How to build a reliable decentralized storage system (2/2)
The key to decentralized storage
The key to decentralized storage lies in 2 points: 1. To unleash the full potential and advantages of decentralized storage; 2. Make up for the disadvantages of decentralized storage.
How can we achieve them?
1、QoS(Quality of Service)
QoS is about provisioning network services to the level of guarantee required by the application layer services.
QoS mechanisms provide means for ensuring network resources required by applications when delivered to achieve the expected level of user QoE(Quality of Experience).
On the basis of good QoS, quality of experience (QoE) of DApp based on decentralized storage is also very important.
QoE(Quality of Experience)
The degree of delight or annoyance of the user results from the fulfillment of his or her expectations with respect to the utility and / or enjoyment of the application or service in the light of the user’s personality and current state.
What are the specific metrics of QoS and QoE? As mentioned earlier, decentralized storage has the advantage of being cheap enough, fast enough, and secure enough, and QoS is to quantify and optimize these metrics until they eventually go far beyond the QoS of centralized cloud storage.
Existing decentralized storage chain projects put the main focus on how to be decentralized, how to reach consensus, how to do rigorous mathematical proof, so as to prevent cheating, etc., (especially FileCoin, almost 80% of the content of it’s white paper is on mathematics), completely ignore the importance of QoS.
However, QoS cannot be done from the very beginning. It needs to be gradually optimized and improved step by step in the process of project development. So early on we had to make sure that the project could be updated flexibly, because we had to try to calibrate and adjust the algorithm.
- Excellent economic model
As mentioned earlier, one of the difficulties and challenges of decentralized storage is stability. When the storage provider (miners) is unstable, it could lead to file loss, which reduces the availability of the entire platform (like Amazon S3 SLA, also one of the QoS indicators). A good economic incentive model can encourage more stable storage suppliers (miners) and punish or even eliminate unstable miners. As long as storage suppliers (miners) are stable, the availability of services (SLA) remains very high.
Not only affecting the stability of storage providers (miners), economic models also encourage new storage providers (miners) to join or participate in solving the aforementioned demand and supply balancing problem of the sharing economy.
The economic model is the same as the quality of service (QoS). It is not done right from the beginning. It needs to be gradually optimized during the development of the project and changed step by step. Thus, in the early days we had to ensure that the project could be upgraded flexibly because it is necessary to try to adjust the economic incentive strategy.
- Early anti-cheating is difficult
Storage proofs are less mathematically rigorous than bitcoin’s mining and hash calculation. Only the proof of the idling hard disk of the BurstCoin is very strict, but the BurstCoin consumes hard disk space with unmeaningful hash data and it is unable to provide storage and upload services. Storage proofs such as PoRep(copy proof) and PoSt(space-time proof) introduced by FileCoin in the white paper are mathematically not as simple and rigorous as Bitcoin, but are very complex and undermined by low-performance. And subjected to the Impossible Triangles in the blockchain, that is, security and performance are at odds. If behaviors need to be rigorously proved, performance will be very low and the comparative advantage of centralized storage will be lost. The source code of the FileCoin has not been opened for so long, partially due to the rigor of the proof.
Therefore, it is difficult to consider the rigor of cheating in the early days. It is proved that the method of spot check and miner mortgage punishment should be adopted, but there is no guarantee that the hacker cannot find loopholes to take advantages of. In order to reduce such risks, the most effective way to achieve service with some super nodes, then realize full decentralization service. Service with super nodes refers to the way in which the supervising organization of the certification adopts the mode of alliance. There’s threshold to enter and obtain permission, and the nodes guarantee not to do evil. This can greatly simplify the difficulty of authenticity. This is more conducive to ensuring the performance of the system. We have an idea that to keep the code of the proof algorithm not opening source for a period of time in the early stage. After verification by the market and a period of time after the algorithm matures, it will gradually open source and eventually be fully open source. This is also an effective method to reduce early risks. It is believed that Filecoin will probably adopt this scheme.
……
- Flexible and easy to upgrade architecture and governance mechanisms are very important
Since both quality of service (QoS) and economic models require an iterative process, it is important to build a flexible and easily scalable architecture.
Decentralized storage is not used for projects such as Bitcoin, Ethereum, etc., because they are projects that write all information to the blockchain. As to the decentralized storage, only incentives and contracts need to be written to the blockchain, and the actual stored files do not need to be written to the blockchain. As a result, there is no need to be completely decentralized in one step.
I think the appropriate path to decentralized storage is: Firstly, to build a centralized network, then multi-center, and finally a decentralized network. In phase one, services other than users and storage providers (miners) are central, such as scheduling and indexing, similar to Bittorrent’s Tracker, which allows rapid iteration of P2P algorithms and economic model policies. After a period of verification, it can become multi-center, inviting honest and powerful nodes to join and become an alliance chain. Finally, when the algorithm is proved to be efficient and safe, it goes from the super node mode of the alliance to the decentralized stage where everyone can participate.
Because of the above considerations:
Based on the above considerations, I found Bill and started up the PPIO decentralized storage public chain project together. Our goal is to develop a decentralized storage public chain that takes the quality of service as the primary factor. I will continue to write articles about the public chain of PPIO to explain the characteristics and principles of PPIO. Please look forward to it.
Article author：Wayne Wong
If you want to reprint, please indicate the source
If you have an exchange about blockchain learning, you can contact me in the following ways:
Github: https://github.com/omnigeeker
Telegram: @omnigeeker
Twitter: @omnigeeker
Medium: https://medium.com/@omnigeeker
Steemit: https://steemit.com/@omnigeeker