Database Reference
In-Depth Information
The remainder of this chapter is structured as follows. Section
7.2
introduces the
architecture of the adaptive framework. Details of the experiment implementation
of the different components of the framework are discussed on Sect.
7.3
. Then,
the results of the experimental evaluation for the performance of the approach are
presented in Sect.
7.4
, followed by discussions and conclusions in Sect.
7.7
.
7.1
SLA Management for Virtualized Database Servers
Cloud-based data management poses several challenges which go beyond traditional
database technologies. In principle, outsourcing the operation of database applica-
tions to a cloud provider who, on the one hand, takes responsibility for providing the
infrastructure and maintaining the system but, on the other hand, this cloud provider
need to pool resources and operate them in a cost-efficient and dynamic way promise
cost savings and elasticity in usage. In practice, most customers of cloud services
will be willing to move their on premise setup to a cloud-hosted environment only
if they can guarantee that their data are kept securely and privately as well as non-
functional properties such as availability or performance are can be maintained.
An SLA is a contract between a service provider and its customers. SLAs capture
the agreed upon guarantees between a service provider and its customer. They define
the characteristics of the provided service including service level objectives (SLOs),
such as maximum response times, minimum throughput rates, and data freshness,
and define penalties if these objectives are not met by the service provider. In
general, SLA management is a common general problem for the different types
of software systems which are hosted in cloud environments for different reasons
such as the unpredictable and bursty workloads from various users in addition to the
performance variability in the underlying cloud resources. In particular, there are
three typical parties in the cloud. To keep a consistent terminology through out the
rest of this chapter, these parties are defined as follows:
Cloud service providers
: They offer the client provisioned and metered com-
puting resources, such as CPU, storage, memory, and network, for rent within
flexible time durations. In particular, they include: infrastructure as a service
providers and platform as a service providers. The platform as a service providers
can be further broken into several subcategories of which database as a service
provider is one of them.
Cloud customers
: They represent the cloud-hosted software applications that
utilize the services of cloud service providers and are financially responsible
for their resource consumptions. Most of software as a service providers can
be categorized into this party.
End-users
: They represent the legitimate users for the services or applications
that are offered by cloud customers.
While cloud service providers charge cloud customers for renting computing
resources to deploy their applications, cloud customers may or may not charge their
