Provisioning for interdomain quality of service: the MESCAL approach
- ISSN: 01636804
- DOI: 10.1109/MCOM.2005.1452841
Abstract
This article presents an architecture for supporting interdomain QoS across the multi-provider global Internet. While most research to date has focused on supporting QoS within a single administrative domain, mature solutions are not yet available for the provision of QoS across multiple domains administered by different organizations. The architecture described in this article encompasses the full set of functions required in the management (service and resource), control and data planes for the provision of end-to-end QoS-based IP connectivity services. We use the concept of QoS classes and show how these can be cascaded using service level specifications (SLSs) agreed between BGP peer domains to construct a defined end-to-end QoS. We illustrate the architecture by describing a typical operational scenario.
Provisioning for interdomain quality of service: the MESCAL approach
ENABLING INTERPROVIDER SERVICE QUALITY
INTRODUCTION
Network quality of service (QoS) is a key consider-
ation for future multiservice networks, as the
demands placed on the Internet continue to
increase with deployment of multimedia applica-
tions and distributed data retrieval systems. Extend-
ing the current best effort Internet to support QoS
is thus recognized as an important next step [1].
Most research to date has focused on sup-
porting QoS within a single administrative
domain. However, delivery of end-to-end QoS to
support end-user applications requires that
autonomous systems (ASs) administered by dif-
ferent organizations cooperate to deliver the
required level of service. The problem we there-
fore seek to address is how to provide QoS
across multiple domains in a way that takes into
account the commercial Internet’s multi-organi-
zational structure, builds incrementally on exist-
ing protocols and approaches, and is scalable.
Compared to the intradomain case, the abili-
ty to deliver interdomain QoS requires different
IP network providers (INPs) to negotiate service
contracts with each other and to engineer their
networks to provide the required level of perfor-
mance. The service contracts are called service
level agreements (SLAs), and they include a
technical component that is called a service level
specification (SLS). The contracts specify the
relationship between an INP and either its cus-
tomers or a peer INP. A key aspect of the SLS is
definition of the QoS classes an INP can offer
for its customers’ traffic. Each INP then has to
provision and configure its network resources so
that traffic is forwarded in accordance with the
agreed QoS levels. Thus traffic engineering (TE)
plays an important role in achieving end-to-end
QoS, enabling the network to deliver defined
performance (measured typically in terms of
throughput, delay, and packet loss) while also
optimizing the use of network resources.
The ability to support interdomain QoS pro-
vides a number of challenges: awareness of QoS
capabilities in other domains (for both long-term
planning and provisioning, and short-term
dynamic response); the ability to engineer
domains to deliver QoS; and to achieve all this in
a way that is scalable. In this article we propose a
framework that brings together all the functions
described above in a way that meets these chal-
lenges. The framework therefore encompasses
business-related processing of service planning
and exchange of QoS capabilities between pro-
viders; QoS-based inter- and intra-domain TE in
the management plane; QoS-enabled routing at
the control plane; and traffic enforcement in the
physical network at the data plane. Our approach
does not require that automated processes always
be used to implement the functions described
here; many of the management functions could
be implemented by manual processes, or by man-
ual processes with automated support.
Some work on interdomain QoS provisioning
exists in the literature. Key components of an
interdomain QoS architecture have been
described in [2]; at the service management
level, [3] proposed SLA policies to enable INPs
to agree how to distribute QoS across multiple
domains. At the control level, QoS extensions to
the underlying border gateway protocol (BGP)
Michael P. Howarth, Paris Flegkas, George Pavlou, Ning Wang, and Panos Trimintzios, University of Surrey
David Griffin and Jonas Griem, University College London
Mohamed Boucadair and Pierrick Morand, France Telecom R&D
Abolghasem (Hamid) Asgari, Thales Research & Technology Ltd.
Panos Georgatsos, Algonet S.A.
ABSTRACT
This article presents an architecture for sup-
porting interdomain QoS across the multi-
provider global Internet. While most research to
date has focused on supporting QoS within a
single administrative domain, mature solutions
are not yet available for the provision of QoS
across multiple domains administered by differ-
ent organizations. The architecture described in
this article encompasses the full set of functions
required in the management (service and
resource), control and data planes for the provi-
sion of end-to-end QoS-based IP connectivity
services. We use the concept of QoS classes and
show how these can be cascaded using service
level specifications (SLSs) agreed between BGP
peer domains to construct a defined end-to-end
QoS. We illustrate the architecture by describing
a typical operational scenario.
Provisioning for Interdomain Quality of
Service: the MESCAL Approach
HOWARTH LAYOUT 5/19/05 11:32 AM Page 129
Sign up today - FREE
Mendeley saves you time finding and organizing research. Learn more
- All your research in one place
- Add and import papers easily
- Access it anywhere, anytime


