Documente Academic
Documente Profesional
Documente Cultură
ABSTRACT
In this paper, we present a reactive control policy which adapts the source bit rate to the
reserved resources in order to ensure performance guarantees for multimedia applications.
The proposed method called flatness based trajectory tracking deals with drastic traffic
flow rate changes and limits the traffic in order to respect the time constraints. We show
the contribution of the reactive control and the dynamic regulation using purely control
theoretic approaches which stabilize the network and avoid undesirable oscillations for
the transmission of such critical flows. Here, we present a performance analysis for such
rate control mechanism, and illustrate its feasibility through its implementation on
MPLS-TE control plane of SSFnet/Glass simulator.
The control is performed by adjusting the Classically, most papers consider the problem
network inputs, i.e. application’s input rates which from the network dimensioning point of view: given
affect buffer sizes with respect to the buffer planning. an input stream and a scheduling policy, what is the
The need for regulatory control arises first to track worst case buffer requirement and what is the nature
the reserve capacity, and second to enforce the of the output stream? However, in our case the
application service guarantees. output stream and the buffer size are given and
described by the trajectory. So, we the input stream
Since the measured outputs are used to to meet buffer constraints and to maintain QoS. We
determine the control inputs, and the inputs then use a nonlinear theoretical approach to resolve this
affect the outputs, the mechanism is called feedback reverse problem, named flatness based trajectory
or closed loop. Feedback control based on flatness tracking control [14], [15], [16].
notion is a powerful tool 1) to ensure that the
measured output (buffer utilization) tracks very In the following, we first state the problem
closely the reference input (buffer planning) even in definition, and then briefly recall flatness notion and
the presence of disturbances, 2) to deal with network expose the trajectory tracking control method with
stability and other aspects of control performance, reference to the target environment of figure Fig.1.
especially when changes in network workloads and The packet stream is processed and controlled at the
configuration occur. node denoted Rc and, the router resources considered
here are the buffer size and the bandwidth.
2.3 Our Work Objective
Depending on network variations, we use a 3.1 Trajectory Tracking Control Methodology
feedback mechanism to inform the sources when QoS control requires an understanding of the
they exceed their profile and regulate their input quantitative parameters at the application and
rates in order to match the reserved network resource network layers. For example, from a traffic
and to meet their QoS requirements. As a result, descriptor which describes the requested service, we
dynamic adaptation is provided between clients and deduce what will be the node buffer variation of the
their service provider. corresponding LSP and we define the buffer
trajectory, denoted as qref (example in section 5.1).
The purpose of this work is to present this new
Traffic Engineering approach, which aims to Recall that we assume a MPLS-TE architecture.
optimize the network utilization and performance by For each incoming stream an LSP is established and,
intelligently handling the buffer reservations at the a buffer queue is created to support the
routers. We take benefit of traffic descriptors to corresponding packets. The critical router Rc of the
model communication behaviours and QoS figure Fig.1, collects the packets generated by n
requirements by using trajectories. earlier routers in its buffer q(t). The incoming rates
Trajectory would be a new way of mapping traffic of these packets are denoted ui(t). The aggregated
demands over networks. The trajectory establishment packets are served with some service bit rate r(q(t))
or network resource planning translation into to the next hop following their LSP. If Rc is not able
trajectories is out of the scope of this paper. Here, we to handle all the incoming packets, the packets are
are interested mainly in source bit rate regulation to either buffered to wait for transmission service or
meet QoS requirements and bounding packet delay. rejected in case of buffer saturation. The lack of
feedback between adjacent routers may cause an
3 RATE REGULATION SCHEME excessive data loss and bad transmission service
(delay violation) for these critical flows.
Our method called Flatness-based trajectory
tracking control is a network-driven intra-domain
and inter-domain layer 3 bandwidth provision
approach. We aim to prove its efficiency, when
applied at the hotspot nodes such as the ingress
nodes of large scale networks like Internet, and also,
the nodes that aggregate flows of several LSPs
(Label switched Path), in case of an MPLS-TE
domain (as in figure Fig.1). Network resources
q& i (t ) = ui (t − hi ) − α i µ
qi (t )
(2) = − K i (qi (t ) − qiref (t ))
n
1 + ∑ qi (t )
i=1 qi (t + hi )
ui (t ) = − K i e(t + hi ) + α i µ n
qi (t )
Figure 7: Queuing delay variation at the
ri (qi (t ))
Figure 4: LSRs buffer reservation qiref LSPs buffers
5. REFERENCES
Figure 9: LSPs bandwidth utilization ratio [3] K. Lee. Modèle Globale pour la QoS dans les
Réseaux FAI: Intégration de DiffServ et MPLS-TE.
The flatness control mechanism successfully PhD thesis, Ecole Centrale de Lille, (2006).
prevents the router's buffer oscillation and the
queuing delay variation by tracking the reserved [4] S. Black, D. Black, M. Carlson, E. Davies, Z.
buffer size. Our approach is very efficient for Wang, and W. Weiss: An Architecture for
managing network resources as it maximizes the Differentiated Services, Network Working Group
buffer utilization (see figure Fig.9), in comparison to IETF-RFC 2475, (1998).
the reserved buffers.
[5] Y. Berny, S. Black, D. Grossman, A. Smith: An
6. CONCLUSION Informal Management Model for DiffServ Routers,
(Appendix. A Discussion of Token Buckets and
The new traffic engineering approach presented Leaky Buckets), IETF-Network Working Group,
here, aims at optimizing the network utilization and RFC No.3290, (2002).
performance by intelligently handling the buffer
reservation at the routers. We take benefit of traffic [6] W. Ashmawi, R. Guérin, S. Wolf, M. H. Pinson:
descriptors to model communication behaviour and On the Impact of Policing and Rate Guarantees in
QoS requirement by using trajectories. We show that DiffServ Networks: a Video Streaming Application
the reactive control and the dynamic regulation using Perspective. In ACM SIGCOMM 2001, pages 83–95,
purely control theoretic approaches stabilize the San Diego, CA, USA, (2001).
network and avoid undesirable oscillations for
critical flow transmissions.