This work presents a simulation-based tool for dependability-oriented design of Video on Demand (VoD) applications. The tool is organized in a layered architecture, so that simulation models can be built and detailed according to a hierarchical and modular
retransmit it. Instead, if at some time t1 a fault corrupts amessage traveling on a TCP Link Object, the message isretransmitted at a later time t2.
Figure 2 – Dependability analysis features of
Network Level objects
As shown in the pictures, all objects provide features toinvestigate the effects of both transient and permanentfaults. They also provide support to readily evaluate theimpact of hardware redundancy techniques on systemdependability.
Since the tool allows for the use of arbitrary C++ code,the tool’s basic objects can be augmented, combined, andcustomized, in order to more accurately model applicationand/or platform specific characteristics.
3. Case study
The system under study is a prototype Video onDemand (VoD) distributed application currently underdevelopment at the University of Naples. The applicationis called DIstributed Video Architecture (DiVA) [9] and itcan be operated both in a WAN and in a LAN scenario.
3.1 System Architecture
The DiVA system relies on three distinct kinds ofprocesses, the Video Server, the Video Client, and theFilm Server. Such processes interact according to a client-
server paradigm and communicate via an IP-basednetwork infrastructure. The system allows for retrievaland real-time delivery of video and audio documents. Thearchitecture of DiVA is schematically represented inFigure 3.

Figure 3 – Components of the DiVA systemVideo/audio documents are distributed over the network.At each computing node, documents are managed by theFilm Server (FS) process, running on the local host. TheVideo Server (VS) process, which is – in the most typicalconfiguration – a centralized entity, provides the VideoClients (VCs) with functions to search the database andselect the clips. Users may require the list of availablefilms and then choose one of them for display. Multiplesessions may be active concurrently. Video is stored inMPEG format [7]. Clients and servers are synchronizedthroughout the transmission of audio/video data.
An experimental testbed of the DiVA system was used asa case study for demonstrating some of the potentialitiesof the simulation tool we have developed.
The testbed, illustrated in Figure 4, was configured asfollows:
A Video Server process - running on a Sun Sparc 4 -located in Milan.
Four Video Client processes - running on two Sparcs
and two Pentium II PCs - located in Milan, and twoVideo Client processes - running on two Sparcs -located in Naples.
Two Video Archive processes - running on two
Sparcs - located in Milan, and one - running on aPentium II PC - located in Naples.
The two Sparcs located in Milan, hosting the Video Clientprocesses, were equipped with Parallax cards, forhardware decoding of the video streams. The computers inMilan and Naples were connected via a 100 Mb/sec and a10 Mb/sec Ethernet switch, respectively. A 2 Mb/secnetwork infrastructure connected the two LANs.

for the Design of Integrated Dependable Real-Time...A simulated annealing approach is utilized to ?nd...Each application introduces system level constraints ...
simulated fault-injection, Video on Demand INTRODUCTION...dependable applications must be retargeted, in ...(VoD) distributed application currently under ...
a Video on Demand (VoD) distributed application....fault injection experiments on a simulated model ...Tool for dependabilityoriented Design of Multimedia...
fault injection, where data is altered and/or timing of an application is...Two techniques have been proposed to inject faults into a simulated model [...
simulated fault-injection, Video on Demand INTRODUCTION...dependable applications must be retargeted, in ...(VoD) distributed application currently under ...
热门文档