The biggest impediment to the broad adoption of decentralized applications (dApps) is the requirement that participants onboard to a cryptocurrency first. As a consequence, interested parties often do not adopt because the ‘sign up’ or ‘create an account’ flow involves learning about wallets and keys. Then, participants find that they need to purchase tokens to pay for gas, which leads them to try to understand exchanges. By this point, it is understandable why the average participant already wants to give up. Overall, the contemporary dApp user journey is closer to about as painful as opening a bank account when it needs to be as simple as signing up for Instagram.

This issue is well-documented for many platforms. For example, there are several issues filed for the Ethereum Improvement Process (EIP) discussing workarounds and solutions for this issue:

“*Alternatively, a 3rd party may wish to subsidize the gas costs of certain contracts. Solutions such as those described in EIP-1077 could allow transactions from addresses that hold no ETH. The gas station network is an EIP-1077 compliant effort to solve the problem by creating an incentive for nodes to run gas stations, where gasless transactions can be ‘fueled up.’”*

Ethereum’s EIPs and the broader dApp developer community have known of this issue for some time. The community has identified potential ways to address the problem, but all of the solutions remain either centralized or speculative.

Kadena’s smart contract language Pact presents a novel approach to overcoming the impediments associated with onboarding to cryptocurrency. Pact has always supported general multi-sig, and its signing system is the first smart contract language to bring a generalized capability-based security model to market. This allows both protocol and app developers to go much further than previously possible. To solve the onboarding problem, Pact uses something closer to co-signing transactions.

Through co-signing, dApp operators can have participants locally create their first wallet keys and use them to construct pre-determined transactions (e.g., a transaction that creates the user’s Kadena account and establishes the user’s dApp account simultaneously). These keys can even be made in the user’s secure enclave, making the experience seamless. The user then sends the transaction to the dApp operators. This transaction specifies the dApp operator’s wallet as being the gas payer. Before co-signing, the operator verifies that the user’s transaction is correct.

While this is a technically simplified example, it does require an identified dApp operator to co-sign transactions. The upside is that dApp users don’t need to be made aware of any of the implementation details done by the dApp devs. All users see is a frontend and user experience that they are used to: “I get an app and use it.” For a more decentralized approach, we need gas stations.

We recently upgraded Pact’s capabilities implementation to allow for the creation of accounts that can only be used to fund gas payments. Gas stations are the logical conclusion of this ability: an immutable resource, in the form of an account that exists solely to pay for gas fees encountered during smart contract execution involved with a specific contract.

The technical specification of gas stations can get complex and will be the topic of a separate post, but the high-level summary is that once a gas station is created, it is a non-transferable resource. A gas station can only be used to redeem gas fees encountered by specific contract logic. The refund occurs only when a transaction is successful. The participants would pre-pay the gas, which gets entirely consumed only if the transaction fails. Otherwise, the gas station keeps the used gas while refunding the remainder.

Up until now, we’ve been discussing this concept with the assumption that dApp developers would be co-signing transactions for those using the dApps, fronting the gas fees, and getting refunded through gas station accounts. However, our approach is general enough to permit incentivized anonymous co-signing of transactions. This could function similarly to how mining rewards work, except that the dApp itself would fund the rewards for anonymous co-signers. Experimentation on the economic side will be needed to see what works here.

With Pact’s capability system, dApp developers could potentially allow these anonymous co-signers to consume, verify, co-sign, and submit a transaction that will refund their fronted gas, as well as give them a small bonus for helping facilitate the transaction execution on-chain. The co-signer only co-signs the transaction, making the decision point simple. Is co-signing in the co-signer’s economic interest? What’s happening under the hood matters only in that the co-signer would want to check that the transaction is valid and will succeed.

This mechanism makes it feasible that by mid-2020, we may start seeing the first mempools that gather not yet co-signed transactions specific to a given dApp. There will be anonymous parties tapping into these mempools and taking transactions that they are willing to co-sign.

The first gas stations will be built on Kadena in 2020, and deployed in a very limited way as the economic impact of them is still unclear — we’ll need to build a few, put them on mainnet, and see what happens. The key takeaway is that we’ll be able to test this approach for the first time and experiment to figure out what works best.

Most of the data we have on “how gas works in practice” comes from Ethereum and this data differs significantly from how it was “supposed to work in theory.” ETH gas prices should theoretically fluctuate like fees in BTC, but in practice, there is little evidence to suggest that ETH’s gas prices are determined by market forces most of the time. Instead, it seems that ETH miners view smart contracts (and thus gas) much the same way that BTC miners view transactions: most miners would rather mine empty blocks most of the time, but doing so undercuts the value proposition of the network.

We need to do some experiments and determine if gas stations are seen as honey-pots or something closer to a take-a-penny-leave-a-penny cup. Miners could pad a block with gas-station funded fees, but if we’re to use ETH’s data as a guide, then the overall mining rewards on Kadena will likely dominate even a block padded with coins from a gas station. So, it is possible that gas stations are seen as part of the value proposition of Kadena. Or, perhaps, an open gas station is seen as a take-a-penny cup and not something worth plundering.

Gas station creation will be demand-driven, and 100% of gas station tokens will go toward paying for the fees associated with dApp transactions. We are devoting tokens that other projects would have burned toward the gas station concept, encouraging usage in a transparent and decentralized way. No one has ever tried to just pay for smart contract fees across an entire network before; we’re excited to see what happens.