The proliferation of wireless broadband and fast networks is driving data into the cloud, decoupling it from specific physical devices and ushering in an era of data access through any screen. The cloud makes these data easily accessible and as a result the need for local storage is increasingly diminished.
Centralized cloud storage systems provided by authoritative and reliable third party organizations are dedicated to meeting these data storage needs. However, that raises the issue of data privacy. For example, do you feel comfortable handing over important data or even data of vital interests to somebody else for its safekeeping? Will they store your data as per your requirements? Can you be assured that your data will be retrievable in its entirety without compromise when you need it, and that your service provider will not expose it to anyone without your permission? Those are all the questions that must be answered before you take the next step.
If the above are your concerns, then the now-popular centralized storage system may not fit you well anymore. You may ask: Is there any alternative secure data storage solution for me? The answer to this question is the decentralized cloud storage (DCS) system. DCS is not a new term, in fact, it was available on the market even during the early days of the digital age. Those who have experienced one or several of those DCS products may have a general idea of what they are. The core idea of decentralized storage consists in taking advantage of decentralized networks known for the improving of privacy, security, and availability. Given the lack of adequate I/O speed and decentralization degree concerning the existing DCS products, MEMO (short for Memoriae) is designed with higher service capabilities through the improvement of the relevant technologies.
As a completely new Blockchain-Enabled Decentralized Cloud Storage system, MEMO is aimed at integrating low-cost edge storage devices by mobilizing idle resources fully and effectively so that they can be used by those in need worldwide. The testing network of MEMO has been up and running for about a month, and its volunteer recruitment plan has also been launched recently. Your participation will be most welcome. We will keep you informed of the architecture and operation of the MEMO system through articles, community interactions, and video courses.
The details of the MEMO project are described below.
What Is MEMO?
MEMO, short for Memoriae, is an open-source blockchain-enabled decentralized cloud storage system developed by Memolabs with a devoted R&D team of well-known experts, engineers, and PhDs in the storage field around the world with multiple exclusive patents. Currently distributed all over the world, all team members have years of Internet and development experience in the blockchain industry, and they have designed MEMO to provide a completely new storage network that can meet the demand for more security, reliability, and cost-effectiveness.
Three System Roles in MEMO
● User: The User of the storage service in MEMO. The User configures the smart contract parameters and indicates the desired service level based on which the appropriate Providers are selected. The User can generate a smart contract, upload the data to MEMO and use MEMO tokens to pay for the services. The User can also download and manage data stored in MEMO.
● The Provider: The Provider of edge storage devices in MEMO. The Provider uses the smart contract parameters to verify its service capabilities, store uploaded user data according to the smart contract, executes and responds to user data access and management requests, responds to challenges from the Keepers, cooperates with the Keepers to repair or recover data and receives MEMO tokens as income for the data storage services provided.
● The Keeper: They are the guardian of data security, reliability and availability in MEMO. The Keeper is responsible for preserving data mapping information and reaching a consensus. The keeper uses the smart contract parameters to validate its upkeep capabilities, challenge the Provider, reaches a consensus on both the Provider’s reliability&availability, and on when to repair/recover damaged/lost data. Besides, the Keeper validates, confirms the smart contracts between the User and Provider, validates&transacts on the blockchain, calculates&reaches consensus on the storage space and time utilization, and receive MEMO tokens as payment.
In the MEMO ecosystem, the three system roles interact with each other to guarantee the security of data storage–the User puts forward the storage requirements, the Keeper finds the suitable storage nodes for the User, and the Provider stores the data as the storage node for the User.
Technology Breakthroughs Brief Overview
● Security: Hidden flow of cryptocurrencies and increased transaction anonymity; encrypted data in accordance with security requirements from the source; controlled data access throughout the process.
● Reliability: appropriate redundancy mechanisms by design; guaranteed availability of data service with reduced redundancy overhead; further reduced risk of data loss and service interruption through repair scheduling technology.
● Affordability: reduced construction costs by reusing edge resources; reduced operating costs by decentralizing operation and maintenance; storage services provided at affordable costs.
How does MEMO achieve a truly decentralized cloud storage based on blockchain?
Compared with the existing DCS product offerings, MEMO can realize the true peer-to-peer decentralized cloud storage mode through its many innovative solutions to avoid centralized management nodes. Decentralized cloud storage systems make full use of the storage nodes scattered on the Internet to provide the necessary storage space. Since the storage nodes may delete the data, go offline, or experience hard disk failures, effective measures must be taken to prevent active malicious behaviors on the storage nodes, and to detect occasional data loss in time to complete timely data repair. Therefore, economic mechanisms and verification schemes are designed to ensure that the nodes actually store the data the way they are supposed to.
To ensure data reliability, MEMO incorporates BLS signature aggregation, vector commitment, and probability sampling mechanisms to design an efficient data integrity verification scheme. This scheme only requires a constant-level communication overhead to verify the data with a low computing cost, thus cost-effectively preventing the storage node from discarding or tampering with the data.
While the pursuit of the highest possible level of blockchain security and reliability remains our top priority, we are also mindful of the enormous pressure it brings to MEMO. To resolve the potential bottleneck issue of performance and system cost, we have designed MEMO to record only the most crucial information–e.g. role information and smart contracts–on the blockchain due to its proven security, availability, and reliability, and to store all other data on the edge devices.
Moreover, to strike a balance between reliability, availability, and traffic repair consumption in the decentralized cloud storage system, MEMO has developed a Risk-Aware Failure Identification Strategy, referred to as RAFI. RAFI is more efficient in addressing security and reliability risks arising from the unpredictability and unreliability of edge devices. The key in RAFI is to use different confirmation time for failed data blocks in strips at different risk levels so that high-risk strips containing multiple failed data blocks will be allotted a shorter failure confirmation time, and low-risk strips containing a small number of failed data blocks will be allotted a longer failure confirmation time.
Specifically, in the erasure coding system, RAFI accelerates the identification of failed data blocks in high-risk strips, as well as the repair of high-risk strips, thereby improving data reliability and availability; it also delays the identification of failed data blocks in low-risk strips as well as the repair of low-risk strips, thereby reducing the traffic in network transmission arising from unnecessary repairs, thus improving the system serviceability. In the process of repairing high-risk strips, the confirmation of invalid data blocks in the strips takes priority, so that RAFI can effectively speed up the repair of high-risk strips. As the strips containing a small number of missing data blocks make up the vast majority of all the strips that require repair, delaying the confirmation of the failed data blocks in the low-risk strips can effectively reduce the bandwidth consumption arising from the repair.
Its most noteworthy innovation lies in the improvement of erasure coding and copy redundancy. Most existing DCS systems share the common problems of high cost from erasure coding and copy redundancy. Although erasure coding can reduce redundancy cost, it also leads to excessive consumption of the repair function, with some as high as almost 10 times the cost of multi-copy repairing.
It is crucial to control the level of redundancy so as to minimize the cost arising from excessive redundancy. With the pros and cons of multi copies and erasure coding fault tolerance in mind, we adopt a design strategy based on multi-copy fault tolerance with high redundancy in the MEMO ecosystem, which is supplemented by erasure coding fault tolerance with low redundancy.
Different fault-tolerance methods are employed based on the volume of data in MEMO– erasure coding is used to the greatest extent possible for large-volume data, while multi-copy is used for small-volume data, which can minimize system redundancy. Therefore, the User can choose different fault tolerance methods based on their specific needs. In the absence of a specific fault tolerance mode chosen by the user, the data is processed by default through erasure coding as the fault tolerance mode to reduce the storage overhead incurred by the Provider.
Our Value Proposition
With our independently-owned intellectual property rights, MEMO is committed to developing the next generation of blockchain-enabled decentralized cloud storage system by integrating globally available low-cost and easily-overlooked storage resources to build a safe and reliable high-quality storage ecosystem. MEMO’s mission is to develop a ZB-level data storage system to outlast human civilization!
Official Website: http://www.memolabs.org
Company Name: Memolabs Ltd.
Disclaimer: The views, suggestions, and opinions expressed here are the sole responsibility of the experts. No Bengaluru Bytes journalist was involved in the writing and production of this article.