Since the launch of the Meson Testnet, we have gone through several iterations for the version. The number of miners and network usages have far exceeded our expectations. As a result, we are ready to release the Tokenomics-Beta1 and apply it on the Meson Testnet. The following contains information on several key modules and issues. Since the crypto world is still relatively early and rapidly progressing, our model may be subject to further improvements and parameter adjustments based on actual data collected throughout the testing period. Enjoy.

Decouple the supply/demand and the speculation

We have noticed some design problems in the existing economic system, which we will discuss in further details here.

Problems in Ethereum

The Ethereum token model is designed while taking into consideration the two major factors that can influence a token economy - the supply/demand of the token and the level of market speculative activities (speculation is a neutral term here, expressing situations outside of supply and demand, such as secondary market operations).

When we use a service, the pricing of the service should be primarily determined by supply and demand. However, when using Ethereum's services, we find that the price (in stablecoin terms) is very cheap in the early days of use. As Ethereum's popularity grows and its applications become more popular, the price (in stablecoin terms) that users pay for the same service rises dramatically. If we only look at the supply and demand aspect of the service, the magnitude of the price impact should be much smaller than the actual price fluctuations that users now perceive. However, the mechanism is designed in such a way that a large amount of market speculation is directly reflected in the ether price, and the demand side uses the service in ether as the local currency, resulting in users being forced to bear the price premium to the services.

From this we can see an interesting phenomenon. When Ethereum is not well-known, the service price is very cheap. When Ethereum became more famous, the token price rose sharply. At this time, users found that they could not afford the service and were forced to look for alternatives (e.g., Polygon). When the usage and popularity of the alternatives increase, users find that they cannot afford to use the alternatives anymore and are forced to continue this counter-productive cycle.

More applications and more market acceptance can lead to higher token prices, but higher token prices drive up the price of services and, to some extent, suppress demand, which in the long run causes the network to lose value and leads to lower token prices. The paradox is thus formed.

How to pay for the fees

Regarding the question of how users pay, we tend to be product and service-oriented with two custom-designed characteristics. First, the product is denominated in stable currency, and users do not need to worry about drastic changes in service prices due to currency price fluctuations. Second, users can use any token they like to pay for and use the product, including ether and various ERC20-compatible tokens. In the future, we will gradually support more non-ERC20 standard tokens as well. When users use Meson native token to pay, they can be granted a discount on service.

How to distribute the fees (demand side)

Regarding the question of how to allocate the fees being generated from the demand side, let's first look at the types of allocation methods. The first one is allocated to the direct provider of the service. The second one is to convert the revenue and token in a certain way.

For the first approach, we can look at a platform like Uber. The passenger pays the fare, of which about 80% goes to the direct service provider (driver), and 20% flows to the platform and is reflected in Uber's share price to some extent. For the second conversion method, several existing models can be referred to. One is the BNB token model, where the revenue from the product (the bulk of which is the exchange fees) is bought back and burned every quarter. The second is the ETH2.0 token model, where the system calculates the number of tokens required for users to use the service, burns them in real-time, and then gives certain tips directly to the service provider. The third is the approach adopted by Helium, where the $HNT assumes the role of a channel fee. Users pay stablecoin $DC to use Helium services, and $DC can only be exchanged by $HNT. The number of generated $DCs and the destroyed $HNTs have the same value in the stablecoin niche.

The method we adopt is based on the Burn & Mint mechanism, combined with the characteristics of the Tips from ETH2.0. For the demand side, users' basic fees are denominated in stablecoins, and the system calculates in real-time. Users can use a variety of tokens to pay, and this part of the fee will be converted into Meson Token and destroyed to generate equivalent Meson Credit. Meson Credit and USD have a fixed conversion ratio. Users can get a 5% discount if they pay Meson Credit directly. At the same time, users can also choose to pay a certain amount of Tips in exchange for more benefits from the service provider (e.g., higher processing priority, longer cache time, etc.), and the Tips are paid directly to the node hosting the service.

How to decouple the supply side with native token

In Meson's system, we further discuss the type of currency paid to the supply side (node). All payment fees, all token payments, or a certain mix. We first consider two edge cases: 1) If there is no use on the demand side, then only tokens may be valuable to the supply side, similar to the situation of Ethereum in 2015-16. 2) If the demand side is enormous, then under a reasonable economic model design, the service fee on the demand side will be reflected in the token price. At this time, for the supply side, only tokens can be good enough. It can be seen that in two edge cases, it’s possible to pay no fees. For the middle part, is it necessary to do a certain mixing ratio, such as 50% token and 50% fee?

We can change our thinking on this problem. Under the background of the above mechanism, if we need to achieve the goal of a 50% fee, we can actually solve it through the inflation rate. If the total amount of tokens issued to miners after convergence equals the number of tokens in the genesis block, miners will get 50% token + 50% fee in disguise. Therefore, the solution we adopt is only to pay tokens to the miners and add a small part of the tips mechanism in exchange for additional services (such as longer cache time, higher priority cache queues, etc.).

The price for the Service

The pricing of services, we address the issue in two phases.

In the first stage, a fixed price is adopted. Similar to mainstream cloud vendors, we will provide a price based on bandwidth usage and volume in different regions. We expect the price to be less than 1/10 of mainstream manufacturers.

In the second stage, market matching is adopted. This method is more similar to the orderbook, where both parties in the market quote on demand and match on the Meson side. There are significant benefits to be gained from using the orderbook matching model. For example, we cannot accurately calculate the cost price of users on the supply side. When the demand side quotes 100 units, all supply nodes can benefit from it and provide services; when the demand side quotes 50 units, some supply users are forced to leave but the cloud Manufacturers can continue to make profits; when the demand side quotes 10 units, most users cannot cover the cost but operators (Telecom) can still continue to make profits. Thus, the model essentially provides a free market where prices for services are matched by supply and demand. In view of the performance factors and user convenience of existing on-chain matching, we first adopt a fixed pricing model in this version.

One token or two tokens model

We use the scheme of the single token model.

The inflation rate for the supply side

Regarding the inflation rate, the network will issue a certain percentage of tokens to miners every year. We have observed that the inflation rate sometimes has problems deviating from actual development. For example, some projects issued most of the tokens to miners in the early stage. As a result, when the actual adoption of the project broke out, due to the small proportion of tokens to be distributed, it was not easy to increase the volume of supply-side nodes. Therefore, we decided to match the inflation rate with the actual network conditions and modify the rate in due course. Then how to modify the inflation rate becomes the next problem.

A relatively simple solution is to do a governance poll. The token holder participates in voting according to certain rules and modifies the parameters regularly. This program looks democratic at first glance, but ‘Where one stands depends on where one sits.’ For example, miners will choose to increase the inflation rate, while other token holders will decide to reduce the rate. Another solution is that we set the rules for the change, just enter the parameters regularly and modify the inflation rate. For example, the token price and burn rate are used as parameters, and the fixed function Func(token price, burn rate) = K * burnRate / tokenPrice is modified yearly according to the calculated value. We will adopt the latter scheme, which combines the inflation rate with the actual development on the one hand and avoids the concentration of rights on the other hand while making the rules fairer.

https://blog.meson.network/blog/2022-12-07-tokenomics-beta1

Abstract

As the digital economy advances, data centers have become necessities, the core infrastructure that makes modern-day living possible. Per Moore’s Law, large data centers are known to consume more energy as data transmission increases. However, at the same time, the world is threatened by climate change, and data centers are accelerating the deterioration of our environment.

Through careful investigation of peer review journals and efforts to control global warming worldwide, the MESON team hereby proposed Data Center Carbon Credit(DCCC) as a medium to support data center carbon trading. The MESON network can better utilize the stock of idle resources and coordinate bandwidth resources across borders, regions, and projects through token incentives. It also encourages the construction of mega data centers. Meson is an effective platform that empowers global data centers to meet the development of modern technology and the needs of energy conservation and emission reduction.

Carbon emissions and IDC background

How important is the data center in the digital economy

The data center can be considered the “brain” of the Internet. Their role is to process, store, and communicate the data behind the countless information services we rely on every day, whether it’s streaming video, email, social media, online collaboration, or scientific computing. Digital services are the main reason for the increase in energy use. The demand for digital services has been steadily increasing, followed by the development of its infrastructure. As infrastructure increases, so makes the energy demand.

How Much Energy Do Data Centers Really Use?

Data centers use an estimated 200 terawatt hours (TWh) each year. That is more than the national energy consumption of countries such as Iran, half of the global transportation electricity consumption, and 1% of global electricity demand. All in all, data centers contribute around to 0.3% of global carbon emissions.

Information and Communication Technology(ICT)’s carbon footprint is on a par with the aviation industry’s emissions from fuel.

One of the most worrying models predicts that electricity use by ICT would exceed 20% of the global electricity consumption by the time a child born today reaches her teens, with data centers using more than one-third of that.

Servers will be major end-use and most of which will all be integrated into hyperscale data center.

Many energy consumption reports have been published in the past few years. The International Energy Agency (IEA) reported that workloads and internet traffic would double in 2021, yet data centers’ energy demand will remain flat due to increased efficiency.

However, some reports refuted the claim. Uptime Institute Intelligence believes that there will be strong factors driving data centers’ energy consumption. Some data even contradicts that of IEA.

For example, the IEA reported that worldwide data center energy consumption was 197.8 TWh in 2018, with a slight drop in 2021. But the European Union Resource Efficiency Coordination Action (EURECA) Project said that European data centers consumed 130 TWh last 2017, while Greenpeace estimated that China’s data centers consumed 160 TWh in 2018. This means that both China and Europe alone have consumed 290 TWh which is far higher than the data provided by the IEA.（https://journal.uptimeinstitute.com/data-center-energy-use-goes-up-and-up/)

In the US, data centers have diminishing returns, which may limit the impact of energy savings. For example, at the data center level, best practices such as hot/cold aisle containment, installation of blanking plates and raising set point temperature have already been widely deployed; this can be seen in the substantial drop in power usage effectiveness (PUE) between 2011 and 2014. However, since 2014, PUE has not dropped much. In 2019, a slight increase in the average annual PUE was reported by respondents to a global data center survey. Similarly, with IT hardware, Moore’s law has slowed down. Newer servers are not maintaining the same efficiency improvements seen in the past.

If hardware optimization has been yielding diminishing returns, then optimization in the resource distribution system is the main source of efficiency improvement in the future.

Uptime Institute expects the strong growth in the IT sector to be sustained over the next five years, given the well-understood demand patterns and the existing technologies coming into large-scale adoption.

What drives the increasing energy use of data centers?

In addition to the demand from cloud, colocation, and enterprise data center, 5G is another rising sector that requires large scale infrastructure.

While it will take a few years for 5G to mature and become widespread, it is widely expected that the rollout of 5G will substantially accelerate the data growth, with new types of digital services in domains such as smart cities, IoT, and transportation.

The larger bandwidth compared with 4G will lead to a growing demand for higher resolution content and richer media formats (e.g., virtual reality) starting in 2021.

Social media also contributes to the explosive growth of energy use. Research by Uptime Intelligence shows that every time an image is posted on Instagram by the Portuguese soccer star Cristiano Ronaldo (who at the time of writing had the most followers on the platform), his more than 188 million followers consume over 24 megawatt-hours (MWh) of energy to view it.

Media streaming, which represents the biggest proportion of global internet traffic, has become the energy guzzler of the internet. Streaming a 2.5-hour high definition (HD) movie consumes 1 kilowatt-hour (kWh) of energy, but for 4K (Ultra HD) streaming — expected to become more mainstream in 2021 — this will be closer to 3 kWh, a three-fold increase. To utilize the global existing data center bandwidth resource pool to the fullest extent can also minimize the global carbon footprint of data centers. The most viable way right now is to build more efficient distributed systems for bandwidth resource allocation.

Protocol Design

The Meson Network designs a protocol that accommodates both the demand of the data center and the demand to reduce global carbon emissions.

Data centers generate carbon emissions throughout their lifecycle. During the construction process, data centers consume raw materials (such as steel, cement, etc.), staffing, and the destruction of the natural resources of the local land and vegetation, which can be abstracted into the carbon emissions of building construction and comply with the standards of various countries.

The data center consumes energy in several aspects during operation, ranging from IT equipment, cooling systems, lighting, and other ancillary parts.

The operation of the server needs to be kept within a certain temperature limit due to the constant heat generated from computing. The cooling system is essential because it discharges the heat and ensures the safety of the entire operation. For data center energy efficiency indicators, there is a unified parameter called PUE:

***Power Usage Effectiveness (PUE)***A metric that is often used to describe the efficiency of data centers. It relates the total energy consumption of a data centers to the energy consumption of the IT equipment: PUE=(total data center energy consumption)/(energy consumption of IT equipment)

For the cooling system, there are roughly three mainstream solutions:

Overview of different cooling systems for data centers

Air based cooling systems* In these systems, cool air is supplied to the server rooms. The server racks are often arranged in so called ‘cold’ and ‘hot’ aisles, to control the airflow and eliminate mixing of cold and hot air. Due to low heat capacity and heat transfer coefficient, air is not a very good medium to transfer heat, which results in high energy consumption, limitations to how compact the servers can be placed and a relatively low waste heat temperature. Liquid based cooling systems use a liquid such as water to dissipate heat. This can be done by circulating water in micro channels and exchanging heat in cold plate heat exchangers which are in direct contact with the server components. Water and liquids in general have significantly better heat transfer properties compared to air. Liquid based cooling systems allows for more compact datacenters, reduced energy use for cooling and higher waste heat temperatures. Two-phase cooling is an emerging form of data center cooling technology. Here, the liquid coolant evaporates in the cold plate heat exchanger and the dissipated energy is stored as latent heat. This allows for even greater heat fluxes, and coolant return temperature and makes systems with even higher computational densities possible.

Focusing on the core energy consumption parameter of PUE, we hope to encourage the world to build, use, and upgrade data centers that are more environmentally friendly, and limit nodes with poor energy structures.

Therefore, the Meson Network proposes a set of Data Center Carbon Credit(DCCC). For data centers that do not meet the PUE index, a certain amount of DCCC needs to be purchased to offset their carbon footprint, while for nodes with better PUE, DCCC can be sold in the secondary market.

The second parameter is the utilization rate. The utilization rates of data centers built around the world vary greatly. For data centers with decent PUE and low utilization rate, idle resources can be connected to resource-sharing markets (such as MESON) to acquire DCCC and additional benefits(Revenue, e.g., Tesla earned around $1.6B through selling the carbon credit) can be obtained by selling DCCC. Data centers with poor PUE need to purchase DCCC from the market or share idle resources to offset the difference in DCCC.

In addition to the two parameters mentioned above, another important parameter is the energy structure supplied to the data center. In the future development of DCCC, the energy structure of the power supply could also become a key parameter.

DCCC Use Case

MESON will become the first market to support the use of DCCC. MESON attempts to solve the problem of idle bandwidth resources by building a bandwidth aggregation and transaction market from the bottom up.

Nodes that meet the PUE standard can contribute resources to the MESON network to obtain a certain amount of DCCC, and nodes that do not meet the PUE standard can contribute resources to the MESON network to obtain the DCCC balance.

DCCC will become an essential parameter of trading resources in MESON. Nodes that fulfill the DCCC standard get priority in the market of resource replacement. Nodes that do not fulfill DCCC standard have limited income and ranking of transactions. We hope to create a platform for the world and hope that this platform can contribute to the sustainability of mankind.

Pay by the demand side or the supply side?

The debate on whether the Supply-side or Demand side should pay carbon emissions has been ongoing. The question such as: should airline companies pay for carbon emissions or passengers? Should the power plant pay for carbon emissions or the electricity user; Should Bitcoin miners pay for the carbon emission or those who use the network? (Bitmex buys carbon credits). We believe that both ends must be responsible. For the demand side, voluntary encouragement is the main focus, and certain mandatory constraints are required for the supply side.

Conclusion

We propose establishing an agreement on carbon emissions trading in the data center dimension, introducing indicators such as PUE, utilization rate, and power supply structure, and using DCCC to make specific measurements. The agreement itself is still in its early stages, and we hope to attract more people with lofty ideals to participate in the discussion and governance of the agreement.

Join the talk: https://github.com/daqnext/DCCC

Appendix

• https://www.tesla.com/ns_videos/2020-tesla-impact-report.pdf

• https://blog.bitmex.com/bitcoins-carbon-footprint

• https://blog.sintef.com/sintefenergy/this-is-how-we-reduce-data-centers-carbon-footprint/

• Bill Gates:《How to Avoid a Climate Disaster》

• https://www.mdpi.com/1996-1073/10/10/1470

• https://www.nature.com/articles/d41586-018-06610-y