A Network Resource Management Framework for Multimedia Applications Distributed in Heterogeneous Home Networks

Journal of Systems Architecture

2013

Self Cite

International audienceThe home network is an open, heterogeneous and distributed environment where ensuring multimedia ap-plications' quality of service is a main concern. Mechanisms to reserve resources (CPU, memory, network) and architectures using them already exist. However, they require to modify the devices or the applications and they do not take into account the heterogeneity of the home network. This paper presents a non-intrusive and adaptable resource management framework. This framework is developed upon an architecture, customized for the actual devices. This architecture uses global components that delegate to local components the management of local resources. These components rely on the resource reservation mechanisms provided by the Linux operating system in order to guarantee the resources to the applications. The framework has been implemented on real devices (PCs, laptops and embedded multimedia devices), bridged with wireless and Ethernet networks. The evaluations of the framework show that reservations are guaranteed even if noise is generated on the resources, which also guaranties the expected quality of service

Section: Introductionmentioning

confidence: 87%

Resource management for multimedia applications, distributed in open and heterogeneous home networks

Louvel

Plantec

Journal of Systems Architecture

2013

Self Cite

“…The reservationPolicy controls the network interface of the server device as described in [12]. It limits the default outgoing traffic to the capacity of the network interface minus the reservation made for each streaming application.…”

Section: Methodsmentioning

confidence: 99%

“…Two trivial solutions exist to perform such a solution. The first one is based on a constant reservation which equals to the maximal value of necessary bandwidth [12,14,17]. But because VBR includes a lot of peak rates, significantly larger than average rate, this solution leads to an over-reservation of the resource.…”

Section: Introductionmentioning

confidence: 99%

“…It is the role of a central controller [14,17] to manage resources and QoS. Having a strong control on resources behavior, it performs a bandwidth admission test and reservations on the network interface of the server and on the network itself [12,16,17]. The reservation for the server is done by scheduling and limiting the outgoing traffic.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

VBR Video Abstraction for Home-Network Reservation

Lemarchand

Louvel

Lecture Notes in Electrical Engineering

2012

Self Cite

“…But, in the specific context of closed systems, since we have a full knowledge of contents and architecture, we argue that we can control resource usage precisely using a centralized architecture and an accurate knowledge of bitrate to transmit [4]. In this approach, the network admission is classically solved by limiting the outgoing traffic on server devices with a token bucket that shapes the outgoing traffic.…”

Section: Introductionmentioning

confidence: 98%

Dynamic Server Configuration for Multiple Streaming in a Home Network

Lemarchand

Mensah

2014 12th IEEE International Conference on Embedded and Ubiquitous Computing

2014

International audienceIn home network, to manage network bandwidthusage, one solution is to control the server outgoing traffic witha token bucket policy. Hull-based token bucket parameter settingallows the guarantee of Quality of Service for variable bitratevideo streaming. Hull is an abstraction of the bitrate, followingthe bandwidth requirement evolution dynamically. In the case ofmulti-streaming, we investigate a shifting technique to reducethe peaks impact. Postponing the streaming starting time ofa video helps to decrease the maximum required bandwidth.The technique is then mixed with the hull-based reservation.Simulations show the effectiveness of the combined approachesto optimize bandwidth usage, guaranteeing the best QoS forstreaming. Online utilization is also discussed