The technology industry has made rapid progress and development in the past year, and more and more entrepreneurs, aiming at the data storage industry with huge development potential, have seen an unprecedented rise in market interest in the data storage industry.
Due to the strong demand, semiconductor prices have been soaring.
Server prices continue to be high and a number of projects have already turned their attention to decentralized storage protocols. However, the blockchain storage protocols that exist today are all more or less problematic, and being able to store and read data safely and quickly is still a pressing issue for many developers and users.
Problems with Distributed Storage Protocols.
Back around 2018, blockchain technology ignited the market and became all the rage, with all walks of life starting to use blockchain decentralized peer-to-peer technology for industrial upgrading. Now, three years on, there are not many blockchain projects that actually have application scenarios that can be put into practice.
Refined to the data storage industry, the storage protocols that are currently widely applied in the market mainly have the following problems.
1. Long data transformation time for the purpose of block generation.
This type of protocol is represented by Filecoin. The data stored in Filecoin nodes is all for the purpose of block generation. The larger the amount of data stored, the greater the arithmetic power, and the greater the arithmetic power, the greater the chance of block generation.
However, arithmetic power cannot be created and disappeared out of thin air, so when users store data and read data on these products, it creates a long data transformation time and the speed of storing and reading data is very slow.
What’s more, Filecoin itself is not responsible for the stored data, the data storage is all the responsibility of the miner’s node, Filecoin does not provide any system-level guarantee function, if the miner’s node fails, then the data is lost.
2. Pseudo-decentralized protocols with centralized nodes
This type of protocol is represented by Storj. In this type of distributed storage project, a node or nodes take on the function of a “centralized node”, which links all the data, and the data stored by the user has to be sent to the “centralized node” first. The data stored by the user has to be sent to the ‘centralized node’ for validation before it is stored in the other nodes.
When the “centralized node” is overloaded, the data storage speed and reading speed will become slower and the service will not be able to keep up with the requirements. When the “centralized node” is compromised, the source data is also lost and the data security issue is not well addressed.
3、Inefficient operation and high cost of redundancy
For a blockchain used for data storage, any data written to it requires a consensus process, which can cause a pause of a few seconds or minutes. And even if smart contracts are used to define a flexible data structure, the data structure becomes homogeneous and retrieval is limited, preventing updates to the data model and data size. And any data written to the chain would be synchronized across thousands of nodes, imposing a very high cost of redundancy and leaving the blockchain with much less performance and capacity.
The new and upgraded decentralized storage protocol MEMO
MEMO, whose full name is memoriae, is a project that truly uses blockchain peer-to-peer technology for data storage and is dedicated to creating a new generation of secure, efficient, large-scale decentralized storage system based on blockchain.
Through a three-year-long research and innovation, MEMO proposes a new decentralized storage concept that maximizes the three problems mentioned above and can achieve low-cost simple, secure and efficient data storage functions.
MEMO achieves decentralized peer-to-peer storage through the following two main innovations
1. Decentralzed distribution of three roles, subject to smart contracts
The reason why MEMO is truly decentralized is that there are three main roles in its system: the storage demander “User”, the storage coordinator “Keeper” and the storage provider “Provider”. Provider”, in which all three roles are subject to a smart contract to store, maintain and read data.
The source data uploaded by the User is stored in the smart contract in the main chain, and the User will find a suitable Provider with the help of the Keeper, who will provide the space and time to store the data. The repair function will be activated in case of data loss or corruption. In this application scenario, the User pays the Provider for storing and reading the data, and the Keeper for coordinating the management.
The three types of roles in the MEMO system, under the constraints of smart contracts, constitute a three-legged battle, with no real centralized information processing node, in terms of data privacy, data security, data stability, etc.
2. Innovations in corrective code and copy redundancy
Most of the existing distributed storage has the common problem of high redundancy cost. Although the corrective code can reduce the cost of redundancy, it also leads to excessive consumption of repair functions, some even up to ten times more than the repair overhead of the copy.
In MEMO, only the most critical data in the storage system, such as role information and smart contracts, are stored in the main chain of the blockchain, while other data are placed in ordinary edge storage devices. At the same time, MEMO uses a combination of both corrective coding and multi-copy redundancy, with the default being corrective coding mode, which greatly reduces redundancy costs.
In addition, MEMO also has an original data recovery method, RAFI, which supports snapshot backups of multiple points in time, allowing multiple samples of time nodes to be extracted and recovered at the same time. In other words, even if one of the storage nodes fails, it can be recovered quickly and will not affect the remaining nodes to continue working, ensuring data security and system stability.
Future vision
The MEMO project was created in 2017, and the team hopes to make MEMO a global large-scale storage pool that can hold ZB-class data in a stable and efficient manner. It will be flexibly adjusted according to users’ actual business needs to meet performance and at the same time both cost-effective advantages. It will continue to expand and improve from on-chain data storage, off-chain data storage, IoT data storage, enterprise cloud storage, decentralized cloud storage, etc., with an eye on landing applications and multiparty cooperation.
As Tim Ding, founding partner of MEMO’s investment arm Commando Capital, said, “Most of the storage products on the market today are products that use the principle of distributed technology, while MEMO can achieve true peer-to-peer decentralized storage, and they have made many creations and innovations, such as They are one of the more cutting-edge decentralized storage projects, such as “public verification mechanism” and “data repair mechanism”.”
