2 System key technologies
2.1 Transparent proxy that encapsulates multiple communication protocols
The communication protocol agent process is a protocol encapsulation for data access of various business systems on the intranet, including database transfer agent, file transfer protocol (FTP), data collection and forwarding agent (104/101 protocol), etc. For business systems that use relational databases, use the Java database connectivity (JDBC) standard or open database connectivity (ODBC) for interface access; for business systems that use file transfer protocols for transmission, access them according to the FTP standard. File data; for business systems based on real-time data, data forwarded by the real-time system is received according to the IEC60870-5 series of protocols. The core feature of the communication protocol agent is to provide access interfaces for different data sources and convert the received data into standard XML files according to the unified data format formulated by the data unification platform.
2.2 XML-based data unification platform
The unified data platform develops a unified XML-based storage and presentation method for the heterogeneity and looseness of data in various business systems. First of all, XML files are interpretive data files based on full text, which naturally meet the data transmission regulations for isolation devices; secondly, XML files are not only conducive to data display, but also can encapsulate business logic structures, which are useful for the actual data sources of each business system. It is well-structured and can perform good data mapping; thirdly, the biggest difference between XML and other data representations is that it is extremely simple, easy to read and extensible for external sharing of data.
The data unification platform has developed its own data dictionary and data description, and built an XML-based database. In actual applications, dictionary mapping is first performed for different businesses to ensure the uniqueness of keywords and data. Dictionary data tables mostly use whole table storage, while various raw and cooked data tables mostly use incremental storage, that is, each time new or changed data is stored after the last data storage. Data storage includes a variety of storage modes: storing dictionary (i.e. table structure) and data, only storing data, only storing as local files, only storing in local database and transmission rule table, etc. [6]. Then the data communication agent initiates a connection with the external network database and performs one-way data transmission; after the external network platform receives the data and successfully verifies it, the parsed data is merged into the external network database, that is, the external network data sharing library.
2.3 Isolation device crossing
For physical isolation devices, the data exchange between the internal network port and the external network port is serial port protocol data transmission based on virtual address mapping. The clear regulations on the isolation device are: for forward transmission (internal network to external network), the response message is prohibited from carrying application layer data; for reverse transmission (external network to internal network), the use of general network services is strictly prohibited.
During the process of transmitting data from the internal network to the external network, the external network cannot use the transmission control protocol (TCP) method to respond to the data transmission request and establish a connection. Therefore, during forward data transmission, the forward isolation device must be installed. Mapping the physical address of the internal network data proxy platform and the physical address and target port of the external network data proxy platform, using the user datagram protocol (UDP) protocol, the internal network data is transmitted to the outside in the form of messages through the forward isolation device. net[7].
The specific implementation is: using the Socket (4BDS UNIX process communication mechanism, socket) communication method, establish a Socket client on the intranet data proxy platform, and establish a Socket server on the external network data proxy platform. During forward data transmission, the intranet Socket client packages and compresses the transmission data and sends it to the external network Socket server in the form of UDP data packets; the external network Socket server does not need to shake hands with the internal network Socket client and only needs to receive broadcasts. The data message can be decoded to extract the data[8]
When data from the external network is transmitted to the internal network, the reverse isolation device blocks general network services and fundamentally blocks the TCP/IP protocol stack. Therefore, during the reverse data transmission process, in addition to mapping the physical address and target port of the intranet data proxy platform and the physical address of the external network data proxy platform on the reverse isolation device, it is also necessary to use the data transmission tool of the reverse isolation device itself. Transmit external network data to the internal network in the form of text files.
2.4 Data management Data management mainly includes task scheduling, data maintenance management, data monitoring, etc. Task scheduling is responsible for the access frequency, storage interval, task execution and stopping of data in various business systems on the intranet; data maintenance management is mainly responsible for the maintenance of warehousing data, including data backup, data increase and decrease, access interface start and stop, data Increase or decrease proxy protocols; data monitoring provides functions such as data section monitoring, internal and external network interaction data, communication monitoring and logging.
3 Application Case 3.1 Lianyungang Load Forecasting System Deployment Structure In the process of building a prefecture-county integrated load forecasting system [9] in the Network Security Zone III of Lianyungang Power Supply Company, the cross-region data sharing platform was used to undertake most of the data transmission tasks. This system is Typical cases of using cross-region data sharing. The load forecasting system is a daily business system for full-time personnel in the dispatch operation mode. It obtains the network supply load data of the day from the SCADA system, performs prediction calculations through the clustering optimization [10] algorithm, and sends the prediction curve back to the SCADA system for future reference. SCADA system curve monitoring module display. The system diagram is shown in Figure 4. 3.2 The real-time data transmission load forecasting system requires the SCADA system to provide real-time network supply load data of Lianyungang and its four subordinate counties as the basic data for ultra-short-term and short-term load forecasting. The system design transmits the real-time network supply data sent according to the 104 protocol from the SCADA system to the intranet data unified platform through the server, and updates it to the real-time database. Then, the data is transmitted across regions through the forward isolation device to the real-time shared library of the external network data unified platform [11]. The load forecasting system server will access the real-time shared library of the external network to obtain real-time data of the network supply load, and perform online load monitoring and ultra-short-term load forecasting. 3.3 The external network historical data storage load prediction system server will transfer the real-time data of the external network real-time database into a load history curve every 1 minute. After removing or filtering the abnormal data in the load history curve [12], the server will store the load into a load history curve. The forecast history database is used by the system for daily load forecast calculations, and provides users with load history curve query and the generation and query of various reports. 3.4 Intranet Forecast File Upload After using the load forecast system for forecast calculation, the user needs to send the generated load forecast curve back to the SCADA system for display in the SCADA system. At this time, the load prediction server sends the prediction curve to the non-real-time library of the intranet through the external network file proxy and through the reverse isolation device. The data communication agent will regularly take out the prediction curve of the non-real-time library according to the task scheduling plan and send it back to the SCADA system server through the FTP agent in the communication agent.