Part II in a new series discussing the challenges and the design choices that represent the backbone of the Dusk Network protocol.
Part I of the series discussed the basics of applied economics in distributed systems as well as outlining the game-theoretic principles indispensable to the Bitcoin protocol.
The dichotomy (i.e. two contrasting propositions) of commodity scarcity and availability extends to the auction theory. Any conceivable system can only allocate a finite amount of resources leading to natural scarcity when the demand exceeds the theoretical maximum supply on offer. A game theoretically-sound incentive-based distributed system also defines artificial scarcity, which either acts as a deterrent to distributed denial of service (DDoS) attacks or as an incentivization mechanism for honest (participants, adhering to the predefined protocol rules) protocol participants (or both). Dusk Network protocol is no different, having a finite supply of computational power and storage on offer per block. The cost of computation performed on cryptographic systems, the speed of instructions requiring the Virtual Machine to fetch data from the global state is one of many constraints imposed on the throughput of the Dusk Network protocol.
How to tackle?
Bitcoin utilizes a generalized first-price (GFP) auction model to enforce the fee policy on-chain. The model variant deployed in Bitcoin compels users to propose fees derived from a non-trivial assessment of constantly-evolving factors, including the average proposed fee in the previous b blocks, the contents of the mempool, etc, rather than a fee derived from the honest valuation of the importance of the transaction to the user.
A number of papers published in recent years (e.g.  ) propose a generalized second-price (GSP) auction alternative to the generalized first-price auction model deployed in the exceeding majority of the existing digital currencies.
In simple terms. There is the generalized first-price (GFP) models used by Bitcoin, where those who are prepared to pay the most for their transaction to be processed, are processed first.
And there is the generalized second-price (GSP) where those who are prepared to pay the most for their transaction, are processed first, but pay the gas price that those who were outbid by the former, were willing to pay.
A generalized second-price auction model enforces the winning n auction participants to pay the fee proposed by the n + 1 participant (i.e. the participant with the highest proposed fee to not win the auction). An application of the GSP model in the distributed systems requires multiple modifications to the original algorithm, including the enforcement of the fee proposed by the nth participant (i.e. the winning participant with the smallest proposed fee) rather than the n + 1, an even distribution of block rewards amongst the generators responsible for forging blocks per epoch and a penalty enforced on the generator for not reaching a predefined minimum threshold for computation. As many of you might have already noticed, GSP is only possible when the demand for the commodity is higher than the supply, meaning that the utilization of GSP is contingent upon a stringent enforcement of a computational limit or mass adoption driving the utilization of the protocol to the theoretical limit, unlikely to be reached in the infancy of the protocol.
The abstract transaction fee structure
Dusk Network protocol will utilize a fee payment structure similar to the Ethereum protocol’s gas model to administer the fee policies. The gas model requires the transacting user to specify the cost in DUSK at which he/she is willing to purchase gas and the maximum amount of gas he/she is willing to allocate for the transaction execution. The gas is intrinsic to the Virtual Machine and does not exist externally. Similarly to Ethereum, the block will not have a physical size limit, rather having an adaptive cumulative gas limit per block due to the computational power being a more scarce commodity in the Dusk Network protocol than storage and network throughput.
Making stuff more concrete
As outlined above, the deployment of GSP is non-trivial without imposing an artificial bottleneck on the protocol when launching a brand new offering. The decision was made to proceed with GFP while imposing a lower bound on the cost of gas until the protocol reaches the levels of adoption capable of pushing the system closer to the theoretical limit, at which point the protocol will transition to the GSP model.
So what is next?
Part III will outline the use of fees in terms of the Virtual Machine and how certain fee structures are necessary to avoid the “halting problem”.
Finally, in Part IV we will take a look at the ways in which the network participants are incentivized to retain the security of the protocol as well as the measures deployed to penalize malicious behaviour.
Dusk — Technology for Securities
Dusk Network is an open-source and privacy-oriented blockchain based on years of academic research. You can use Dusk Network to create smart contracts that control digital assets and securities.