Server Architectures Using a Fat

Server Architectures

Using a Fat Client/Server Architecture in Computer-Aided Design and Drafting Workflows

The continual debate between the inherent value of thin client/server architectures that can be configured relatively quickly to the more intensive levels of planning, implementation and continual maintenance of thick client/server networks need to be delineated by use cases and applications used. Thin client network's speed to configure and ubiquity of support needs to balanced against the bandwidth requirements that more data-intensive applications require over time (Guynes, Windsor, 2011). The intent of this analysis is to illustrate how an engineering consulting and design organization will require a thick client/server architecture given the size, complexity and continual editing necessary to get projects done on time across their global operations. Computer-Aided Design (CAD) files often require intensive levels of cross-integration for support and reference files including symbol libraries as well. These prerequisites for effective CAD system use make a thick client/server architecture ideal for this specific need.

CAD Designers Need Real-Time Data Integration To Complete Tasks

The driving catalyst of having a thick client/server architecture to support teams of CAD designers globally is the intensive level of data integration inherent in their specific design files and supporting imagery, vector data and supporting application designs. A thin client network, predicated on small, relatively easily scaled transactions, is well suited for conversational and quick interchanges as occur in e-mail and low-end collaboration systems (Lee, 2002). A thin client network also requires continual contact with servers to keep applications continually performing as well (Schmerken, 1997). Both of these aspects of a thin client network architecture would be disastrous for a network of CAD designers working for an engineering and design consulting firm.

A network of CAD engineers and designers would require a more intensive level of transaction support and scalability of network performance to support shared visualizations of their models as well. One of the criterion for thick client architecture performance is the need for optimizing multimedia performance (Lai, Nieh, 2006). For the teams of CAD designers located globally, the ability to literally walk through each of their models and show the outcomes of design decisions could save literally thousands of dollars in travel costs per year. Technologies exist today that can accomplish this, and they are predicated on the advantages of a thick client/server architecture.

As CAD designers and developers rely on a highly coordinated, synchronized approach to completing projects called concurrent engineering, thick client networks fit with this requirement much more effectively than their thin client counterparts. Thick client technology can support more efficient asynchronous transfers and the independence of data models at the individual worker level can significantly improve the overall performance of a given project. Thick client/server networks are especially well suited for concurrent engineering tasks given their data management and overhead requirement specifications (Lee, 2002). In the field of engineering consulting and design, it is also critically important for companies to have secured networks capable of collaborative design sessions and concurrent design sessions. The thin client technology alone can't scale to this requirement, and when companies have pushed this technology to this level of performance, security compliances have become commonplace (Vlissidis, Hickey, 2010). In other words, even if an engineering and design company strove to create a thin client/server network to support its collaborative engineering and concurrent workflows, it would fail on the security aspects of performance alone.