Preliminaries - Real-Time Rendering: Computer Graphics with Control Engineering - page 16

TABLE 2.1
Results from the Research Review Classification

Polygon-based Rendering Non-Polygon-based

With Frame Rate Data [62] [64] [65] [67] [72] [74] [77] [81] [84]

[85] [92]

[69] [73] [80] [83] [86] [87]

Without Frame Rate Data [61] [63] [66] [68] [75] [78] [79] [82] [89]

[90] [91]

[70] [71] [76] [88]

The next step in this comparative study is to select literature which provides
experimental results on frame rate control since we need to conduct a qualitative
analysis on them. For this purpose, these results should contain a history of data
points in the time domain that demonstrate certain desirable qualities such as stabil-
ity, offset errors and smooth frame rate transitions. These features would be com-
pared to the results we obtain from the experiments conducted using the techniques
proposed in this research with both qualitative and quantitative perspectives.
While research papers could be found relating to the topic of interactive rendering,
however many of them alluded to concrete experiment results on sustainable perfor-
mance as shown in the references from the bottom row of Table 2.1. In some cases
[61] [71] [76], only static frame rates are given as an approximation to the interactive
requirement. Furthermore, other researchers have chosen to work on volumetric [71]
[76] [80] and image-based rendering [68] [70] [83] techniques which are prevalent
in medical and large-scale visualization research but they differ from polygon-based
rendering vastly. As a result, it is not straightforward to establish a direct comparison
on the benefits offered by our research with these techniques. Despite these differ-
ences, we strive to provide a detailed qualitative and quantitative analysis on the
aforementioned rendering architectures and their respective performance with our
rendering framework in Chapter 8, Section 8.1.

2.3.2 c ontRol -t heoRetic a PPRoaches to c omPuteR s ystems
As computer systems become increasingly complex through advances in hardware
and software technology, traditional approaches to providing performance guaran-
tees have become inefficient. In recent years, control engineering principles used
successfully in real-world applications such as mechanical and electrical systems
and process control have emerged as promising solutions to meet performance con-
trol challenges such as real-time scheduling, network bandwidth control, and power
management in complex computer systems.
The comprehensive framework presented by Abdelzaher et al. [2] introduces
feedback performance control in software services. The authors emphasised the
importance of guaranteed quality of service (QoS) in modern computer software
and systems that indicates the need for robust frameworks to achieve certain perfor-
mance objectives. They further defined and explained the attributes of a QoS-aware
service consisting of performance metrics such as queuing delays, execution
latencies, and service response times. They also demonstrated that a software