Demand case
The following requirements are adapted from real situations.
-
Monitor the production process (MES), after-sales management (CSM), inventory management (WMS, TMS) and other manufacturing links at the factory side, and monitor and warn inefficient operations, abnormal operations, bad information, and key actions at some key nodes to help advance Identify problems and assist in formulating treatment strategies in a timely manner.
-
The monitoring solution needs to provide support for timeliness and event-based dimensions
-
Monitor tasks that are not completed within the required time and give early warning, such as inspection unqualified orders, inspection abnormal orders, process abnormal orders... and other business operations.
-
Monitor the bad indicators of the corresponding indicators to be greater than certain values to give early warning, such as test failure rate, aging failure rate... and other business operations, to supervise the poor quality and find problems as soon as possible.
-
-
Support the cloud (group) to monitor the quality of all factories
-
Support the factory side to view the data Kanban and data analysis (BI) of the factory manufacturing process in real time
-
Support the cloud (group) minute/hour/day level to view the data Kanban and data analysis (BI) of all factory manufacturing links
-
etc.
demand analysis
-
What is the background of the requirement? Is it part of an overall solution? Do you need to consider the smart factory business and technical architecture?
-
How many factories are there? What is the system architecture (network bandwidth, IT system deployment, communication protocol, hardware resources) of the factory and the group?
-
How about the data scale, data format, data dimension (business system), and data operation (CURD) of factories and groups?
-
What is the budgeted cost to implement the requirements (hardware budget, labor budget, software budget)?
-
What is the required manpower input and delivery time? Who is in charge? Who are the people involved in each function?
-
etc.
Sort out program constraints
According to 需求调研the collated information, complete the combing of the constraint conditions of the scheme. For example:
-
In the communication network between the factory (side) and the group (cloud), 5 factories communicate through VPN with a bandwidth of 10MB. Then this kind of restricted network will affect a large amount of data transmission. Will the private network be upgraded in the future? If you do not upgrade, how to design a solution to ensure the stability of data transmission?
-
Deployment of hardware and software at the factory and resource budget. The factory only has one common configuration server, and the business IT system + PostgresSQL database is installed. In order to save costs, it is not planned to expand a large number of servers in the factory, and a common configuration server (X86 IBM 48GB 4TB)? Budget around 3-4W? . Such constraints require that the scheme cannot use complex data processing, monitoring and early warning schemes that rely on multiple servers.
-
etc.
Scheme business logic diagram
Through the overall logic diagram of the scheme, the data flow direction and key technical points of the scheme are preliminarily sorted out. Through the diagram, business stakeholders, demand proposers, solution designers, and solution implementers can quickly understand the overall design solution without going into technical details.
The overall logic diagram has multiple presentation methods and dimensions. For specific theories, please refer to the software design UML diagram. For agile development, it is not recommended to spend too much time in the early stage to draw UML diagrams from various perspectives according to the theoretical process. Draw an overall flowchart/framework diagram first, so that relevant parties can reach a consensus on demand points, technical points, etc. on the same diagram. For example:
Associate the identified constraints with each link in the business logic diagram. For example:
-
VPN network-related constraints affect the design of real-time data return schemes and technology selection
-
The budget constraints of factory-side hardware resources affect the design of data processing schemes and technology selection
-
Business system type (PG/Oracle) constraints affect data processing scheme design and technology selection
-
etc.
Scheme design and technology selection
After identifying the demand constraints and the relationship between scheme design and technology selection through the business logic diagram, key technical points and alternatives for key technical points can be sorted out based on industry experience, team experience, and architect experience.
Then, according to the alternatives, the matching of the constraints is verified one by one. At the same time, it is also necessary to consider some technical constraints, such as: stability, operation and maintenance, security, etc., which should be supplemented according to the template and specifications of the enterprise.
The common technology selection table is as follows, taking the options of key technical points of data processing in the scheme as an example:
Scheme verification
Set up the experimental environment, prepare the constraints in the requirements, and conduct business logic testing and program verification.
Program realization
Follow the implementation path of the verification environment, consider the production situation, and complete resource application, environment preparation, system deployment, parameter tuning, code development, application deployment, and document supplementation.
Program launched
According to the implementation path of the verification environment, combined with the production environment, system deployment, parameter tuning, monitoring and early warning.
Solution iteration
After going online, iterate the solution according to the problems found in the operation and maintenance process and the adjustment, change or addition of requirements.