase the processing capacity of Ethereum. Plasma is a Layer 2 (i.e. off-chain) scaling solution for Ethereum, which enables a very high throughput of transactions, with a lower cost per transaction compared to Ethereum. This is possible because transactions (selling, trading, transferring, etc.) happen on a child chain (referred to as Layer 2) instead of the Ethereum root chain (referred to as Layer 1). This relieves pressure on the Ethereum blockchain by condensing all transactions that happen across multiple child chain blocks into a single, validated Merkle proof that is submitted to the root chain. Interaction with the Ethereum root chain is only required for depositing funds to the child chain or exiting funds from the child chain back to the root chain when you withdraw or when anything goes wrong. This kind of construction enables a throughput of thousands of transactions per second and results in 1/3rd of the transaction cost for the end-users, while still preserving the high security aspects of the Ethereum root chain, as we will describe later in this post.
The Bitfinex use-case
The main use-case of the OMG Network right now is to support exchanges and wallet providers for cheaper, faster and more efficient value transfer (P2P transfer). The best example of this is Bitfinex, which will move some of the Tether (USDT) transactions onto the OMG Network’s layer 2 solution. Transactions will have a non-custodial settlement on OMG Network, while having a centralized order matching on the exchanges.
By migrating USDt value transfers to the OMG Network, we save costs, drive performance improvements, and relieve pressure on the root chain. This is good for Bitfinex, our users, and the entire Ethereum ecosystem.
(Paolo Ardoino, CTO at Bitfinex)
As shown in this article, USDt is by far the most dominant medium of exchange on Ethereum. The OMG Network will facilitate seamless value transfer between wallet A and wallet B.
The limited speed and the congestion of Layer 1 is a major hurdle for institutions to move in the space
All people who were moving funds during the spring of 2017 faced problems of liquidity and were not able to get transactions through the network cheaply and within reasonable times. Currently, Ethereum is reaching the same congestion status as back in 2017. During the flash crash on Ethereum in March 2020, the price of Ethereum crashed from $240 to $107, liquidating more than $10 billion from the market. People paid absurd amounts of money to access their funds, because the network was fully congested and paralyzed. The average network waiting times were rising from 15 seconds to 44 minutes. This is a major hurdle for large institutional investors to move into the space.
Why companies should move to Layer 2
In the words of Alexei Schaller from OMG Network, being able to access and transfer liquidity is like pumping lifeblood through a system. This is the core of all financial infrastructure. When moving over to L2, individual businesses can still compete by geography or any other differentiating factors, but they all benefit from Layer 2 because of:
- faster settlements,
- greater throughput (thousands of transactions per second versus 12–15 transactions per second on the Ethereum root chain),
- and the gas fees that are only 1/3rd the average cost on the Ethereum network.
For the last example, in the last 30 days, USDT tokens accumulated over $1mil in transaction fees. The network can cut this down by 2/3rd today.
Network congestion happens more and more often, resulting in fee spikes, delays in finality, and other undesirable effects. There is a huge drag on the growth of the decentralized ecosystem if we don’t shift to Layer 2.The entire DeFi crypto value framework is based on reliable, fast and efficient value transfer, and OMG is a major step forward in solving that problem.
We’re talking about core value infrastructure which will move the entire space forward.
How difficult it is to move to Layer 2?
The protocol of the OMG Network didn’t change any of the signing mechanisms or constructor transactions to submit to the network. The network-level transaction formats for L1 are the same on the OMG network, so it’s relatively straightforward for any wallet provider to connect to the network. Moving to the OMG network is straightforward, through well-documented SDKs.
The long-term solution for trading venues & wallet providers
The partnership with Bitfinex puts OMG Network in a good position to scale further and attract other exchanges and partners to the network. The network is actively discussing this integration with other wallet providers and exchanges. In the early stages, we might observe the so-called ‘chicken-and-egg’ problem: enough companies need to move to L2 in order for it to be appealing. In these markets, liquidity attracts liquidity.
The more partners that are connected to L2, the more reasons there will be for other businesses to join.
What you can transact on the network
Currently you can transact any ERC-20 on the OMG Network. This allows the network to support cryptocurrency and tokenized fiat transactions, and transactions involving any tokenized value such as loyalty points or in-game currency. But part of the work of Q4 2019 (initial alpha and closed beta) was building the ability in the child chain and the contracts to support multiple token standards, so it’s relatively easy to add ERC-721 (NFT’s) and any other applicable root chain token standards. The network will definitely support those as we go forward. This will unlock support for transacting provably-unique digital items such as tokenized real estate deeds, video game items, and much more
How fast are transactions?
The OMG Network is bound by the protocol and block production times of Ethereum, but it can process 2^16 transactions per single block. With a 15 second interval on Ethereum, this means the network can handle about (2^16)/15 = 4369 transactions.
How much do transactions cost?
The network currently uses a fixed fee model. A 100K$ transaction will cost only 0.059 OMG. However, the OMG Network is planning to change its fee model this week. The new fee model will still require fees to be paid in OMG tokens, but will be changed from a fixed fee into a dynamic fee based on one-third of the ETH gas fee. For investors holding the token, this is an important step, because it will remove the incentive to keep the price down. The token price can increase and no longer raise the cost of transactions on the network.
How will fees be paid?
Currently, OMG is the only accepted fee, but the platform is designed to eventually support multiple fee tokens (e.g. ETH, DAI, etc.). They will be added soonTM.
What happened to the initial idea of a full on-chain DEX?
By looking at the wider L2 community, and looking at both centralized and decentralized exchanges, the OMG Network realized that on-chain order matching really doesn’t scale. It has both technological problems and economical ones. For instance, it is very easy to front run orders that have not yet been matched on the DEX. Frontrunning (also known as tailgating) means a broker or trader could enter into a trade because he has foreknowledge of a large pending transaction that will influence the price of the underlying security. EOS has on-chain dexes, but there are a lot of limitations. Especially with high-frequency trading, it’s close to impossible to avoid this issue.
With the OMG Network, the order matching (order management and trade execution) happens on the venue, but the settlements are conducted on the network. The venue provides the order book with sufficient liquidity, and the settlements can be on-chain on L2 and are then pushed back to L1 for finality. This hybrid model gives you most of the benefits of a DEX, without the order matching being decentralized, as you cannot afford to have any time delays there.
The OMG Network also hopes to see a lot more centralized exchanges do their order matching on L2, as a lot of them today are off-chain because of the problems mentioned above. Currently their order matching is done off-chain in a custodial way, and eventually the settlement is done either at the end of the day or when you withdraw. With a high-performance and trustless non-custodial layer 2, a lot of that trade activity can move on-chain so that exchanges holding custody of user’s tokens are no longer necessary for speed.
The incentive for regular users to join the network
The benefit for the users is the speed of the network and cheaper transaction fees, while retaining full custody of their funds. The more scale we reach, the more users will join. These benefits will accumulate for these participants and will unlock a lot of capital or funds that would otherwise have been used to pay transaction fees, which now gives holders the possibility to reinvest them. It provides more value to the ecosystem because less is paid for the transfer rails.
Fast finality in parallel with Layer 1
Finality is the guarantee that past transactions can never change. The deposit finality period refers to the number of Ethereum block confirmations required before a deposit can be used on the network. It is in place to mitigate the risk posed to the OMG Network by Ethereum chain re-organizations.
In the OMG Network, finality cannot be very fast, as it is one of the security measures of the network. A Plasma child chain block is produced for each Ethereum root chain block, so while the transaction processing is much faster, the finality block time is still 15 seconds. If there are transactions on the OMG network, those are batched up, a Merkle proof is created, and this is submitted in the root chain in a mainnet transaction. This means finality is still parallel with Layer 1. Finality cannot be faster than the root chain block, because you’re trusting the child chain and those commitments haven’t been published widely yet. With Ethereum 2.0 it will be faster (probably 4 seconds) and this will benefit OMG as well.
Staking and Proof of Guarantee
Proof of Guarantee (PoG) with a single block producer
Currently the live network uses Proof of Authority (PoA), which is a consensus mechanism that gives one node, called an Operator or single block producer (in this case, the OMG team’s node), the sole right to make all of the blocks in the Plasma blockchain. Other nodes called Watchers, which oversee the transactions processed by the Operator, can be run by anyone. There is no incentive yet in this stage for people to run a Watcher, though after this initial stage, the intention of the network is to go to Proof of Guarantee (PoG), where people are financially incentivized to run a Watcher. The Watchers will check for blocks that are submitted to the root chain, pull the Merkle proof from the operator, and then validate that the transactions that are supposed to have happened, have indeed happened. In this system, the network is semi-decentralized through the Watchers that can identify nefarious behavior by the Operator, including if it becomes unavailable or if blocks are being withheld or if transactions are manipulated, and flag all funds on the Plasma chain to be withdrawn back to Ethereum (this is called a “mass exit”). Watchers are able to do this because they have visibility to data on both the L1 and L2 blockchains.
The PoA/PoG design was built to minimize trust and allow for faster, cheaper, and non-custodial value transfer on Layer 2. The Watchers are there to keep the single block producer honest, and will be rewarded for that, which brings back some of the decentralization.
The way Plasma is constructed, your funds are always secure because of the exit game that everyone has to play. An Operator could get compromised or become malicious — in both cases there is no direct punishment for the Operator, but the consequences are implicit. After a Watcher-initiated mass exit, nobody would trust or use the Operator anymore.
How mass exits work
The easiest way to perform an exit from the OMG Network back to Ethereum is using the UTXO information that is held by a Watcher, to start to perform an exit. Technically you don’t need a Watcher for an exit, as all you need is data availability, but Watchers exist to constantly provide this data. If a Watcher is not available, there are other backup mechanisms that can be used to exit funds back to L1.
OMG never takes custody
Exits are possible because nodes on the network never take custody. This means nodes on the network are not storing the private keys of the users. The user still has his private key to sign the transactions and to prove ownership of the tokens and the value, and they are secured on L1.
Maximize for decentralization and security
The fact that root chains are slow is not a mistake, as they try to maximize for decentralization and security. The idea that OMG Network is not taking custody while still preserving Layer 1 security softens the criticism of being a single block producer. It’s all about trade-offs. Theoretically, the network can become fully decentralized as more Operator nodes are added to the network, in addition to a wide array of distributed Watchers.
The reason the OMG Network shifted from PoS to PoG
The OMG Network believes that PoG is the better way to go compared to PoS, for multiple reasons:
- PoG is less expensive. If Plasma used PoS, for every submitted block that is pushed out the root chain, each node would need to pay a significant gas fee. Rather than having every node act as an Operator submitting blocks to the root chain AND validating other block producers, only some nodes will have this responsibility while being under the scrutiny of the Watcher nodes, achieving the same security effect but with increased speed and lower overall network costs.
- PoG on Plasma is faster than PoS on Plasma. As mentioned above, the fewer the block producers, the faster the network. However this carries risk to decentralization, such as with EOS’s 21 supernodes. PoG aims to achieve the same speed effect using fewer block producers than PoS, but uses the system of Watchers to retain the decentralized security aspect. In PoS, all nodes are both “Watchers” that call each other out for malicious behavior and “Operators” that submit blocks, whereas on PoG these roles are separate.
- The solution is green: it mirrors Ethereum’s graduation into 2.0, which reduces electricity usage per transaction by up to 99% compared to Ethereum 1.0’s Proof of Work system.
It will take a while to implement PoG into the network. OMG Network has internal specifications about PoG, such as rewards and staking mechanisms, but it is not yet ready for release. The documentation is being finalized, and will be shared with the community soon (TM).
Running a watcher
Currently you can run the Watcher on your laptop or you can run it on a VPS. It’s a background service that can be run anywhere, although it’s preferred to run it on a VPS or dedicated server, because the internet connection and uptime will be a lot better.
The Watcher is quite light-weight, so it doesn’t require much storage or CPU. Eventually, when there are more transactions on the network, the hardware requirements will be more strict.
On the documentation portal, there is already a guide for running a Watcher locally, and a basic guide for running it on a VPS.
Fees will be divided when staking is launched
Right now the Watcher is just a service that watches the transactions that are happening on the child chain, and is syncing with the Ethereum root chain so that everything is verified and confirmed. The network is providing all of the infrastructure for the entire setup (the child chain, the connected API’s, the infrastructure in front of it, as well as OMG’s own Watchers that are validating OMG’s Operator).
Currently, you do not receive any rewards yet from watching the network, but this will change when staking is launched. Stakers on a Watcher will be rewarded proportionally to the amount of tokens that they have staked. In OMGpool’s staking rewards calculator, you can estimate the rewards a user could get.
There are operational fees that the Operator has to pay as well. For this, a small percentage of the fees will be retained to cover the costs. Remaining fees will be distributed among Watchers.
Today the fees are covering the costs of running the network
The fees that are collected from running the network currently go to covering the costs of the infrastructure as a single block producer (including gas fees to submit blocks to the root chain), which is quite costly at the moment. There are no token burns as well. As soon as staking rolls out, the rewards are there to be shared.
Looking at the (bright) future
How future updates can be done on the network
The smart contracts in the OMG Network are built using an Abstract Layer Design (ALD). This introduces the concept of Transaction Type and Transaction Output Type. Each Transaction Type and Transaction Output Type can define different rules about how to spend funds. This means that new business logic can be introduced without upgrading the core protocol, so no hard fork would be required. For instance, new transaction types (like ERC-721s) can be implemented without asking people to exit. While new transaction types can be introduced, existing ones cannot be changed.
For both asset types and applications, OMG Network will be led by demand and concrete use cases as they speak to more partners and customers.
So what now?
Many things can begin using the OMG Network:
- Centralized and decentralized exchanges
- Wallets and interoperability between wallets
- Point-of-Sales systems, using QR codes (either on the POS screen itself or on the paper receipts)
- Rewards and loyalty programs with the eWallet SDK
- Lending
- Payroll use cases
- dApps
- and more!
The OMG Network is building the basic roads and infrastructure into an open landscape. CurveGrid for instance utilizes this core infrastructure to make it easier for third parties or users to spin up their own Watcher.
Although the current throughput of thousands of transactions per second will provide sufficient capacity for a while, the volume could start to increase quickly when there is a large demand for Layer 2 scaling. Discussions are already on-going on how we could scale the OMG Network even further. One option is by introducing a third layer: a new child chain rooted to the L2 child chain. Already 32 bytes of metadata per transaction is supported so the engineering effort wouldn’t be too big to include metadata of a third child chain in the L2 transactions, which then resolve their finality on L1. These nested Plasma chains could theoretically scale the OMG Network infinitely, so long as the layer below them is sound. Hopium.
What’s on the roadmap
OMG wants to introduce more Watchers and more volume onto the network. They are currently in the process of solving some key elements that will improve the network even further:
- Staking specifications (as described above)
- ERC-721 and custom transaction types (as described above)
- Fast Exits: currently an exit takes 2 weeks (proof of concept stage currently), but this will be shortened. In short, someone who is watching the network will constantly be assessing a challenged exit, and could prefund a “fast exit” to the person who is requesting the exit. Fast exits will be a gigantic win for the network. A summary of Plasma’s Implementer Call 21 can be found here, with interesting opinions about Plasma roll-ups and fast exits.
- If the Ethereum network remains congested and there is a need to move more transactions from Ethereum to OMG Network, it takes quite some time to do that. During such a migration period when the root chain is congested, you basically need to adjust the gas fee. OMG is looking to onboard more and more transactions onto the network to help reduce this congestion, though this requires the bizdev team to support partner adoption. Tether is currently the largest cause of this congestion, so it is fantastic that Bitfinex is first to adopt the network.
- The concept of an Enterprise Plasma has been mentioned during the AMA’s, though the details have been sparse. We speculate it may entail allowing enterprises to form their own Plasma chains — perhaps Layer 3, resolving to the OMG L2 chain before establishing finality on L1. If this is the case, it would allow enterprises to maintain their own self-contained system on the network, which provides them greater control for their use-cases. We’ll keep our ear open for any new information about Enterprise Plasma chains.
- Privacy on Plasma (such as ZK Proofs) is currently not yet in the pipeline, but the OMG Network will watch these developments on Ethereum 2.0 closely and might use some of its approaches, although it will take some time to develop.
More info
- See the OMG FAQ for more information about the topics in this article.
- The transactions on mainnet can be followed on the Block Explorer.
- The currently accepted fees can be seen here.
- The steps to deposit funds from the root chain to a child chain web wallet are described here.
Edit: formatting fix
Many thanks to the contributors of the AMA’s: Jonathan Mander, Eddie Fulton, Jim Eneman, Ghaffar Glenn, Apple Agrawal, Steve Paduchowski, Jim Dawson, Andrew Tu, Po-An-Lin, the speakers and moderators from OMG Network (Alexei Schaller, Dmitry Baimuratov, Stephen McNamar, Jake Bunce, Bowen Shen, Pong Cheecharern, Gerhard Wegner, Victoria Daet, Jeremy Lam, Jet86, …), the folks from Messari, as well as tousthilagavathy, Sir-Kao-Pad, ii_OiO_ii , and many other members of the community for their contribution.
OMGpool (omgpool.org) is building the first open-source, community-driven staking pool for the OMG Network.
Be sure to check out our staking rewards calculator on our website omgpool.org
If you’d like to be the first to hear about the staking pool when it launches, sign up for email updates here or in the form below.
Because we love hearing from you, we’ve provided a feedback form here where you can share your thoughts with us. We’re looking forward to your messages!
