Development of Geographic Information System Architecture Feature Analysis and Evolution Trend Research

Abstract

A geographic information system (GIS) is a technical system which is supported by computer software and hardware systems. It focuses on the geographical information related to the whole or part of the earth’s surface. It is used for collecting, storing, managing, calculating, analyzing, displaying, and describing geographical information. It has inherent advantages in processing geographic data and plays an indispensable role in the sustainable detection of natural resources, natural disaster risk management, urban sustainable development planning, etc. With the continuous development of technology, the integration of GIS with emerging technologies such as big data, cloud services, and artificial intelligence creates new geographic information systems and entirely new development directions. The GIS architecture is of great value for the efficient execution of GIS systems. In this process, as the organizational form of GIS systems, the GIS architecture is also constantly evolving with the intersection and integration of GIS and other technologies. This research reviews a large amount of literature on component technologies, 3D technologies, cloud computing, big data, artificial intelligence, and so on, at home and abroad and analyzes and elaborates on the current development status and trends of GIS software architecture. It discusses in detail the characteristics and future development directions of different GIS software architectures in different periods and makes delicate descriptions of their hierarchical features. This study aims to summarize the advantages and disadvantages of architectures in different stages, the interactivity from the user’s perspective. On this basis, it studies the development trends of GIS integrated with big data and artificial intelligence, summarizes the laws and experience of the evolution of its system architecture, and analyzes the technological drivers of each evolution and their impact on GIS applications. Reviewing the evolution history of GIS frameworks is expected to provide guiding references for more efficient GIS system architecture research in the future.

1. Introduction

A geographic information system (GIS) is a computer system that collects, stores, manages, analyzes, displays, and applies geographic information, as well as a general technique for analyzing and managing mass spatial data [1,2,3,4,5]. GIS has a wide range of applications. It plays a key role in resource management, land planning, natural disaster management, transportation planning, environmental protection, social service planning, and decision support. GIS provides spatial data analysis and visualization tools that help to better understand environmental and social issues, including resource protection, sustainable land use, natural disaster mitigation, transportation efficiency improvement, ecosystem protection, social equity, and smart decision making. Many cases in the 2023 Sustainable Development Goals (SDGs) report, especially SDG 11, the “Sustainable Cities and Communities” section, are mostly based on GIS platforms. Studying the evolution of GIS architecture to make it better serve the whole society in terms of information services and data sharing will help support the realization of sustainable development goals. The GIS technique began in the early 1950s. Canada’s CGIS, the first operational GIS in the world, debuted in 1972 and achieved significant technical advancements such as the scan-in of map data and the conversion of raster and vector data formats [6,7,8]. The 1960s was the expansion period of GIS, focusing on the geoprocessing of spatial data, but the scope of application is still relatively narrow [2,9,10,11]. GIS technique has significantly improved with the rapid development of computer technology in the 1970s [12,13]. The era of application started to push forward, leading to the creation of commercial GIS in the 1980s [10,11]. In the 1990s, it was the socialization stage, and GIS entered a phase of fast growth in China, paying more attention to enhancing the user experience and spatial analysis function [2,6,7,8,9,10,11,12,13,14,15].

The structure of the GIS technique has also unceasingly improved with its continuous development. As the socialization of GIS continued to develop and new architectural systems were launched in the 1990s, component-based GIS [16,17,18,19,20] and WebGIS [21,22,23,24,25,26,27] were extensively adopted by the GIS industry. In the following years, because of the strong mobility of mobile devices, real-time services, and the diversity of information carriers, mobile internet technology has become popular, and mobile GIS has gradually become the mainstream of the GIS field [28,29,30,31,32,33,34,35]. As technology develops and society expands, and more diverse information and data of all types become available, people start to consider how to preserve resources more effectively, save expenses, and access data more efficiently, and the introduction of cloud computing technology offers a solution to this problem. Meanwhile, cloud GIS technology has matured with the constant growth of cloud computing technology [36,37,38,39,40,41,42,43].

GIS software architecture represents the organization mode of the GIS system in the whole GIS software system and throughout the entire life cycle researched and developed by the GIS system. Accurate architecture mode plays an important role in effectively executing GIS spatial operations. Currently, the popular GIS architecture organizations include Components GIS [44,45], Web GIS [46,47], mobile GIS [48,49], grid GIS [50,51], new 3D GIS, Cloud GIS [52,53], big data GIS [54,55], AI GIS [56,57], etc. Based on the above-mentioned architecture, this paper will illustrate and discover the status quo and trend of the GIS architecture system at home and abroad.

At present, the relevant literature on GIS architecture is relatively scarce, especially with the lack of a comprehensive review and systematic study of the development process of GIS architecture. Most of the studies focus on the development of GIS and one-sided analysis of the characteristics of a single architecture. There is little literature to summarize the inherent laws of GIS architecture evolution at different stages. Through the form of a literature review, this paper systematically combs the technological evolution process of mainstream GIS architecture in different stages from component GIS to AI GIS and makes up for the literature gap in this field. The purpose of this study is to explore the current situation and development trend of GIS architecture at home and abroad in recent years. Through the comparative analysis of GIS architectures in different stages, it is expected to summarize the advantages and disadvantages of different architectures and provide a reference for the design of future GIS architectures.

2. Methodology

We conducted a comprehensive search and statistical analysis of Chinese and English academic papers on GIS architecture published between 1991 and 2023. For Chinese papers, we chose the China National Knowledge Infrastructure (CNKI) database as the paper source, which is widely recognized as an authoritative source in academia. CNKI includes a large number of high-quality papers published in core Chinese journals that adopt strict peer-review systems and are widely recognized for their quality. For English papers, we chose the well-known Google Scholar database, which extensively covers various high-impact factor English journals and contains papers that have gone through rigorous peer review, ensuring high quality. The richness of these two databases in terms of quantity and quality of papers meets the needs of our research in investigating GIS architecture literature.

After obtaining the paper samples, we carefully reviewed the authors, abstracts, introductions, and conclusions of the papers to eliminate duplicated content, ensuring the accuracy of the research papers. We used keywords related to GIS architecture such as “GIS architecture features”, “GIS architecture evolution”, and “emerging technologies in GIS” to search for literature. Eventually, we identified 127 high-quality research papers from the two databases as the literature basis for this study. Among them, there are 66 Chinese papers and 61 English papers. Furthermore, 25 papers directly studying GIS systems account for 20% of the total. There are 102 papers on different types of GIS, classified by category. Figure 1 shows the proportion of the number of different types of geographic information system (GIS) studies in the sample of papers included in this review.

We also used statistical analysis methods to calculate the number of papers in every 5 years from 1991 to 2023 and visually presented the results in statistical charts. Figure 2 clearly shows that during the 2011–2015 period, the number of papers on GIS research surged to a peak. In the following nearly ten years, the number of GIS research papers showed a gradual downward trend. This statistical result can provide valuable information for our analysis of GIS architecture research’s historical context and development trend.

3. Research Status

The architectural framework of GIS assumes pivotal significance as the bedrock upon which GIS systems are built, serving as a guiding force for their ongoing development and enhancement. GIS plays a pivotal role in the management and analytical processing of geospatial data, wielding the capacity to harness and dissect spatiotemporal information in the contemporary era of information technology. This capability lends invaluable support to decision-making processes in urban development and human activities. The integration of GIS with cutting-edge technologies such as cloud computing, artificial intelligence, and other emerging paradigms has given rise to specialized GIS subfields like Cloud GIS, Web GIS, and AI GIS, among others. This convergence of nascent technologies has ushered in profound transformations within the GIS domain and fostered extensive interdisciplinary synergies.

3.1. The Status Quo of Components GIS Architecture

In the computer field, Microsoft introduced the COM component technology in the 1960s and has become one of the trends in today’s software technology due to its strong independence and free composition, flourishing the component technology and greatly accelerating the progress of computer software development. Similarly, the development of component technology promotes the improvement of GIS technology, and the application of component technology enables GIS developers to develop a variety of commercial GIS software quickly, efficiently, and flexibly [58,59]. The core of the GIS is the GIS software [2], and the component GIS applies object-oriented technology and computer component technology to the GIS software development process [17,18]. Using the component object platform as a foundation, programmers incorporate GIS features into a system by creating a collection of components with common communication ports and multilingual applications, and the components communicate with one another via interfaces [19,60,61].

Before the emergence of components GIS, GIS was in the stage of expert GIS, in which GIS architecture mainly uses modular GIS, integrated GIS, modular GIS and core GIS [62], and GIS software primarily utilized desktop GIS along with a specific secondary development language. This method is challenging for users to understand and use, is difficult to expand, and cannot be connected with other systems like MIS (Management Information System). Modular GIS is the application model for the early-stage GIS, which divides GIS functions into several modules to realize corresponding functions, but they do not have the ability to work together harmoniously with each other. Integrated GIS incorporates each module function as an enormous GIS software package. The modules of modular GIS are purposefully segregated to achieve fine functionality, so it is challenging to integrate them with management information systems and specific application models. Core GIS develops a series of dynamic chained libraries on the operating system, and developers can start developing through API [62]. This non-modular GIS architecture can satisfy users’ demands in the early phase, but its flaws and shortcomings are gradually revealing with the unceasing development of computer technology: the development burden is heavy, and the system is growing larger; the integration of spatial data, GIS software, and application models is challenging; development languages cannot be translated between each other, etc. Because of these flaws, GIS applications are confined to mapping, land, and a few sectors, which cannot serve more fields and the public.

Components GIS has flexible, low costs, directly embedded in MIS development tools, robust GIS functionalities, easy development, and other features [18,19], which can effectively solve the above-mentioned problems. Components GIS is designed into three-layer structures: the basic component, advanced common component, and business component [19,20,61,62,63,64]. The basic component is located at the lowest layer of the platform, providing the interactive process between the system and spatial data management and achieving the connection with the database; the advanced common component is formed by the basic component and is packaged with message control; the business component achieves industrial support by encapsulating the feature functionality components for particular functions. This type of architecture enables the development of GIS software to contain a stronger configurability, expandability, and openness and greatly reduces the research and development costs, which has already become one of the essential trends of GIS development. Subsequently, many successful component GIS applications started to appear, including ArcObjects [65,66] released from ESRI and the SuperMap iObjects from SuperMap [67,68], etc. The major architecture of the Components GIS system is shown in Figure 3.

Compared with developed countries, the development of GIS software in China has fallen behind for several years, but the emergence of Components GIS enormously promotes the evolution of Components GIS in China. The Components GIS software technology based on COM has still been the mainstream at present. However, there are still diverse problems that need to be solved such as the component version conflicts brought on by “Dynamic Library Hell (DLL Hell)” and the restrictions on secondary development extensions brought on by the inability to inherit objects, which have an impact on the advancement of COM-based GIS technology [16].

3.2. The Status Quo of Web GIS Architecture

After experiencing the single PC application environment, there appears a novel GIS architecture model in the GIS field—Web GIS. Network environment GIS applications develop from the local area network C/S (client/server) structure of the application to the Internet environment B/S (browser/server) structure of the Web GIS application [21,23]. This structure changes the data processing method of GIS, extending GIS functions on the traditional single PC to the Web. Client application program utilizes network protocols to operate the GIS system on the Web based on the network platform, and users can log in at any node, browse spatial data on the website, perform various map operations and GIS spatial analysis, etc. [24,25,69]. It is the production of network technology applied to GIS development and is a significant constituent part of modern GIS.

Compared with the C/S structure, Web GIS has several characteristics such as convenience deployment, ease of use, and low requirements for networking [70,71,72], which laid the foundation for the geographical information service [73,74]. Its development went through three stages from C/S to B/S [75,76]. The initial C/S mode is split into client and server, which is a loose coupling system, and the client is responsible for the logical expression of the display layer and application layer, and the server in charge of the database and file serves the data preservation and management. This structure causes a reasonably large client load and a small server load, which has a significant negative impact on the system’s overall performance. The server business and data services are separated with the continuous expansion of network services in order to facilitate system maintenance and logical clarity. The calculation and data of application software are then fairly distributed between the client and the server, balancing the load on both ends and improving operational efficiency. But this structure still requires the client to download the application program, and the linking number and communication traffic are limited, which is only applicable for small-sized local area networks. After that, the B/S pattern is introduced to cover the flaws of the C/S pattern, and it deploys the business logic of application software to the server side, and the client only needs a browser for business processing, realizing the function of accessing operations with different people and different place, and also effectively ensure the security of server database. The common Web GIS development software is the Super Engine Figure. It is an interactive, distributed, and dynamic GIS and consists of multiple hosts, wireless terminals of various databases, and the connection between clients and servers (HTTP servers and application servers). GIS extends through the WWW function to become a tool for the public.

B/S architecture applications move from the three-layer architecture: Browser/Web, Server/Data, and Server to the four-layer architecture: Browser/Web, Server/Application, Server/Data, and Server [72,76]. In the new four-layer architecture, the Web Server and Application Server are divided; meanwhile, the secondary development and extended function can be inserted, where the Application Server is generally a components GIS platform supporting the remote procedure call or is packaged by the components GIS platform. Transferring the GIS complex data analysis and processing function (editing, construction of topological relations, automatic maintenance of object relations, mapping) to the GIS Application Server can reduce the data transmission of client and server to the lowest degree, providing the possibility for achieving cineplex and large-scale GIS on the Internet. This architecture brings an enormous advantage which is the extreme expandability of the server; hence, the functions that are available in the Components GIS as an application server can be realized through the B/S structure. The architecture of the Web GIS system is shown in Figure 4.

Web GIS is no longer a simple software for satisfying map browsing and searching but is a Server GIS with an advanced system and stronger functions. IT technology, particularly the rapid growth of mainstream software development technology, requires GIS software to keep in step with the trend of technological development.

3.3. The Status Quo of Mobile GIS Architecture

The research related to mobile GIS began in the early 1990s, and it gradually became an essential component of the GIS field with the growing use of mobile devices [30,31,34]. It expands GIS’s powerful spatial analysis function to the mobile terminal to provide users with position-based information exchange, containing, sharing, and releasing [77,78]. But its system architecture used the mainstream three-layer architecture, which are pattern client, server, and data server, corresponding to the presentation layer, logic layer, and data layer, respectively [33,35].

The presentation layer is the support layer of the mobile GIS server, which directly displays the operation interface of the software to users and supports multiple mobile terminals such as mobile phones, tablet computers, vehicle terminals, etc., and also supports the data synchronism and association of the desktop side and mobile side. The logic layer consists of a wireless gateway, Web server, GIS application server, and mobile tracker gateway. The main function of the wireless gateway is the processing capability for extending mobile devices; the Web mainly manages the requests related to HPPT; the mobile tracker gateway is responsible for obtaining real-time location information from wireless networks. The GIS application server, as the core part of the logic layer, is responsible for both managing and spatial analysis and displaying the spatial database, and it shields various underlying network details, enabling developers to concentrate on a software system and business logic. The dataset in charge of the management of spatial data and attribute data, as the data pump, renders the mobile terminals to interact with other diverse data sources. The architecture of the mobile GIS system is shown in Figure 5.

3.4. The Status Quo of Grid GIS Architecture

Realizing the common interconnection between GIS systems is crucial given the ongoing development of GIS, and because the geographic information of traditional GIS systems that accumulate for a long time exists in the heterogeneous architecture, data sharing and service sharing are hard to achieve. But, the later emerged Web GIS can realize the connections between systems by its flexible service-oriented architecture, but when facing intensive GIS data, it will meet some problems such as loading imbalance [82,83]. The appearance of the grid technique brings hope for solving this problem, in which grid calculation provides high-performance parallel computation, and offers effective colony and load balancing through resource sharing, indicating the route to thoroughly realize the GIS common interconnection [64,82,84,85]. The architecture of grid GIS also exhibits a diverse development trend with the ongoing advancement of technology, in which the current standard grid GIS architectures have the “middleware-based technique” grid architecture, three-layer grid GIS architecture, and five-layer grid GIS architecture [82,86,87].

The “middleware-based technique” network architecture is based on the middleware to construct the network computing environment and GIS spatial service system, and under the support of network computation technology and GIS spatial technology, this technique researches the spatial entity object and the organization and calculation of spatial process simulation to realize the process simulation and active computation representation of complex spatial patterns and provide a knowledge-driven approach to spatial modeling and collaborative network work [88]. Three-layer grid GIS architecture mainly contains a data resource layer, grid service layer, and application layer, and dividing the three-layer architecture can achieve the acquisition and processing of spatial information, distributed user concurrent operations, and multi-level collaborative work mechanism in a heterogeneous environment. The data resource layer is responsible for several functions related to data resources and provides a data interface for upper-layer calls. The grid service layer consists of a series of network agreements and distributed software, providing different types of grid data applications for components and interfaces. The application layer is user-oriented, applying architecture and serves to finish the research and development [64]. Five-layer grid GIS architecture is further subdivided based on the three-layer architecture and enables the logical expression between each layer to be more distinct based on the five-level sandglass architecture. This architecture contains the bias layer, resource layer, control layer, realization layer, and application layer [84]. The basic layer realizes the grid’s basic structure by developing agreements; the resource layer is in charge of the management of data resources; the control layer as the core of the entire architecture becomes a connecting link through customizing different interfaces; the realization layer completes the design of functional interfaces through middle wave; application layer is responsible for user interface management. These three types of architecture have different emphases, which all have important contributions to the development of grid GIS.

Moreover, there are some other architectures, for example, a “two-dimensional dual GIS grid architecture” that focuses more on the characteristics of GIS systems proposed by Xianying Pan, which views the GIS grid architecture as two-dimensional and divides it into a base layer, a distributed database layer, a network operating system intermediate layer, and an application service layer [82]. Wu et al. suggested a “Service-Oriented Grid GIS architecture” to solve problems with GIS integration, cooperation, and resource sharing [85,89]. It can also be divided into four layers: a resource layer, a monitoring layer, an interface layer, and a client layer. The architecture of the grid GIS system is shown in Figure 6.

Grid technology is a burgeoning distributed network technology, and its combination with GIS will provide service efficiency, enriched services, and GIS support. Grid GIS is currently in the exploring phase, but the future holds the promise of far more extensive development.

3.5. The Status Quo of New 3D GIS Architecture

In the face of the burgeoning challenges posed by multi-source, heterogeneous, spatiotemporal big data, and intricate spatiotemporal geographic issues, conventional two-dimensional GIS have exhibited certain limitations. Their inherently low-dimensional nature constrains our comprehension of geospatial environments, restricting geographic information to a static, two-dimensional plane. While the advent of 3D GIS with the advancement of computer technology offered enhanced visualization, early 3D GIS software primarily emphasized visual representation and lacked robust spatial analysis capabilities, thus limiting its utility. The next generation of 3D GIS transcends the constraints of the 2D plane by seamlessly integrating both 2D and 3D components. This advancement enhances the processing capacity of spatial geographic information, resulting in more precise and authentic data visualization. Furthermore, it stimulates the development of innovative processing technologies and methodologies. Contemporary 2D–3D integrated GIS software has made considerable progress in data sourcing, data modeling, and 3D conversion technology. This progress extends beyond mere visualization to encompass novel data model construction, data storage, and management, as well as spatial analysis capabilities, significantly augmenting its practicality [90].

The architecture of new 3D GIS can be broadly categorized into four layers: The first layer is the data layer, which involves the fusion of diverse data sources, including oblique aerial imagery, LiDAR point clouds, 3D field data, building information modeling (BIM) data, etc., with traditional image, vector, and fine-modeling data. This integration enhances the authenticity and accuracy of the 3D scene. Additionally, the data structure facilitates the transformation from irregular triangular networks (TIN) to irregular tetrahedral networks (TIM) and from raster to voxel grids, enabling the faithful representation of the real world in three-dimensional space. The second layer is the service layer, which is responsible for publishing data services, and the service layer encompasses 3D scene services, spatial analysis services, and Web services. These extensions enhance the functional capabilities of GIS. The third layer is the business layer. This layer realizes the 2D-3D integration technology, including the integration of data storage, analysis functions, and cross-platform support. It enables the conversion, analysis, and visualization of data in both two and three dimensions within a single software environment. In addition, this layer supports three-dimensional spatial operations, relationship assessments, spatial analyses, and other functions. The last layer is the performance layer, which utilizes 3D interaction and output technologies. The performance layer integrates cutting-edge technologies such as WebGL, virtual reality (VR), augmented reality (AR), artificial intelligence (AI), and 3D printing. It offers users a more immersive 3D experience, enhances visualization and interaction capabilities, and showcases the processing results generated by the business layer. The architecture of the new 3D GIS system is shown in Figure 7.

The new 3D GIS achieves the integration of two- and three-dimensional components in data modeling, scene construction, spatial analysis, and software architecture. It amalgamates heterogeneous data from multiple sources and aligns with the objectives of constructing smart cities and digital twins. This technology enables intelligent perception across entire territories, spaces, and elements. It faithfully recreates real-world scenes, furnishing users with visual, intuitive, and three-dimensional environmental and spatial geographic information. The future of the new 3D GIS technology holds promise for significant advancements in standardized construction, massive data processing, and visualization [91].

3.6. The Status Quo of Cloud GIS Architecture

A significant amount of data is generated along with the creation of a new generation of large-scale Internet applications, including social messaging, e-commerce, online video, and several other applications, which brings a huge challenge to hardware and software devices in companies and enterprises [21,71,72]. Therefore, Amazon and other companies posed the idea of “Cloud Computation” in 2006, and as an innovative technology in the information industry, cloud computation gained widespread attention once it was proposed. In the GIS field, with the various map applications, it is required for GIS practitioners to start to consider introducing cloud computing technology to GIS. After that, Cloud GIS technology emerged, which applied cloud computation to every aspect related to GIS, like modeling, storage, and management of spatial data [92,93,94,95]. The development of Cloud GIS cannot be without the support of architecture, and after cloud computing technology has gone from a process-oriented architecture to a system-oriented architecture, it has now developed into a service-oriented architecture. Cloud GIS utilizes a service-oriented architecture [39,41,96,97], and its architecture mainly contains three parts: core service layer, service management layer, and user access interface layer.

The kernel service layer mainly includes SaaS, PaaS, and IaaS [39,94,97]. SaaS is an application program developed based on the cloud computing foundation platform, in which users can join the network without downloading the application program and use map resources and spatial analysis functions through the browser or server. The PaaS layer provides application deployment and management, such as GIS Server, database, etc., which is the abstract package for the development loop. The IaaS layer is user-oriented and uses several servers to construct “Cloud” bias facilities, providing the basic server, storage, and network for users [98]. At present, data management and optimization are still hot issues for this layer. The optimal allocation of hardware resources can be achieved with the introduction of virtualization technology, and the reliability and expandability of services can be provided with the help of virtualization tools such as Xen and KVM. After that, the service management layer is responsible for managing the reliability and security of the core service layer, applying functions such as safety supervision and resource supervision to users [99,100,101,102,103]. Users can access cloud GIS services via the interface using the user access layer, which offers access methods like command line and Web services [104]. The architecture of the Cloud GIS system is shown in Figure 8.

3.7. The Status Quo of Big Data GIS Architecture

Because of the rapid development of internet technology, the volume of big data changes from GB (Gigabyte) to TB (Terabyte) and then jumps to PB (Petabyte). It is challenging for traditional GIS to store and manage such a large amount of data with high refresh rates and to extract more useful information from it because of the growing amount of data, rapid data turnover, and multiple sources of data spawned by various industries, such as healthcare and transportation [105,106].

GIS is a relation-type database at present, but the relation-type database has the disadvantages of low efficiency in storing and querying mass data, difficulty in processing multiple processing requests in parallel with highly concurrent read and write processing, and the expandability of adding service nodes and hardware can only be achieved through downtime, and these problems are critical defects for big data processing and analysis [107]. On the other side, the typical GIS is structural data, but the current big data types are unstructured and irregular data structures that originate from a wide range of sources and have various forms, such as the spatial model, picture, and other report forms, and is hard for typical GIS to dispose and recognize. How to train the parameters, how to optimize and excavate big data algorithms for spatial from different perspectives, and how to choose a diverse spatial statistic model different from typical GIS are the directions that need to be tackled in the current development. The aforementioned issues can be resolved because of its expandability for analyzing big data models [108].

Big data GIS architecture can be divided into four layers. The first one is data storage, which uses a distributed storage system to place data into several mutually independent electronic equipment to reduce the pressure of continuous high-intensity access of the network, resulting in a significant reduction in server load rate. The second layer is the spatial big data GIS components, which can be categorized as data stream processing, data management, spatial analysis, and other components, and these components are the key factors for completing GIS tasks. The third layer is the GIS application server, and the data catalog service, distributed analysis service, data stream service, and framework are located in the middle layer; the operation and management of components all rely on it. The fourth layer is the GIS side, which includes all types of devices, including browsers and mobile terminals.

The emergence of big data urges GIS to develop toward stronger processing ability and accept multiple sources of data. The research on big data GIS development should replace model-driven new spatial analysis methods with data-driven ones and continue intensive research in the direction of real-time dynamic processing and analysis of big data iterations from the data off-line analysis and static management. The SuperMapGIS 9D, for example, is shown in Figure 9.

3.8. The Status Quo of AI GIS Architecture

The generation of AI combining everything has become an irresistible trend, and the machine studying on the branch direction of AI allows computers to make judgments on new sample data by learning to make misjudgments on historical data and other functions that fit the processing and analysis of spatial data in GIS, with the benefit of significant labor cost savings and improved accuracy [109]. Training AI via vast sample models can render AI automatically reveal the unknown objects in the geographic data and discover the rules of known objects. AI GIS is a set of technique systems fitting AI technology and GIS functions [110].

For the AI GIS technique system, the data layer is the bottom layer covering non-relational data, relational data, and file-based data measured on the ground; the domain base layer contains a sample and a model and constructs the model with different data types by integrating diverse samples, moving up to the framework layer, to minimize unnecessary repetitive development tasks, and appropriate computer technologies are employed. This layer involves extracting shared characteristic features from a multitude of entities and organizing the obtained information into a coherent whole. Lastly, it manages external encapsulated information while facilitating internal operations, thus amalgamating various AI-based framework structures. The uppermost layer is the function layer of AI GIS architecture, empowered by GIS to AI, and AI manages the non-geographic spatial data results for spatial analysis and visual representation; enabling GIS by AI improves and strengthens GIS functions and the ability to analyze and manage data; AI GIS consists of three parts. As the fundamental underpinning tools for AI, AI workflow tools provide training tools for GeoAI algorithms. Using data to create a more accurate model, training models may be applied to many scales and geographies. However, deep-learning and spatial-learning algorithms construct GeoAI, and the AI workflow tool and GeoAI product construct the AI GIS algorithm [111]. The SuperMapGIS AI, for example, is shown in Figure 10.

The combination of AI and GIS makes a significant breakthrough for the typical data processing and analyzing, and the served areas at present belong to a weak artificial intelligence field that possesses a certain intelligent behavior in a specific area to handle something dangerous and tedious in place of someone. It is still far from having strong artificial intelligence with autonomous choice behavior. Based on the current 2D extraction on the flat images, the intelligent acquisition of three-dimensional information, artificial general intelligence systems, and the direction of combining brain-like and acquired training are still difficulties and the emphases of AI GIS development [112,113].

4. Review and Discussion

After decades of development, GIS has been integrated into the mainstream of information technology and will continue to be an important part of IT. Its significant advancement could present both opportunities and challenges to GIS. In addition to IT, the application of artificial intelligence could also drive the development and innovation of spatial information science, such as the application of GeoAI based on machine learning and deep learning in target detection, binary classification, feature classification, and scene classification. In GIS-center geoinformation science, artificial intelligence and remote sensing big data have facilitated the evolution of geoinformation science towards intelligence and promoted the reform of the geoscience research paradigm. In addition, the use of digital technology to create virtual models of objects throughout their life cycle enables predictive and control operations on physical models in the real world. Digital technologies such as digital twin technology, metaverse technology as a network of three-dimensional virtual worlds, and depth mapping where brain science intersects with cartography [114], all play a significant role in contributing to the various directions of GIS development. The application of equal distance weight feature space optimization model [115,116] and other applications of space calculation algorithm is a multi-directional expansion of spatial big data, which further makes the development and application of GIS have a deeper expression.

High-performance GIS, both at home and abroad, is an important frontier direction of geographic information system technology. The rapid increase in the level of computer hardware has led to the simultaneous improvement of computer performance. Among them, the development of parallel computing, distributed network, and cluster technology, which uses multiple processors to integrate computing resources to solve problems, has pushed the performance of computers to the level of high-performance computing and applied it to traditional geographic data processing. In the Internet era, it is an important direction for the development of GIS to apply the hardware system of high-performance computing to the field of GIS to solve the problem. The development of GIS is inseparable from high-intelligence spatial computing. It promotes the evolution of GIS architecture, which can process a large amount of data, models, and algorithms more efficiently and drive more accurate and rapid spatial analysis, more efficient spatial modeling, and more intelligent spatial prediction.

As an important part of high-performance GIS, high-performance GIS algorithms enable the GIS system involved in the design to simulate and efficiently process large-scale, massive spatiotemporal phenomena and data. A new exploration of spatial analysis and the optimization of high-density computing are two main research aspects of high-performance GIS algorithms. As two kinds of geospatial storage formats, raster data and vector data are topics that GIS cannot handle. Due to the difference in the data structure between the two, there are differences in parallel computing, but from the perspective of division strategy, they are divided into static and dynamic parts. Trajectory big data spatiotemporal analysis and spatiotemporal index technology are two important parts of high-performance GIS algorithms. Parallel computing is another important part of high-performance GIS. Its essence is a super computing method. It will use the efficient data transmission between computing nodes connected by the network to improve computational efficiency. The three typical hardware architecture models of parallel computing have their own characteristics: the shared memory model of multi-task parallel computing and storage address sharing, the message passing model of mutual communication, and a stream processing model for increasing the number of processors. High-performance memory computing is another important part of high-performance GIS [117,118]. It consists of three parts: distributed cache computing to improve access efficiency, computing network for local execution data, and distributed in-memory database system. High-performance in-memory computing is a hot issue in high-performance GIS and solves many problems in the direction of geographic computing. Core computing is the last component of high-performance GIS. It refers to the integration of a processor by multiple independent engine cores so that it can perform parallel operations to improve computing performance.

Massive GIS big data have promoted the continuous progress of high-performance GIS research, made significant contributions to geographic phenomenon analysis and geographic data processing, and greatly improved the efficiency of geographic analysis. At present, the development of high-performance GIS has been significantly improved in high-performance algorithms, index mechanisms, parallel computing, memory computing, and spatial data storage. It optimizes the operation process through a more concise model, completes high-performance computing within the basic system, reduces costs, and makes it possible for ordinary computers to participate in high-performance computing.

5. Conclusions

The evolution of GIS architecture is an ongoing process driven by application demands and technological advancements [119]. From the early days of component-based GIS to today’s high-performance geographic information systems such as AI GIS, big data GIS, and cloud GIS, significant progress has been made. This evolution is reflected not only in the complexity and performance of the systems but also in their ability to handle diverse and complex geographic data, supporting decision making. Traditional GIS architectures faced challenges such as the complexity of data content, diverse service systems, and the demands of large-scale users. These challenges are gradually being addressed in the introduction of new technologies. The integration of cloud computing technology has enhanced resource management and data maintenance capabilities, enabling GIS to efficiently process large-scale data and support real-time analysis. Concurrently, the application of artificial intelligence and deep learning has made GIS systems more intelligent, improving spatial and temporal analysis as well as visualization capabilities. This trend of technological integration will continue to shape the future of GIS development, providing powerful tools for addressing a wide range of complex geographical problems. The evolution of the GIS architecture and the direction of its development is shown in Figure 11.

In summary, the development of GIS architecture has made it a more scientific, efficient, engineering-oriented, and application-focused technology [90,120]. In the future, GIS will continue to adapt to evolving demands, fostering interdisciplinary collaboration and cross-domain integration to better serve various industries. The direction of GIS evolution will prioritize cost savings, improved processing efficiency, and user-friendly operations to meet the needs of its users.

6. Future Research Directions

In the era of the digital economy, the development of deep learning, big data, and cloud technology has promoted the innovation of GIS technology. These new technologies are further applied to GIS to solve various problems in GIS data processing and analysis, such as calculation, network communication, and spatiotemporal data-intensive issues [121,122,123,124,125]. In the future, in the face of the increasing complexity of geographic data and the upgrading of computer hardware, the integration of parallel spatial databases into GIS systems will move towards new areas and solve new problems. At the same time, it is still necessary to continuously improve the algorithm, improve the index mechanism of the traditional algorithm, and focus more on the independent platform’s own redundancy control to improve the index effect and expand the performance of the GIS system.

The rapid development of information technology will lead edge computing GIS, cloud-native GIS, geographic blockchain, digital twin, and metaverse to become research hotspots for GIS. The framework of the GIS software system runs through the whole life cycle of GIS development and represents the way GIS is organized. Studying the GIS framework system is of great significance to improve the operational efficiency of the GIS software system. It is inevitable for a multidisciplinary GIS to cooperate with metaverse, digital twin, and other emerging technologies. In this paper, the characteristics of the current mainstream GIS framework system and its development trend are comprehensively studied and analyzed to provide a guiding reference for GIS system framework research [126].