A Visual Language for Modeling and Analyzing Printer Data Path Architectures

The data path is an important printer component that performs real-time image processing. Because of the large data sizes and high throughput requirements in high-performance multifunctional printers, the data path is usually implemented as a hybrid software/hardware system. Designing the architecture of a data path is a nontrivial problem, because of the many tradeoffs involved and because it is difficult to analyze how well a design conforms to many of the important quality attributes. One quality attribute that is difficult to deduce from a data path design by hand is its throughput, typically expressed as the amount of pages that it can handle per minute. Because this is hard to analyze by hand, it is also difficult to predict how well a data path will perform when requirements change. This limits how structured and flexible the data path design process can be. In order to improve this process, a solution is needed that helps architects to analyze the behavior of a model of a data path architecture, find its throughput and identify bottlenecks. The solution must be sustainable, so that support for increasingly detailed architecture models can be added. Our aim is to enable these goals by providing a domain-specific modeling language for data path architectures, a modeling environment that supports this language, and a set of tools for model conversion and analysis. The language also allows transformations to timed automata, colored Petri nets and synchronous data flow graphs. This paper concentrates on the modeling language and its design rationale. 1. Domain analysis The data path of a printer consists of the image processing steps that need to take place on the stages between for example network input and the print head (network printing) or between the scanner and the print head (copying). Figure 1–1 shows the place of the data path in a heavily simplified illustration of a multifunctional printer. Figure 1–1. A schematic illustration of a multifunctional printer, showing the place of the data path with respect to the print process, the paper path, the scanner and the network. To the right, a typical multifunctional office printer with a similar layout is shown. The data path performs necessary image processing steps to improve quality, to conform to the user’s settings but also to overcome limitations of scanning and printing hardware [3]. For example, when scanning a white page with text, the user expects an image that has a white background. Scanners, however, typically detect a light shade of gray, rather than white. Figure 1–2 shows how image processing can help to produce the desired results.