A 2016 survey of industrial leaders found that 67% of them prioritized investments in data analysis technologies to ensure success in the competition, even at the expense of cutting costs in other areas. Today, when the issue of collecting data is no longer an issue, their storage and analysis are becoming increasingly important. That being said, there are a number of challenges that stand in the way of anyone who has decided to benefit from the collected data:
- Expectations of quick results of the analytical project. Most often these are overstated requirements for CDO to develop a solution for specific client tasks. Complicating this challenge is the changing contractors responsible for digital transformation.
- A large amount of data in a variety of information systems. All of them can be a good basis for making decisions, but it is difficult to extract knowledge from them.
- Inaccessibility of data quality control. Bad or insufficient data negatively affects the entire further process, including the decision made on the basis of them.
- "Patchwork" automation and the lack of a unified system for working with data. Although many processes are already automated in an organization, the complexity arises when it is necessary to combine the results of the work of these disparate systems.
To overcome these barriers, Factory5 created a platform based on technology and software components to efficiently collect, process and store data. In fact, it is a pre-configured set of tools that allows you to process big data and reduce labor costs for developing and implementing business applications. What makes the solution effective for data analysis?
1. Connectors to main sources
The F5 Platform is designed to collect and process industrial-scale data, therefore it accepts data from the main known sources: traditional databases, non-relational databases and storages, cloud storages, connected devices and Historian and file sources. Data comes to the platform in three ways:
- Automatically from sensors in real time.
If the equipment is equipped with sensors that transmit telemetry in real time, then through the connectors it goes to the platform in real time. Also, data can come to the OPC server, which is a standard for data exchange between industrial equipment and information systems. - Manually from files.
The data is accumulated on the equipment, and when the equipment is ready to transfer this data, the files are manually transferred to the platform using flash drives. - From external systems without manual intervention.
The platform integrates with external systems such as 1C, Galaktiki EPR. Therefore, data comes from external systems using REST API requests or is unloaded, providing REST API for external systems.
This allows you to use the entire array of data circulating in the company. And the connectors built into the Platform allow you not to waste time on their development and speed up the process of working with data.
2. Data markup service
In order to be able to analyze the incoming data, they must be marked up. In practice, this means building a model of a real object - equipment or process - and linking it to the incoming data. The output is a digital twin. The platform provides functionality for creating such a twin: a convenient interface that allows you to visually see all sensors and their parameters and allows you to link them to equipment.
For example, temperature data comes from a sensor on the engine. Thanks to the created double, they can be written into the database as "sensor # 123 transmits parameter X". In order to be able to write the rules, it remains to explain to the system that sensor # 123 is a sensor on the engine, and parameter X is the temperature. After that it will be possible to write rules like "if the temperature is more than 50 then ...".
Without the data markup functionality, these parameters would fall into the database cell # 456. And the rule for them would look like "if the value of cell number 456 is more than 50 then ...". This would require specially trained personnel, whereas with a data markup service, equipment engineers can write expert rules.
3. Service of creation and library of expert rules
Expert rules are used for data analysis. The F5 Platform allows you to write your own rules in a simple language that engineers understand - Domain Specific Language. You don't need to know the programming language for this.
In addition, the Platform has a library of expert rules for industrial equipment, which contains about 2,000 pre-written rules.
4. MX-editor and library of mathematical models
Also, the data is processed using mathematical models. They examine the flow of received data for deviations and thus look for anomalies not yet known to the engineer and not described in expert rules. The platform contains about 100 pre-built mathematical models for working with different amounts and quality of data.
5. Microservice architecture
Each functional module of the F5 Platform is separated into an independent service, which is located in its own "container", which is not related to others. Containers contain everything the application needs to run. Each container is isolated and is a secure base for the Platform's services.
Expert rules and mathematical models are just such services. In addition to them, there is a service for building cubes, reports, a pipeline manager and a service for managing user rights. The microservice architecture of the F5 Platform allows you to update services if they are improved, without affecting the rest. This simplifies the management of the Platform.
6. Scaling of computing power
Based on the architecture of the Platform, it is possible to increase the computing power according to actual needs. These are the properties of horizontal scaling: there is an even load of application servers, it is possible to connect additional servers, as well as distribute data across different servers. This ensures high system performance and does not reduce fault tolerance.
7. Pre-configured set of tools for creating business applications
The F5 Platform contains and pre-configured all tools for collecting and analyzing data, as well as visualizing reports:
- connectors to various DBMS: PostgreSQL, ClickHouse;
- connectors to industrial protocols: OPC-UA, REST API;
- Pipeline Manager for configuring and executing arbitrary scenarios for data processing / analysis;
- service of expert rules;
- service of mathematical models;
- various widgets for presenting information;
- constructor and designer of reports.
The F5 Platform is designed to simplify the process of working with data, as well as speed up the process of creating business applications. The platform implementation reduces labor costs and allows companies with little experience to enter the analytical solutions market.