Blockchain circle

One stop hot information platform

About us:

Blockchain circle provides the latest information about blockchain, digital currency, digital wallet, exchange, metauniverse, bitcoin, Ethereum, contract, financial management and so on, and always pays attention to the latest market...

What are the key indicators of blockchain network performance

Time : 17/03/2022 Author : zfjat0 Click : + -
        We will explain below why it is best not to shelve performance evaluation and what metrics are best to use and make full use of them. Let's find out. TPS measurements are from distributed databases. They are typically performed using standardized transaction types or collections (e.g., some inserts, updates, deletions, and constant selections) and configured for specific clusters or individual machines. Such "comprehensive" indicators cannot reflect the actual performance of the database or blockchain under discussion, because the transaction processing time may be different in such a system. Consensus oriented databases (see "cap Theorm") will not submit transactions until a sufficient number of confirmations are received from other nodes & mdash& mdash; This is very slow.
 
        In the availability oriented database, if the transaction is simply written to disk, it is successful. They immediately provide updated data & mdash& mdash; This is very fast (although this transaction may be rolled back in the future). If the transaction updates only one data unit, the TPS will be higher. If a transaction updates many data units (rows, indexes, files), they will block each other. To discuss our cherished "transactions per second", you need to describe all network conditions, parameters, and benchmark logic. In the blockchain, applying a transaction to an internal database does not mean that the consensus will accept it.
 
        In the pow consensus, the transaction will never be finalized. If a transaction is contained in a block on one machine, this does not mean that it will be accepted by the entire network (for example, in the case of another fork winning). If the blockchain has an additional algorithm to ensure finality (e.g., EOS, Ethereum 2.0, polkadotparachains use the method common with the grandfather's finality), the processing time can be regarded as the time when the node sees the transaction and the next completed block. Such TPS are very useful, but they are rare because they are lower than expected. The number of transactions processed and the maximum / average / minute processing time (on the local node) are very easy to measure because the functions that perform these operations are usually expressed in code.
 
        The transaction processing time is equal to the time required to update the status database. For example, in the "optimistic" blockchain, the processed transactions may have been verified, but have not been accepted unanimously. under these circumstances, The node sends the updated data to the client (assuming there will be no chain bifurcation). This metric is not very reliable: if another chain branch is selected as the main branch, the statistics on the transaction must also be rolled back. In the test, this is often ignored. "Our blockchain received 8000 TPS yesterday." these figures can usually be found in short project reports because they are easy to measure.
 
        Just one run node and one load script are enough. In this case, no network delay will reduce the speed of reaching network consensus. This index reflects the performance of the state database without being affected by the network. This figure does not reflect the real network bandwidth, but it shows the limit it will try to reach if the consensus and network are fast enough. The result of any blockchain transaction is multiple atomic storage writes. For example, a bitcoin payment transaction involves deleting several old utxos (delete) and adding new utxos (insert). In Ethereum, execute a small smart contract code and update several key value pairs.
 
        Blockchain nodes can be implemented in several programming languages & mdash& mdash; Which is more reliable. For example, the Ethereum node has rust and go implementations. Keep this in mind when testing network performance. This simple indicator shows the number of blocks generated by a validator. It relies on consensus and is critical for assessing the "usefulness" of a single verifier network. Because verifiers can make money on each block, they are responsible for the stable operation and safety of the machine. You can determine which validator candidate is the most qualified, protected, and ready to work in a public network with real user assets.
 
        Metric values can be publicly checked & mdash; Just download the blockchain and calculate the number of blocks. Finality ensures that all transactions included in the blockchain will not be rolled back and will not be replaced by another chain fork. This is a way for the POS network to prevent double consumption attacks and confirm cryptocurrency transactions for users. When a block completes the chain containing the transaction, rather than when a transaction is accepted by the node, the user thinks that the transaction is the last block. To complete a block, the verifier must receive the block in the P2P network and exchange signatures with each other. Here we check the actual speed of the blockchain, because the time when the transaction is completed is the most important for the user.
 
        The terminal algorithm is also different, it will intersect and combine the main consensus (read: Casper in Ethereum, the last irreversible block in EOS, grandpa in odd and even Polkadot and their modifications, for example, mixbytesrandpa). For networks that do not complete all the blocks, a useful metric is the delay between the last completed block and the current latest block. This figure shows how far the validator is behind, which is consistent with the correct chain. If the difference is large, the final algorithm needs additional analysis and optimization. Point to point system & mdash& mdash; The middle layer of blockchain Network & mdash& mdash; Often overlooked.
 
        This is due to the fuzzy delay in the transaction between the block delivery and the verifier. When the number of validators is small, they are localized, the peer list is hard coded, and everything works well and quickly. However, just like the verifier, the nodes are geographically distributed and the lost data is simulated. We are facing a serious "TPS" failure. For example, when the EOS consensus is tested using the additional finalization algorithm, the number of verifiers is increased to 80-100, distributed in four continents, and has little impact on the finalization. At the same time, the increased packet loss seriously affects the final result, which proves that additional P2P layer configuration is required to better resist network packet loss (rather than high latency).
 
        Unfortunately, there are many different settings and factors, and only benchmarking allows us to understand the number of validators required and obtain a fairly comfortable blockchain speed. The configuration of the P2P subsystem is clear from the documentation. For example, look at [libp2p], [kademlia] protocol or [BitTorrent]& middot; The number of times the previously cached data block is returned, and the number of times the request needs to be further forwarded in order to find the required data block (cache hit / loss of simulation data). For example, when accessing data, a large number of omissions means that only a few nodes have the requested data, and they have no time to distribute the data to each node.
 
        P2P traffic received / sent allows identification of nodes dealing with network configuration or channel problems. The standard system metrics of blockchain nodes are described in a large number of source codes, so we will briefly introduce them. They help to find logical bottlenecks and errors. The CPU displays the amount of calculation performed by the processor. If the CPU load is high & mdash& mdash; Nodes are computing something, actively using logic or FPU (almost never used in blockchain). For example, the latter situation can occur because the node is checking for electronic signatures, using strong passwords to process transactions, or performing complex calculations. CPU can be divided into more indicators pointing to code bottlenecks.
 
        For example, system time & mdash; Time spent on kernel code, user time & mdash; Time spent on user processes, IO & mdash; Wait for I / O from a slow external device (disk / network), and so on. Modern blockchains use key value databases (leveldb, rocksdb), which constantly store "hot" data in their memory. Any loaded service can be affected by memory leaks caused by errors or attacks against node code. If the memory consumption is increasing or rapidly increasing, it is likely to be caused by a large number of state database keys, large transaction queues, or an increase in the amount of messages between different node subsystems.
 
        The complete node responding to the network client depends on the file cache index. When a client accesses various parts of the status database and transaction log, old blocks in the disk may appear and new blocks may be replaced. Which in turn reduces the response speed of the client. The main network indicators are the size of traffic (in bytes), the number of network packets sent and received, and the packet loss rate. These indicators are often underestimated because the blockchain cannot process transactions at a speed of 1GBIT per second. Currently, some blockchain projects allow users to share WiFi or provide services for storing and sending files or messages. When testing such a network, the quantity and quality of network interface traffic become very important, because a crowded network channel will affect all other services on the machine.
 
        The disk subsystem is the slowest component of all services and often leads to serious performance problems. Excessive logging, unexpected backup, inconvenient read / write mode, and a large number of blockchain volumes & mdash& mdash; All of which may result in significant node deceleration or excessive hardware requirements. The operation mode of blockchain transaction log using disk is similar to that of different DBMS using pre write log (wal). Technically, the transaction log can be regarded as the wal of the state database. Therefore, these storage metrics are important because they can identify bottlenecks in modern key value databases. Read / write IOPs, maximum / minimum / AVG latency, and many other metrics that help optimize disk operations.
 
        Each group is important because it restricts the operation of other components in the event of subsystem errors. Even the deceleration of a small number of verifiers can seriously affect the entire network. In consensus and finality algorithms, the most troublesome errors only occur when large transaction flows or consensus parameters change. Their analysis requires repeatable test conditions and complex load scenarios. Guangzhou, August 2, 2022 / PR News Agency / -- July 29, the opening ceremony and cooperation signing ceremony of Guangzhou Moqi Technology Co., Ltd. was held in Guangzhou Guigu science and technology park. The establishment of Guangzhou company is an important measure for Moqi technology to lay out the Greater Bay area of Guangdong, Hong Kong and Macao.
 
        Murky. Shanghai, August 1, 2022 / Xinhua / -- information innovation refers to the application and innovation of information technology. The development of information innovation industry is not only related to the safety and controllability of national information, but also an inevitable choice to promote the digital transformation of China's economy. In recent years, with the deepening of enterprise digital transformation, the challenges faced by human resource management are becoming increasingly complex. With. Shenzhen, August 1, 2022 / PR / -- recently, Shenzhen Municipal Bureau of human resources and social security (hereinafter referred to as Shenzhen Municipal Bureau of human resources and social security) awarded the plaque of Dongxin post doctoral innovation practice base. Liu Yang, chairman / CEO of Dongxin, personally listed the post doctoral innovation practice base. The core technology.
 
        (Global TMT, July 28, 2022) as the mother city of Chongqing, Yuzhong District is an important metropolitan core for the linkage and expansion of the twin city economic circle in Chengdu Chongqing region. Yuzhong District big data Bureau, together with Inspur Zhuoshu, takes data as the core and intelligent hub as the base to build a grassroots intelligent governance platform with 1 + 4 + n as the framework. Memory chip is the specific application of the concept of embedded system chip in the storage industry. Therefore, whether it is a system chip or a memory chip, the software is embedded in a single chip to achieve multi-function and high performance, as well as support for a variety of protocols, a variety of hardware and different applications. Abstract: in order to meet the process monitoring requirements of sintering production in a steel plant and realize the production management function of ironmaking operation Department, the sintering L2 management system software is developed.
 
        The system realizes the intensive use and management of the system hardware through the HP VMware ESX and vSphere virtual machine technology: through the use of Oracle number. Abstract: combined with the concept of agricultural product e-commerce system and the development status of agricultural product e-commerce system in China, this paper analyzes the basic characteristics of agricultural product e-commerce system, designs the database of an agricultural product e-commerce system from the aspects of function analysis, data dictionary and development environment, and discusses how to better promote the agricultural product e-commerce system. Recently, the Xinhua three group under Ziguang was elected as a member unit of the open source community council of openeuler.
 
        In the future work of the open source community, xinhuansan group will work with other community members to promote the innovation of the open source operating system, promote the sustainable development of the community, and build. Recently, the international data company IDC released the quarterly tracking report on China's enterprise level external storage market, 2022q1. According to the data, with a share of 12.8%, the new H3C group under Ziguang shares ranks the second in China's storage market, and has been ranked the second in the market for 18 consecutive quarters. The year-on-year growth rate of new H3C intelligent storage in this quarter is 15. Chen Xiaojian, general manager of product department of Amazon cloud technology in Greater China, Beijing, July 11, 2022 / Xinhua / -- since the birth of mankind, we have never stopped exploring the data recording and processing capabilities.
 
        For example, the invention and improvement of writing, writing and printing in ancient times can be regarded as an improvement on the human brain. Since the birth of human beings, the exploration of data recording and processing capability has never been stopped. For example, the invention and improvement of characters, letters and printing in ancient times can be regarded as a remedy for the usability and expansibility of the human brain, which is a single and non expandable "original database". Until the advent of computers, human beings have a modern sense of "Empowering academic research with global business data". On July 7, 2022, Shanghai / PR News Agency / -- microcode Deng & Bai announced the official release of the global enterprise information platform (D & amp; bhooversip solution) for education and scientific research.
 
        The platform relies on Deng Bai. (Global TMT, July 1, 2022) recently, Inspur information released four skill models (skillmodels) generated based on the source 1.0 model, namely, the dialogue model yuan Xiaowen, the question and answer model yuan Xiaosu, the translation model yuan Xiaoyi, and the ancient Chinese model yuan Xiaowen. These skills. (Global TMT, June 29, 2022) CCID Consulting (CCID) recently released the annual report on China's big data market research for 2021-2022. According to the report, China's big data market still maintained a high growth trend in 2021, with a growth rate of 32.7%, and the overall scale reached 468.15 billion yuan.
 
        ...。 Stuttgart, Germany and Beijing, China, June 29, 2022 / AP / -- timecho, the provider of industrial Internet of things time series database management system and related services, announced today that it had completed the angel round financing of nearly 100 million yuan. This round of financing is led by Sequoia China, koala fund, Gobi venture capital, Yunzhi. (Global TMT, June 27, 2022) Zhangqiu district is located in the east of Jinan City, Shandong Province. Zhangqiu District selects Shengjing sub district country garden community, Shuangshan sub district sanjianxi village and other communities (villages) as the pilot, and uses Inspur Zhuo digital grassroots data management
 
        
Previous:Algorithmic Politics Series 7: from algorithmic populism to algorithmic Democracy: the balance of democratic politics in the digital age
Next:No more

Related articles:



© 2005-2032 | Blockchain Circle & & All Rights Reserved    Sitemap1 Sitemap2 If there is infringement, please contact us at: