Stir-fried DePIN: immature traditional technology + immature blockchain ≠ the myth of low cost!

Interview | Beichen

Guest|Steven

The Web3 industry is in a very embarrassing stage - the market is going downhill with the tight liquidity, and at the same time there is no new technology paradigm, so it starts to stir up cold food, such as DePIN (decentralized infrastructure network) .

In fact, since 2017, various blockchain-based hardware infrastructures have been emerging. Although none of them have succeeded (it can only be said that they have not failed, such as Filecoin and Arweave, there is still hope in the future), but now they are enthusiastic , There is even a popular opinion that DePIN can save more than 75% of the cost.

“Sharp Whistle” released “Why Helium, which is fundamentally ineffective, and Web3 research methods”, was released last week. Taking Helium, a representative project of DePIN, as an example, it analyzed why it did not work from a commercial point of view.

The reasons can be summed up in two points:

1. Helium’s blockchain only plays the role of issuing coins and order settlement, does not need a public chain at all;

2. The IoT services provided by Helium** run counter to the real development trend of the IoT industry** (manufacturers are more inclined to build their own IoT ecosystems).

In this issue, Steven, an expert in communication technology, is invited to explain why blockchain is not suitable for hardware infrastructure from the perspective of communication engineering.

**1. Beichen: **DePIN is very popular now, and I think it can drive a new wave of physical infrastructure. This is not only whether the business logic is established, but is it also technically feasible? Please first make a statement about DePIN from the perspective of communication engineering.

**Steven: **From a technical perspective, DePIN is currently not feasible and usable, there may be one in the future, but it is far away.

2. Beichen: Why do you think the blockchain industry is now talking about DePIN again? In my opinion, isn’t this the “chain reform” narrative that went bankrupt in 2018? My personal prejudice is that if I can’t get along in the traditional industry, I will come to forcibly integrate with the blockchain. **

Steven: DePIN and chain reform are not a concept yet. The idea of DePIN was actually proposed before the blockchain appeared (for example, some people in the communication industry have made decentralized Wi-Fi networks, such as CDN), but now there is a district The blockchain is relatively easy to finance and has strong financial attributes, so it is naturally combined with the blockchain, which is today’s DePIN, but the current technology cannot support the development of DePIN.

3.** Beichen: **Because there are so many tracks involved in the DePIN narrative, let’s start with the DeWi (decentralized wireless network) that you are most familiar with. Take Helium as an example, is its LoRaWAN and 5G network actually competitive?

Steven: As far as communication technology is concerned, the IoT application scenarios that LoRa can meet are very limited, and Helium also combines blockchain, making it very expensive and not suitable for commercial use at all. Moreover, Helium’s 5G is only a supplement to telecom operators, and it is essentially a 5G subcontractor, which does not make much sense.

Although Helium has officially boasted many application scenarios, the IoT scenarios that LoRa can satisfy are very limited.

According to the coverage area, the wireless physical layer of the Internet of Things can be divided into two categories: long-distance (hundreds of meters to more than ten kilometers) coverage and short-distance (within hundreds of meters) coverage. The Wi-Fi and Bluetooth we are very familiar with are short-distance coverage, and the LoRa used by Helium belongs to the LPWAN low-power wide-area network. This is a so-called ultra-narrowband Internet of Things protocol, which requires very low network bandwidth, and the cost is that the amount of data transmission is also very small, which determines that it is only suitable for application scenarios with light nodes and light transmission, such as thermometers, Sensors such as water meters only need to report data once in a few minutes, and there are only dozens of bytes.

And because LoRa is an unlicensed frequency band for free use, the frequency bands allocated by different countries are different. In Europe and the United States, it is the golden frequency band of 433, 868, and 915MHz (the wavelength of this frequency band has a certain penetration ability, and the speed is also OK, which is the best. Voice service and low-to-medium speed data service frequency band), in China it is 470~510MHz.

(**Beichen: **Does the Helium mining machine also need to be adjusted? Steven: The frequency band and transmission power must be adjusted accordingly, otherwise the signal will be interfered.)

Although LoRa is an unlicensed frequency band, each enterprise can build a LoRa network by itself (just like building Wi-Fi at home), but the network capacity of each frequency band is limited. For example, 5G can theoretically support each A square kilometer is connected to 1 million devices, and we have previously estimated that there are only a few hundred to 2,000 LoRa for actual use and planning.

Helium lets gateways (miners) do proof-of-work based on coverage, which means that it doesn’t care about the actual network requirements in different places, as long as it provides network coverage as much as possible, it can be rewarded (even in no-man’s land). The normal Internet of Things must optimize node distribution based on actual needs.

The Internet of Things requires operation and maintenance, and once a problem occurs, it must be restored in a short time. The normal Internet of Things spends a lot of resources on dedicated maintenance, but it is difficult for the miners of **Helium to do professional network maintenance, not to mention that these mining machines are so scattered. **

4.** Beichen: **After Helium combined the blockchain, they claimed to reduce the cost. Why do you think it increased the cost instead?

Steven: Regarding the cost advantage of Helium, I absolutely disagree.

**First, the cost of Helium network construction is very high. **I disassembled a Helium mining machine before. It was quite expensive at that time, close to 10,000 yuan. I unscrewed the screw and took a look. Counting the cost of the chip, in fact, the same module can be obtained in the market for more than 100 yuan. Yes, but they use a so-called special chip (add an ID to the chip for anti-counterfeiting).

**Second, the structure of the distributed network determines that the cost of transmitting data will be higher. **Each mining machine of Helium must firstly have Internet access equipment, and secondly need monitoring equipment and capabilities. Although the traditional Internet of Things also needs Internet interfaces and monitoring, the Internet of Things and the Internet are strictly isolated, with higher security, and do not require a lot of equipment and monitoring capabilities (for example, Alibaba Cloud only needs a few hundreds of thousands of sensors). Server and platform services are enough), and the cost will be shared equally.

The operation of the LoRa network originally does not require a blockchain. Helium is out of financing considerations. In order to move closer to the blockchain, Helium forcibly increases the blockchain verification, which increases the cost.

In fact, regardless of the blockchain part, if you directly look at the core business of Helium, you will find that it is completely a centralized project, but the transactions that can be completed automatically are placed on the blockchain. The chain is completely a pseudo-requirement for Helium.

5.** Beichen: **If Helium’s approach doesn’t work, what kind of solution does the Internet of Things really need?

**Steven: **The Internet of Things of production-oriented enterprises (especially streaming production and large-scale discrete production) has very high requirements for the reliability of information transmission, and they are more inclined to build their own networks, and they will also do electronic fences and the like Isolation to ensure that all data is only transmitted in the private network, only those with a relatively low security level will use IoT cloud vendor services such as AWS and Alibaba Cloud or NBIOT, and it is even more impossible to casually use **Helium Third-party LoRa networks, even simple infrastructure such as street light control are afraid to use it. **

5G is an important direction of the Internet of Things, it can satisfy a large number of sensors or controllers in a small area, while the LoRa network is not suitable for production enterprises and infrastructure, not to mention that Helium is still more expensive, Less efficient distributed architecture.

6.** Beichen: **In addition, there are distributed storage and distributed computing power concepts. Storage and computing power are indeed in demand in the market, so do they really have advantages compared with centralized services?

**Steven: **Whether the storage and computing done by distributed systems have advantages depends entirely on what problems they are trying to solve-Distributed systems will inevitably reduce efficiency and increase costs, but if The business itself requires multi-node connections (for example, higher requirements for redundancy and reliability, or the need to avoid some legal restrictions), so it is suitable for distributed systems.

Therefore, there is a need for distributed storage and distributed computing power, which is beyond doubt. But how big is this need? And whether it can replace the tasks of the centralized system, this is a question mark.

Let’s take Filecoin as an example. This is a very good project, which solves many technical problems, especially proof of storage and proof of time and space. To be honest, it is not easy to develop it, but the most basic economic model of **Filecoin is unreasonable, and has triggered a series of problems. unreasonable. **

First of all, in Filecoin’s economic model, FIL is a functional token, and its price fluctuation will seriously affect the actual storage demand. In theory, the incentives of ecological participants should not be affected by the token market.

The sunk cost of this economic model makes people who mine FIL dare not use hard drives to mine at all, because if something goes wrong (power outage, network outage, etc.), instead of making money, they will lose the mortgaged FIL. This is the same as Filecoin’s original The concept of “utilizing redundant storage space” is contrary. Moreover, Filecoin’s mining machine also needs to add a graphics card to verify the validity of the storage (replication proof), and you will find that its actual storage cost is more expensive than cloud storage, not to mention the hard disk. **

The high cost has further led to insufficient availability. No enterprise is willing to store core data in the Filecoin network. This is not a problem that blockchain incentives can solve.

7.** Beichen: **So in your opinion, what technical problems need to be solved for a mature “decentralized network architecture based on blockchain technology”? Or is it simply impossible to solve?

**Steven: **DePIN is a decentralized network architecture based on blockchain technology, and **blockchain adopts the triple-entry bookkeeping method, which determines its cost to form a consensus among multiple nodes And efficiency is not as good as traditional digital systems. **

If DePIN is to develop in an ideal direction, three problems need to be solved.

**First, the reliability of infrastructure. **Distributed networks have higher requirements for node maintenance capabilities than centralized systems, and require miners to have strong professional maintenance capabilities. The current hardware equipment and software maturity are far from enough.

**Second, decoupling between software and hardware. **The application dominated by centralized big manufacturers is to bundle software and hardware (such as Mac system). Only when there is sufficient decoupling between software and hardware, when software can define everything, DePIN’s decentralized distribution The ideal of a distributed deployment can only be realized.

**Third, to realize the credible mapping of the whole **** world data. **A reliable connection must be established between the data world and the physical world. For the blockchain, the oracle machine must be very mature.

(**Beichen:**This is also the basis for the realization of the RWA concept. **Steven:**It should be said that it is the basis for whether the blockchain has a future. The centralized system is maintained by the social contract between people, and the district The block chain is realized by code and probability game, if the source data is unreliable, where will all the prospects come from)

Moreover, in addition to the technical bottlenecks that need to be overcome, DePIN also needs to solve community efficiency, which is a big problem.

Take 5G as an example. The latest technical standard Release 18 has 35 million words (Chinese version) in the text alone, and then different manufacturers have derived more product standards based on it. From R&D to operation, this is a very large division of labor.

In terms of the efficiency of the current DePIN communities, it is impossible to complete. You can only wait for the centralized organization to finish all these things and feed it to you. So how meaningful is decentralization at this time? It can only be said that the threshold for starting in the blockchain world is lower, and many things cannot be started in the traditional world, so it is a good thing to come to the blockchain world to try.

8.** Beichen:** What you said really inspired me. DePIN is doing things in other fields under the chain, so the community can only introduce technologies from the external centralized world, only DeFi Only in a crypto-native field can new technologies emerge in the community.

**Steven:**First of all, DeFi is related to finance and naturally fits the technical characteristics of blockchain. Secondly, DeFi uses code to implement projects, and DePIN needs to interact with complex physical devices, and the difficulty is not in the same order of magnitude.

**9. Beichen: **Finally, what do you think of the cycle of DePIN narrative? In other words, how long do you think it will be proven unfeasible by the market?

**Steven: **On the technical level, neither the blockchain technology nor the traditional technology involved in DePIN is mature enough, using these technologies to carry out such a grand vision project , To be honest, there is almost no hope. Only a team with real technical strength like Filecoin will see hope after continuing to work for five to ten years.