Stochastic modeling and performance analysis of multimedia SoCs

Abstract

Abstract-Quality of video and audio output is a design-time constraint for portable multimedia devices. Unfortunately, there is a huge cost (e.g. buffer size) incurred to deterministically guarantee good playout quality; the worst-case workload and the timing behavior can be significantly larger than the average-case due to high variability in a multimedia system. In future mobile devices, the playout buffer size is expected to increase, so, buffer dimensioning will remain as an important problem in system design. We propose a probabilistic analytical framework that enables low-cost system design and provides bounds for playing acceptable multimedia quality. We compare our approach with a framework comprising both simulation and statistical model checking, built to simulate large embedded systems in detail. Our results show significant reduction in output buffer size compared to deterministic frameworks.

Performance evaluation of stochastic real-time systems with the SBIP framework

“…SMC explores a sample of the stochastic model execution traces produced through simulation in order to estimate the probability for the system to satisfy that property. This statistical approach is receiving an increasing attention and is being adopted in a wide range of application domains such as in biology (David et al, 2015a), for the assessment of communication protocols (Basu et al, 2010a), multimedia applications (Raman et al, 2013), and avionics (Basu et al, 2010b).…”

“…In this study, the SBIP framework is used to check QoS properties of an MPEG2 decoder subsystem part of a video streaming application (Raman et al, 2013). This work is about finding a good trade-off, when designing the multimedia system, between the required sizes of the system buffers and the quality of delivered videos.…”

“…This work is about finding a good trade-off, when designing the multimedia system, between the required sizes of the system buffers and the quality of delivered videos. It is known in the literature that an acceptable amount of quality degradation -defined as less than two consecutive frames within one second -can be tolerated in order to reduce buffers sizes (Raman et al, 2013). The study will consist to assess this requirement.…”

“…This component is made of three subcomponents operating as follows. The first component (left) generates frames following the MPEG2 GOP pattern (Raman et al, 2013). Each frame is then decomposed, in the second component, into 330 macro-blocks, which are finally transmitted to the third component that models the macro-blocks stochastic arrival time to the input buffer with respect to ρ mtb .…”

An improved fast MIPv6 using dynamic threshold buffer management

“…PAR and NAR and the remaining one at MN as playout buffer. They work hand in hand and player at MN starts to read Macro Blocks from the playout buffer after a defined initial delay namely playout delay [25]. During L2 switching, the content from buffer at PAR passes to buffer at NAR.…”