Database Reference
In-Depth Information
usually done in an ad-hoc manner due to the complexity and dynamism inherent
in the interaction between the different tiers and components of the system. In
particular, meeting SLAs which are agreed with end-users by cloud customers'
applications using the traditional techniques for resource provisioning is a very
challenging task due to many reasons such as:
Highly dynamic workload
: An application service can be used by large numbers
of end-users and highly variable load spikes in demand can occur depending on
the day and the time of year, and the popularity of the application. In addition,
the characteristic of workload could vary significantly from one application
type to another and possible fluctuations on the workload characteristics which
could be of several orders of magnitude on the same business day may occur
[
83
]. Therefore, predicting the workload behavior and consequently devising
an accurate plan to manage of the computing resource requirements are very
challenging tasks.
Performance variability of cloud resources
: Several studies have reported that the
variation of the performance of cloud computing resources is high [
112
,
172
,
208
].
As a result, currently, cloud service providers do not provide adequate SLAs
for their service offerings. Particularly, most providers guarantee only the
availability, rather than the performance, of their services [
68
,
124
].
Uncertain behavior
: One complexity that arises with the virtualization tech-
nology is that it becomes harder to provide performance guarantees and to
reason about a particular application's performance because the performance of
an application hosted on a virtual machine becomes a function of applications
running in other virtual machines hosted on the same physical machine. In
addition, it may be challenging to harness the full performance of the underlying
hardware, given the additional layers of indirection in virtualized resource
management [
199
].
Several approaches have been proposed for dynamic provisioning of computing
resources based on their effective utilization [
115
,
190
,
232
]. These approaches are
mainly geared towards the perspective of cloud providers. Wood et al. [
232
]have
presented an approach for dynamic provisioning of virtual machines. It defines
a unique metric based on the data consumption of the three physical computing
resources, including CPU, network, and memory to make the provisioning decision.
Padala et al. [
190
] carried out black-box profiling of the applications and built
an approximated model which relates performance attributes such as the response
time to the fraction of processor allocated to the virtual machine on which
the application is running.
Dolly
[
96
] is a virtual machine cloning technique to
spawn database replicas and provisioning shared-nothing replicated databases in
the cloud. The technique proposes database provisioning cost models to adapt
the provisioning policy to the low-level cloud resources according to application
requirements. Rogers et al. [
200
] proposed two approaches for managing the
resource provisioning challenge for cloud databases. The black-box provisioning
uses end-to-end performance results of sample query executions, whereas white-
box provisioning uses a finer grained approach that relies on the DBMS optimizer
