The EOS blockchain went live on June 14, 340 days after ICO, and as it grows, it will naturally face a lot of problems. This Ethereum killer guarantees zero commission and high scalability, allowing thousands of decentralized applications (DApps) to co-exist on the platform. Because of this, EOS won over its overloaded, expensive and slow predecessors.

However, the developers of the first Dapps on EOS were concerned with a different issue: if Ethereum was expensive for application users (who had to pay with gas for any action in it or for another DApp determined by a smart contract and required computation), then EOS was expensive for developers.

One Reddit user wrote: “EOS makes developers bear the cost of transaction and storage. ETH puts the cost on the user.”

According to the CEO of PandaFun, which recently launched on EOS, the cost of resources required for a DApp on EOS is estimated at 21,000 EOS ($122,000 at current exchange rates). In the same case, a smart contract to deploy a DApp on Ethereum would cost $100.

How did the price come about?

At its core, EOS is a “proprietary” network architecture. This is determined by DPoS, the delegated proof-of-stake or Delegated proof of the role-share algorithm. If the equity of a valid proof validator can become any participant in the network, and will create the block probability and encryption in the account currency is proportional to the number, that is, in the network of its equity/share), in the authorized version, the validator producers (blocks) or the super node is simple “selected” by the user. Thus, the EOS network is supported by an “ever-changing network of 21 supernodes (block producers)”. They share networks that are necessary to validate transactions and generate blocks and “voters”. In other words, where the rules are the same, people with more Internet stakes have more power.

The “sharing” mechanism, on the other hand, provides “day-to-day” work for EOS, allowing users (and developers) to open accounts, trade and create smart contracts, while also protecting the network from malicious attacks. This mechanism is called stacking. In fact, it makes possible the disappearance of commissions, because in Bitcoin and Ethernet, all of these operations (except creating accounts, which are free) are “sponsored” by commissions.

Stacking is one of the resources required to “lock up” funds over a period of time in exchange for work on the network.

These resources are listed below:

• Network traffic (NET). Average NET consumption is the number of bytes you have used in the last three days. NET is consumed every time you perform an operation on the blockchain, such as sending a transaction. The more tokens you keep as network equity, the more NET you get to work with.
• Processor time, or computing power (CPU). This is how long it takes the CPU to perform a given operation. Average CPU consumption is the number of milliseconds you spent in the last 3 days. Processor time is also used to execute each operation on the blockchain. The longer the processing takes, the more CPU time it consumes.

These resources are allocated in proportion to the number of tokens you contribute to the 3-day stacked contract. At the beginning of the stack, you specify which should be used for purchasing CPU and which should be used for NET. You can then add funds to the contract or exchange the available resources you gave up for EOS tokens. This means that you don’t lose your funds when stacking: in the contract they are used up, but after 3 days the cost of the contract in EOS is returned to the original Indicator. The only thing that will change is the amount of equivalent dollars.

The essence of the economics of stacking is to ensure that you will not use “secured” tokens for as long as the contract exists. That is, you keep tokens during periods of inflation, where the block producers create new tokens and treat them as rewards for themselves. In this way, you pay for nodes that process transactions and provide processing power.

• Operating memory (RAM). We will discuss it separately because of its acquisition method and mortgage difference, buy it on the internal RAM market, the price adjusts automatically according to supply and demand situation. Storing data on a blockchain requires RAM, which means you pay for a certain amount of memory capacity. Unneeded memory capacity can be sold for EOS tokens at the current conversion price. The amount of RAM is limited (currently 72GB, of which 62% is used, or 44GB is used), but the amount can be increased if necessary. Thus, as soon as the supernodes (block producers) increase the amount of memory once the main network is activated, speculators start buying up RAM and selling it later at a higher price. This pushed the price to 0.94 EOS tokens per KB, nine times higher than the current price. It was decided to double the RAM rollout, adding 64GB per year at a rate of 1KB per block. This step has cooled the market for operating memory

On an industry scale

Developing EOS applications and training new (roughly speaking, attracting) users to the project will require a significant amount of the three listed resources and, therefore, a significant amount of capital. Even taking into account payNET and CPU tokens, in fact, they are not wasted, they should always be “reserved”.

Kevin Rose, co-founder of EOS’s New York block, the Company-Manufacturer, said he was in discussions with a group of developers who wanted to use EOS instead of its existing platform.

Tixico Event platform announces transition from Ethereum to EOS: EOS promises to be scalable enough, even with millions of users, which is critical to designing a platform to maintain large numbers of people at once, especially when ticketing starts and tens of thousands of people enter and trade at the same time.” Among other advantages, Tixico also points out avoiding commissions.

However, these teams have to pay a lot of money for scalability. PandaFun’s CEO has said that about 21,000 EOS tokens were spent on the development of its app, and he also mentioned the distribution of tokens on resources: For example, 10,000 EOS tokens (about $58,000 per hour) were spent on RAM, the same amount was spent on CPU, and 1,000 EOS ($58,000 per hour) were spent on NET. However, the CEO of PandaFund also points out that most of the RAM required is for tokens that will be acquired, and much less is used for the whole thing itself.

On average, each user needs 4KB of RAM to create an account (about $2.70 at current RAM prices). However, many other operations also require RAM.

But in June, when it was cheaper to create accounts (between $0.50 and $1), participants on GitHub had noticed that the RAM model “just doesn’t work because if your goal is to create tens of millions or hundreds of millions of user accounts for your DApp!”

“EOS lets developers pay transaction and storage costs. Ethereum lets users pay.”

Who pays?

The cost of resources could then become an issue for both developers and users of EOS applications.

Thomas Cox, former vice president of product at Block.one, said there could be a usage scenario where developers write dApps and users have to have their own CPU and/or NET and/or RAM to interact with. He also points out that this is one way to write an early version of your DApp so that if the app suddenly becomes popular, you won’t go broke.

Introducing a special commission for users has been one of the main suggestions in the Reddit discussion about how developers can afford RAM, CPU, and NET.

One user, nicknamed MR1ply, wrote: “They can easily charge for using their DApp. Both EOS tokens and their own tokens. The money will go directly to the developer of the DApp.”

SuddenAnalysis, another participant in the discussion, noted that while many apps will clearly start charging, others “will have an inflationary model where developers receive some or all of the interest from inflation to continue paying for resources that are entirely dependent on the value of their platform.”

Ablejoseph writes: “Inflation! You need to design your crypto economics in such a way that inflation covers development costs.”

At the same time, some users argue that getting resources (CPU and NET) by way of collateral is not a big problem for developers:

“If developers have pledged a bunch of EOS tokens, they really don’t have to worry about increased fees. The tokens themselves grew in size as the network expanded. If they have to buy more EOS tokens, then they should be able to sell their DApp tokens in exchange for more EOS tokens for CPU and bandwidth, “wrote another participant in the discussion, adding that” each DApp must have a well-thought-out economic model for its own maintenance and expansion “.

It’s too early to predict costs, but: Right now, the network is just beginning to attract users and developers. According to Dapp Radar, only six EOS apps currently have more than 100 daily visits, and only two have more than 1,000 daily visits. Thus, the cost of CPU time is small, with only 0.00049,966 EOS tokens (equivalent to $0.003) per KB.

In a recent blog post on Reddit, users of the second most popular EOS app (EOS Knight Games) have drawn attention to the unattainable amount of required mortgages for CPUS. A user named AGameDeveloper said he used 10 EOS, equivalent to $59, as collateral, but not enough. According to EOS Knights, a user’s CPU share to start the game is at least 15 EOS (equivalent to $88), but AGameDeveloper says that in reality, a minimum of $500 worth of tokens must be mortgaged.

One user commented on the blog post: “CPU utilization is currently 8%, so figure out a 100% equity price.”

Other problems

In August, hackers used notifications to raid someone’s account, using useless data to take up RAM space. One solution to this problem was proposed by EOS Technical Director Dan Larimer, but it was within a vague and unapproved framework to completely rewrite the EOS “constitution”. Another solution, posted on GitHub, involves sending tokens through proxy smart contracts that do not contain RAM. In any case, the solution to this problem is really the responsibility of the developer of each application.

In addition, the system that was designed to protect against attacks was also attacked because the low price of cpus today makes it feasible and less expensive. The owners of Blocktwitter accounts are happy to “send messages that comprise 192 million actions, which currently account for about 95% of all EOS transactions,” said Tom Fu, a partner at GenerEOS. In particular, all the messages contain a record: “WE LOVE BM” (BM refers to Larimer’s nickname bytemaster). According to Fu, the message is not “significant” but has a negative impact on the network because Blocktwittter owns a large share of the CPU, resulting in less CPU time allocated to other users and developers.

The bright side

In early August, Larimer proposed a rental model for CPU and NET that “will reduce the cost of using the EOS network.”

At the same time, Cox pointed to several distinct advantages of EOS that set the young platform apart from Ethereum. First, the new “crypto-Kitties” won’t stop EOS in its tracks, as it happens with Ethereum: ensuring that the network’s performance will be supported by a certain number of reserved cpus. Second, two of EOS’s three resources (CPU and NET) are “updatable” (unlike gas in Ethereum). That is, after the 3-day mortgage smart contract expires, the funds are released and can be used again for the same (or other) purpose. Unused or freed RAM can also be sold, but perhaps at a lower (or higher) price. In addition, the arbitration system that supports EOS security was considered sufficient to withstand attacks on THE DAO and Parity scenarios, and the developers of Ethereum applications were in “bankrupt error,” according to Cox.

This article is from infoQ by Alex Vet and translated by Yao Jialing with authorization.