Principles, Methodology and Tools for Engineering Cloud Computing Systems - Advanced Design Approaches to Emerging Software Systems - page 266

Figure 6. Sample SLA management system for Cloud service provision

2.3 Service Provisioning

to avoid a possible service outage and greatly
decreasing the offered Quality of Service (QoS);
upon requests shrinks the allocated VMs would
then need to be released to avoid paying for un-
needed elements.
However, available IaaS interfaces are usually
too close to the infrastructure, forcing the SP to
manage manually the VMs assigned to support
the service. See for instance, the typically WSDL-
or REST-based: Amazon's API, GoGrid's API,
Sun's Cloud API or VMware's vCloud (Varia
2009, GoGrid 2010, Sun Cloud API, vCloud API
2010). The available APIs are way too low level
for the abstraction required by service providers.

After developing and defining the service, the
next stage will lead Cloud users to focus on the
way service provisioning is done in current IaaS/
PaaS/SaaS clouds and its APIs (relying on work
done in the previous section).
IaaS Clouds allow for service providers to
quickly arrange new computing infrastructure in
a pay-as-you-go way. This way, virtual hardware
resources can be dynamically provisioned accord-
ing to their services' load. For instance, when the
number of incoming requests grows, new Virtual
Machines (VMs) need to be manually allocated