Selecting the Right Computing Architecture for Your GIS

From centralized to distributed operations, choosing the right architecture pattern for your GIS can improve user productivity and reduce operational costs.
[Note:  This is latest post in our series about Managing GIS.]
There are benefits of centralized operations, and there are some reasons why distributed operations may be preferred by some organizations. We can find reasons why a distributed operation may work better than a centralized one. Distributed operations require more hardware, higher administration cost, higher implementation risk, more data access problems (to centralized data sources), and reduced security. Selecting the right architecture pattern to best meet your business needs can improve user productivity and reduce operational costs.

Centralized vs. Distributed GIS Operations

Centralized operations share a central geodatabase data source. Distributed architectures require replicated copies of business data at remote locations, establishing distributed processing nodes that must maintain consistency with the central database environment.

Many organizations are moving to centralized operations to reduce overall cost and simplify system administration. Distributed solution architecture is important as GIS expands to include a variety of federated operations.
A central database architecture provides one source for the production database environment, minimizing administrative management requirements and ensuring data integrity. GIS desktop applications on the central LAN have direct access to local GIS data sources. Remote user access to central data sources can be supported by central Terminal Server farms, providing low-bandwidth display and control of application environments maintained in the Data Center.
Benefits of a centralized architecture include reduced hardware cost, reduced administration cost, lower implementation risk, improved data access, improved security, and reduced wide-area network traffic .

Best Practice: Centralized architecture generally supports the lowest-cost and lowest-risk operations.

Distributed architectures require replicated copies of business data at remote locations. Data consistency will require controlled procedures with appropriate commit logic to ensure that changes are replicated to the associated data servers. Distributed database environments will generally increase initial system cost (more hardware and database software requirements) and demand additional ongoing system administration and system maintenance requirements.
The most common reasons for distributed architecture requirements are organizational precedence (regional system collaboration), management confidence (critical support requirements), physical security (protect confidential information), and infrastructure limitations (limited WAN connectivity).

Best Practice: When designing a system, centralized operations architecture should be established as the baseline. Distributed operations will cost more to operate and manage, and a business case should be prepared and reviewed for each distributed organization to justify these additional costs.

Enterprise Architecture Deployment Strategies

GIS software and computer infrastructure technology continue to expand GIS deployment capabilities and introduce new business opportunities. New architecture patterns are emerging that reduce administration complexity, provide more adaptive deployment opportunities, and integrate user workflows throughout the organization and the user community. Distributed geodatabase replication technology integrates a variety of desktop, mobile, and server solutions into an adaptive geospatial communications environment connecting operations across the enterprise and throughout the community.
The figure below shows a variety of architecture options available for enterprise GIS deployment.

A variety of system architecture strategies are used to manage enterprise GIS operations.
ArcGIS Server mapping services can be deployed directly from the data center, or geodatabase replication services can be used to provide incremental updates to ArcGIS Server web services maintained within a private or public cloud hosting infrastructure. The Cloud computing infrastructure provides a new adaptive platform environment for managing high capacity map publishing services.
Organizations are expanding operations to incorporate mobile users as an integral part of their enterprise workflow. Improved availability and capacity of wireless technology support mobile communication connectivity for a growing number of GIS users.

Federated Architecture Deployment Strategies

Database and Web technology standards provide new opportunities to better manage and support user access to a rapidly growing volume of geospatial data resources. Web services and rich communication protocols support efficient data migration between distributed databases and storage locations. Web search engines and standard Web mapping services support integrated geospatial information products published from a common portal environment with data provided from a variety of distributed service locations. Federated architectures identified in the figure below provide better data management, integrating community and national GIS operations. Geodatabase replication services and managed extract, transform, and load (ETL) processes support loosely coupled distributed geodatabase environments.
Federated architecture patterns support community and global business requirements, providing layers of managed and published data sources at different levels of granularity.
Multi-layered data architecture patterns are becoming more common. Solutions include the standard desktop, server, and mobile patterns for each implementation level.

Best Practice: Multiple database layers improve security and control level of data dissemination.

Community Deployment Strategies

ArcGIS Online is building community relationships that change the way people work. ArcGIS provides a platform for collaboration, sharing, and community analysis that helps us better define and understand the world.

Cloud-based applications integrate and synthesize information from many sources, facilitating communication and collaboration, and are breaking down barriers between institutions, disciplines, and cultures.
Technology is changing the way GIS is serving our communities in a number of different ways, including:

Service-Oriented Architecture

The figure below shows the basic SOA structure. Technology change is again being influenced by general acceptance of standard Web communication protocols and more stable and available network bandwidth connectivity. Software development is taking advantage of Internet communication standards and network connectivity with an evolving service-oriented enterprise architecture strategy.

Service (REST)-oriented architecture allows programmers to take advantage of Internet communication standards and network connectivity to build applications that share services and data sources on a global scale.
A services-oriented architecture (SOA) is an approach for building distributed computing systems, based on encapsulating business functions as services. Services can easily be accessed in a loosely coupled fashion.

GIS in a Service-Oriented Architecture

Esri embraced open standards during the 1990s and has actively participated in the Open GIS Consortium and a variety of other standards bodies in an effort to promote open GIS technology. The ArcGIS system of software is developed from the ground up, to support interoperability and data sharing.
ArcGIS software promotes sharing from Desktop to Server to a broad range of client devices through a services framework.
ArcGIS SOA framework features an authoring tier of professional ArcGIS for Desktop users, a publishing tier of services, and a presentation tier of viewers with access to available published services.

Best Practice: ArcGIS Online can be used as a portal for publishing enterprise-wide web services.

The ArcGIS system is designed to leverage geospatial operations in the workplace. The SOA framework includes multiple access layers connecting producers and consumers, based on current client/software technology and incorporating web application and service communication tiers. Consumers connect to producers through a variety of communication paths.
GIS is by nature a service-oriented technology with inherent fundamental characteristics that bring diverse information systems together to support real-world decisions. GIS technology flourishes in a data-rich environment, and ArcGIS technology can help ease the transition from existing “stovepipe” GIS environments. The geodatabase technology provides a spatial framework for establishing and managing integrated business operations.

Best Practice: Build enterprise operations that leverage an SOA to author, publish, and serve intelligent maps and create, analyze, and share geospatial information.

Understanding SOA and how it enables business process integration and helps control and manage technology change is important. Organizations must build an infrastructure that can effectively take advantage of new technology to stay competitive and productive in today’s rapidly changing environment.


This post is an excerpt from Dave Peters’ book Building a GIS: System Architecture Design Strategies for Managers.  Extensive information about successful system design can also be found on the System Design Strategies wiki and in the System Architecture Design Strategies training class

Next Article

The Science of Where for Sustainable Development Webinar Series

Read this article