QoS-Oriented Service Computing - Advanced Design Approaches to Emerging Software Systems - page 281

employed for SLA presentation. Especially, the
WS-Agreement is widely used in many Grid com-
puting projects. WS-Agreement not only provides
a specification language for representing SLA,
but also provides protocol for SLA negotiation
and monitoring.
2. While there are many research projects which
involve employing SLA for service and resource
management, a comprehensive reference archi-
tecture which identifies key SLA components and
illustrates how these components are engineered
and interacted to provide a generic overview of
SLA application / service has not been investigated
in great depth. GRIA SLA Management Service
(http://www.gria.org/) is usable, but the high level
SLA architecture is still not quite clear. Web Ser-
vice Level Agreement (WSLA) provides an SLA
framework. It defines how basic metrics should
be measured and how they are aggregated into
composite metrics and SLA parameters. It also
provides expression of the operations for monitor-
ing and managing the service. However, WSLA
was developed to provide an SLA specification
language rather than SLA life cycle management.
Although SLA has been used in many research
projects, these projects are mainly concerned with
one or two functional aspects of SLA (i.e. resource
management and scheduling, negotiation) and do
not provide a high-level view of the SLA manager.
3. It is recommended to incorporate domain
ontology into the SLA to present semantic-rich
policies and requirements. Semantics-enriched
policies can facilitate more accurate SLA life
cycle management than the syntactic approaches.
Therefore, apart from the key SLA manager
requirements identified in the survey, another
requirement of the SLA manager is that it should
be 'semantic-aware'.

computing environment. Cloud computing can
be illustrated from the following aspects:
SPI model - Cloud computing originates from
the concept “Hardware as a Service” (HaaS),
“Software as a Service” (SaaS). Cloud now ad-
vances from SaaS to “Platform as a Service” (PaaS)
and “Infrastructure as a Service” (IaaS), known as
SPI model. In Cloud computing, customers can
avoid capital expenditure on hardware and soft-
ware by renting the usage from service provider
of third parties, rather than owning the physical
infrastructure by themselves. The hardware and
software are rendered to customers as IT services.
Scalability / elasticity - Klems and Gaw claim
that automatic scale of infrastructure for load
balancing is a key element in Cloud comput-
ing (Geelan, 2008). The delivered services can
elastically / dynamically grow its capacity on an
as-needed basis so that the quality of service can
be guaranteed. “On-demand services are all Cloud
computing based” (de HAAf, 2008).
“Pay-per-use” / “Pay-as-you-go” / “Utility
computing” - There is also a vision that Cloud
computing is more like a business revolution,
rather than a technology evolution. Business
model, or we call “pay-per-use”, “pay-as-you-
go”, and 'utility computing' is another feature of
Cloud computing (Kaplan and Cohen in Geelan,
2008; Watson et al., 2008; Buyya et al., 2009). The
usage of the resource will be metered and service
customers will pay bill to service provider for the
actual resource usage.
Data centre - Another view of Cloud is that it
is a powerful computer and the data centre is the
basic unit of the infrastructure (McFedries, 2008).
Data centre can offer huge amount of computing
power and data storage. The capacity of the data
centre can dynamically when handling a task.
According to (Vaquero et al., 2009), this is associ-
ated with the concept “massive data scalability”
proposed by (Hand, 2007).
Virtualisation - Cloud computing can also be
regarded as a “virtualised hardware and software”
(Gourlay and Sheynkman in Geelan, 2008). This

Cloud Computing

Cloud computing has become a new computing
paradigm as it can offer a scalable IT infrastruc-
ture, QoS-assured services and customizable