A few years ago, in the development of cloud computing, Viktor Charypar, technical director of British digital consulting firm Red Badger, wrote on Venture Beat that cloud services will usher in the end, and pointed out that peer-to-peer network will be future development.
He cited the poor performance of cloud computing, the fragility of centralized storage and data trust. He is very confident that most of the problems mentioned above can be solved. Fully distributed, permanently redundant data storage and peer-to-peer network technology are Our inevitable development direction.
The subsequent bitcoin demonstrated the success of digital currency , but for more professionals, its core charm lies in the design of a distributed computing system incentive system for distributed Web3. 0 lay the foundation stone.
What is distributed computing?
Distributed computing is not a new technology concept, and it has been more than 50 years old. Behind all the important trends of the past 10 years, such as 5G, cloud computing, big data, artificial intelligence, blockchain and so on, there is a common technical foundation, that is, distributed computing.
But it is undeniable that many people are still relatively new to distributed computing. In academic terms, distributed computing is a technology that allows independent computers to collaborate across geographies across a network. It makes these computers appear to work in the same environment; they are a group of networked computers. Work together and perform the same workload. A relatively easy-to-understand explanation is that distributed computing is a computational method that divides a huge amount of operations that would otherwise be required by a supercomputer through a long period of work, and is distributed to many computers for processing. A method of calculating efficiency.
Why do you need distributed computing?
In the article by Viktor Charypar, a very vivid example : in the case of existing centralized servers, most of the web applications and services need to be completed through the central server; for example, you and your friends Share photos, this task needs to be transferred to the centralized task keeper for processing, and then returned to your friend’s mobile phone. If the centralized server happens to have problems at this time, you will find that even if your friends are around, you Still can’t finish sharing photos on your device. In the future, with the development of technology, the user may transmit a 4K high-definition video, so the load of the centralized server will be higher, and the calculation amount will need to be expanded at any time, but the centralized computing power is limited by the device itself. The amount of computing power required to move an unlimited amount as needed. In addition to the limitation of computing power, bandwidth will also be an important constraint for centralized servers. Imagine that many computing tasks are transmitted to the centralized server at the same time, which requires a lot of bandwidth to support network communication.
In addition to the constraints of computing power and bandwidth, there are still serious cost problems in centralizing computing power. Because layout cloud computing requires a lot of manpower and material resources, many companies cannot self-operate. Therefore, the cloud computing market is dominated by a few technology giants who rely on themselves. The highly centralized server resource monopoly raises prices and makes cloud computing users miserable. Compared with centralized cloud computing, distributed computing is controlled by the public. The “server” form does not need to pay for physical space, and all users participate at lower cost, saving costs and making resource services cheaper.
As more and more users learn about the value of data, users are increasingly demanding privacy and security for their own data. Traditional centralized databases are intensive and unique, and are vulnerable to attacks. In contrast, distributed databases controlled by network-wide nodes are secure, and distributed has many “storage locations” even in the face of force majeure, still can maintain database integrity and security. On the other hand, companies and customers who no longer need to centrally store large amounts of customer information will be freed from the risk of losing data.
In the era of 5G accelerated connection, 5G technology will promote a process of economic development from global to vertical local area. Traditionally, data centers or cloud computing have great computing power, while for external or edge, they are only distributed geographical locations. They are only responsible for access, such as the ability to cope with new scenes of the Internet of Things, VR / AR, the successful deployment of new applications brought by autonomous vehicles, will require powerful distributed edge computing, sensors and other resources to promote new business applications and improve user experience.
The status quo of distributed computing development?
Since the 1990s, people have been trying to build distributed computing networks; in 1996, the Internet Mason Prime Prime Search Program GIMPS used distributed computing to search for prime numbers. In 1999, Seti used volunteer computing power to search for extraterrestrial life.
Today, with the popularization and development of Internet technology, data transmission solutions centered on network data communication platforms have been gradually formed, and distributed computing has been widely promoted. In recent years, changes have gradually taken shape. The system architecture system with middle ware technology, peer-to-peer technology and grid technology as the core.
Distributed computing brings different effects compared to traditional database storage computing models, especially in the blockchain domain. Distributed computing produces different chemical reactions. Distributed computing infrastructure based on blockchain allows pressing Access, the most competitive computing infrastructure, safely and at low cost. Distributed application DApps can automatically retrieve, find, provide, use, and release all computing resources, such as applications, data, and servers, through a distributed computing platform.
At present, distributed computing based on blockchain is gaining popularity, and many applications using these technologies, such as Golem, iExec, and SONM, have made some progress. Golem hopes to build a decentralized GPU computing resource leasing platform at Ethereum; SONM is building a univer022sal decentralized supercomputer; French blockchain technology company iExec is a provider of all computing resources (computing service providers, data providers) Business, application provider) provides a trusted platform for resource sharing transactions.
By simplifying access to the server, distributed computing greatly reduces the heat loss of the data center while making it easier for data providers and consumers to get the computing resources they need. On the other hand, blockchain-based distributed computing can use contribution proofing protocols, such as providing real time datasets, transferring files, performing calculations, and providing professional services in real time, triggering token transactions between participants, through coding incentives. The system completely revitalized the distributed computing resources.
What is the significance of blockchain for distributed computing power?
Distributed computing power is still in the ascendant, and there is still a lot of room to be explored. However, unlike the traditional centralized data, the centralized computing power can be directly used as a commodity to trade between Party B and Party B. The distributed computing power has multiple computing power providers, and a system is needed to measure and In addition, the current Internet economy is mostly central and group-based development. It does not bring good application scenarios to distributed computing power, and develops appropriate business models to allow distribution. The calculation power is targeted.
The emergence of blockchain solves the above two problems to some extent: the distributed ledger system of blockchain can design an effective and feasible incentive system around distributed computing power, and organize idle computing resources into market operation. So that idle resources participate in mining services and node services to generate market value and receive corresponding rewards.
In addition, the decentralized value network consisting of blockchains will stimulate a large number of decentralized point-to-point value exchange scenarios. Combined with AI, 5G, Internet of Things, edge computing and other technologies, distributed computing power will be found more and more. The space to play, ushered in the spring of development.
However, existing blockchains cannot support large-scale distributed computing and application scenarios. The serial structure of the traditional blockchain determines the amount of computation that can be supported and the nodes that can be allowed to participate. The scale of the calculation can be arbitrarily adjusted according to the demand. It can restrict the scalability of the applications supported by the blockchain itself, and restrict the scalability of participating in distributed computing power.
In this respect, the parallel chain of Buddy consensus algorithm based on Paxos algorithm is adopted, which has natural advantages. Buddy consensus adopts the principle of mutual authentication of nodes, which can form local consensus in parallel, and then form a collection of local consensus and extend to the global consensus. Under such a consensus mechanism, the blockchain can allow multiple sub-chains to coexist, and at the same time, block out in parallel, which means that the block size of the blockchain can be broken in a unit time, and the number of nodes that can participate in the calculation can be sufficient. Mobilize distributed computing power to form a scale that can generate business value; in addition, the Buddy consensus is different from the PoW consensus of traditional blockchains, and it is necessary to compete for power and finally leave only one node with accounting rights. The waste is lost, and the original intention of using distributed computing power to reduce costs is lost. The parallel blockchain adopts Buddy consensus, which can allow nodes to participate in calculation of different difficulty according to their bandwidth resources, connections and computing performance, so as to further utilize distributed Calculate power, make rational allocation of resources and true decentralization.
Driven by blockchain technology, especially parallel blockchain technology, distributed computing will gradually move toward the stage of large-scale commercial application.
Scan and join us !
