Time series data in industrial Internet of things

CnosDB Our R & D engineers are advancing in full swing Rust The development of the version , Here it is Jesse Thank you for your understanding and support to the community , In this issue, we want to talk about time series data in the industrial Internet of things . 

This article only represents personal views , If there is bias , Please ask Hai Han. ～  

The industrial revolution is a watershed in human history . Great changes have taken place in the way humans work —— From piecemeal homework to mechanized manufacturing . since 18 Since the 20th century , A series of innovative waves such as assembly lines and computers are constantly changing the nature of manufacturing . today , We are in the process of industrial transformation in another wave of innovation —— Industry 4.0, It involves providing raw data and trained machine learning models to industrial production autonomous systems to speed up production processes .

say concretely , Manufacturers want more accurate and predictable output . So , They used physical equipment , such as , All kinds of sensors installed on the machine to monitor data . These sensors are the industrial Internet of things （IIoT） The basis of , And recorded a large number of key data about the performance and function of industrial machinery .

Key characteristics of data ： Time

We found that , No matter what type of readings the industrial operator's sensors collect , It always contains a timestamp , This provides a common feature for industrial readings . Time series data with timestamp has also become our processing and understanding industry 4.0 The key fulcrum of IOT data .

Fortunately , Supporting industry  4.0 The basic principle of is consistent with the characteristics of time series data . Industry 4.0 betake ：

（1） interconnection ： Make the device 、 The ability of sensors and people to connect and communicate with each other .

（2） Information transparency ： Interconnection allows a large amount of data to be collected from all points of the manufacturing process , These data provided to industrial operators can provide them with effective understanding , Help identify areas of innovation and improvement .

（3） Technical assistance ： The ability to summarize and visualize collected data using a centralized dashboard , So that industrial operators can make wise decisions and solve urgent problems immediately .

（4） Decentralized decision making ： The system will have the ability to perform tasks independently according to the collected data . We just need to manually enter exceptions into the system . Combine these concepts and goals with some of the time series data IIoT Use case comparison , We will find that time series data touch almost all aspects of industrial operation .

Industry 4.0 ——  Combined with time observation

Industrial operators want greater visibility into their machines and processes , The time series data provides the basis for this . Transform these raw data into insightful industries Know-How It's industry 4.0 One of the key objectives of time series data in . Use reasonable tools to deal with 、 Converting and analyzing these data will become an industry 4.0 The key to the success of the plan . We found that , Many factories and manufacturers are still using traditional in industry 3.0 Industrial real-time database used in （ May refer to ：https://mp.weixin.qq.com/s/QYEZ16sbe4YQuCZY3mXp7Q）, These plans are not suitable for Industry 4.0 Systematic , There are several reasons ：

（1） The high cost ： These solutions are expensive to set up and maintain , And charge license fee and support fee every year . The installation of databases previously used by most industrial operators requires customized development to meet the needs of specific businesses or processes , And may require external consulting resources . The proprietary nature of these systems means that this work is time-consuming and expensive .

（2） Supplier lock up ： These solutions are usually based on Windows, No simple 、 Open API To interact with other software . therefore , Industrial operators need to purchase all integrations and components from a single supplier , This will lock users into proprietary solutions .

（3） Poor scalability ： Previously, operators considered the limited data set when building the time series database . This is introducing artificial intelligence or machine learning  (AI/ML) There will be problems when waiting for advanced functions . These functions need more data to train the model , Traditional systems cannot handle these data .

（4） Poor developer experience ： Previous solution services , Adopt the traditional closed design ,API Limited support . therefore , Implementing or integrating these systems requires a lot of time and money . These closed design solutions provide few built-in tools , There is no developer community , Nor does it support modular development methods , This limits the ability of developers to select tools that best suit their organizational needs .

（5） Orphaned data ： Data acquisition and monitoring system manufacturers may provide data history for their equipment , But most use traditional manufacturing execution systems (MES) Industrial organizations will integrate all their data into a single local data history . however , Due to the lack of microservice architecture and openness API, And the widespread use of firewalls and subnets , Data is usually separated at the site level .

Because it cannot be compared with modern  IT、 Cloud or open source software solution integration , The traditional industrial real-time database previously used by operators cannot provide the flexibility and connectivity needed to develop industrial operations . This significantly reduces industrial  4.0 The operation technology and  IT The effectiveness of the system and the data it contains , Because the lack of interoperability between traditional industrial real-time databases and other systems will inhibit innovation and limit observability .

Replace industrial real-time database

We naturally think about a problem , If the traditional industrial real-time database is not industrial 4.0 The answer , What should we use to replace him ？ We believe that , Most manufacturers may want to choose their familiar technology to replace , For example, relational database , Unfortunately, relational databases cannot be extended for massive data , There are problems of low storage capacity and query efficiency of time series data .

in fact , The most suitable database substitute for these traditional industrial operators or manufacturers should be an open source timing data platform . such as ,CnosDB It is a database that specializes in processing time series data , It has a good compression ratio and supports the writing of massive data . It USES API, So it can be integrated with almost any other connected device .CnosDB It is also a modeless platform , So it will automatically adapt to incoming IIoT Changes in data shape . And international leading manufacturers InfluxDB comparison ,CnosDB The future is not only fully compatible InfluxDB, And the distributed function is completely free , Customers can migrate painlessly , At the same time, it will also embrace cloud ecology .

Manage and utilize time series data

This extensive connectivity also makes it easier for manufacturers to monitor and manage distributed systems and networks on site, as well as remote devices . such as , If industrial manufacturers or operators have three different facilities in different regions ,CnosDB Allow them to collect data from every sensor on every machine in every facility . The data generated by each facility can be summarized and stored on site . These summarized data will be sent to the central storage instance , This example collects data from all three sites and summarizes , Finally, the enterprise can analyze through the summarized data at the company wide level , And then there's a different Know-how. This also applies to any connecting device on the edge , Whether rural solar panels or with  GSM（ Global mobile system ） Connected ocean buoys . No matter how your company defines “ edge ”, CnosDB Can process the data collected by these IOT devices .

CnosDB brief introduction

CnosDB It's a high performance 、 High usability open source distributed time series database , It has been officially released and fully open source .

Welcome to our code warehouse , One key, three links ：https://github.com/cnosdb/cnosdb