Nigel Sheridan-Smith PhD

In a competitive environment, Service Providers wish
to deliver services in a lean and agile manner, despite the rising
complexity and heterogeneity within the network. The desire to
support personalised customer experiences and differentiated
services requires that management systems are increasingly
flexible, adaptable and dynamic. Policy-based Management
(PBM) systems can be helpful in reducing complexity and
enhancing flexibility, but they have not typically been… Show more

In a competitive environment, Service Providers wish
to deliver services in a lean and agile manner, despite the rising
complexity and heterogeneity within the network. The desire to
support personalised customer experiences and differentiated
services requires that management systems are increasingly
flexible, adaptable and dynamic. Policy-based Management
(PBM) systems can be helpful in reducing complexity and
enhancing flexibility, but they have not typically been involved in
end-to-end management of the services, leading to only the
partial management of different network functions.
By using a policy-based service definition language, a Service
Provider is able to specify the structure of services and the
manner in which they are to be dynamically managed for each
customer over time in response to different network events.
Combined with a generic PBM system that supports different
management models, policies can be used to describe all of the
functional aspects of a service. As a result, Service Providers have
the potential to deliver new and modified services quickly and
easily, since the definitions can be altered as their needs evolve in
the medium- or long-term. Show less

Our previous work proposed a simple algorithm for the dis-
tribution and coordination of network management policies across a num-
ber of autonomous management nodes by partitioning an Abstract Syn-
tax Tree into different branches and specifying coordination points based
on data and control flow dependencies. We now extend this work to sup-
port more complex policies containing control flow logic and looping,
which are part of the PRONTO policy language. Early simulation… Show more

Our previous work proposed a simple algorithm for the dis-
tribution and coordination of network management policies across a num-
ber of autonomous management nodes by partitioning an Abstract Syn-
tax Tree into different branches and specifying coordination points based
on data and control flow dependencies. We now extend this work to sup-
port more complex policies containing control flow logic and looping,
which are part of the PRONTO policy language. Early simulation results
demonstrate the potential performance and scalability characteristics of
this framework. Show less

The distribution and coordination of policies is often over-
looked but is crucial to the scalability of dynamic, personalised services.
In this work we partition an Abstract Syntax Tree of the policies to deter-
mine the responsibility of di®erent management nodes in a geographically
segregated network (i.e. management by delegation). This partitioning is
combined with IN/OUT set analysis to determine the required coordina-
tion for policy enforcement of complex policies with… Show more

The distribution and coordination of policies is often over-
looked but is crucial to the scalability of dynamic, personalised services.
In this work we partition an Abstract Syntax Tree of the policies to deter-
mine the responsibility of di®erent management nodes in a geographically
segregated network (i.e. management by delegation). This partitioning is
combined with IN/OUT set analysis to determine the required coordina-
tion for policy enforcement of complex policies with inter-dependencies.
Our simulation results show that this approach is promising, as higher
decision loads can be readily handled by further sub-dividing of the net-
work. Show less

Many existing management systems are not evolutive
or adaptive, leading to multiplicity over time and
increasing the management burden. Policy-based
management approaches may assist in making networks
less complex and more automated, but to date they have
not yet been able to evolve to support new service sets or
provide the capacity for differentiation.
We present the architecture for a policy-based system
named Pronto that helps to deal with these issues.
Layered… Show more

Many existing management systems are not evolutive
or adaptive, leading to multiplicity over time and
increasing the management burden. Policy-based
management approaches may assist in making networks
less complex and more automated, but to date they have
not yet been able to evolve to support new service sets or
provide the capacity for differentiation.
We present the architecture for a policy-based system
named Pronto that helps to deal with these issues.
Layered network and service models are built above an
extensible virtual device model that supports
heterogenous management interfaces. Interchangeable
management components provide the basic building
blocks to construct logical services. The integrated
policy-driven service definition language automates the
management of the services in a manner that is adaptive,
dynamic and reactive to improve the user’s overall
service experience. Show less

Current management systems make it difficult and
often prohibitively expensive for Service Providers to offer
value-added services to large numbers of subscribers with a
concern for each individual’s unique experience. Furthermore,
many Service Providers do not clearly understand what services
will attract subscribers, in order to maximise long-term
profitability and growth.
This paper discusses the concept of Quality of Experience as
a means to come to a greater… Show more

Current management systems make it difficult and
often prohibitively expensive for Service Providers to offer
value-added services to large numbers of subscribers with a
concern for each individual’s unique experience. Furthermore,
many Service Providers do not clearly understand what services
will attract subscribers, in order to maximise long-term
profitability and growth.
This paper discusses the concept of Quality of Experience as
a means to come to a greater understanding of emotive user
behaviour and each individual’s purpose and needs. We combine
this understanding with the use of Policy-based Network
Management to build more automated, adaptable and evolvable
management systems that can respond dynamically to needs.
However, there is still a great deal that must be done to architect
management systems that can handle millions of users. This is
because high performance and availability are essential but
service complexity and user demands will continue to increase,
putting these non-functional qualities at risk. Show less

Experimental design methods have long been used
in scientific areas such as agriculture, biology and physics to
minimise error and assure validity. Although most network
researchers performing experiments with testbeds and simulations
implicitly follow scientific method, until recently there has
been little emphasis on improving experimental design methods.
Traditional experimental design focuses on experiments where
the scope and objectives are relatively constrained… Show more

Experimental design methods have long been used
in scientific areas such as agriculture, biology and physics to
minimise error and assure validity. Although most network
researchers performing experiments with testbeds and simulations
implicitly follow scientific method, until recently there has
been little emphasis on improving experimental design methods.
Traditional experimental design focuses on experiments where
the scope and objectives are relatively constrained, however
network research in innovative areas where there is little or no
precedence often has changing objectives that evolve over time.
We describe the learning-outcome approach to the experimental
design life-cycle that applies the concepts of systems development
life-cycle models used in software engineering, as well as learning
taxonomies used in education. Our approach extends traditional
experimental design by providing a more comprehensive and
efficient way of decomposing an experimental research project
into manageable stages that are designed rather than improvised,
leading to a well-structured way of assessing the knowledge
gained. We provide an insight into our experiences with this
approach in the context of experimental research in management
of Alcatel Australia’s new-generation Enriched Experience
Network. Show less