Database Reference
In-Depth Information
services on EC2. The Elastic Load Balancer service automatically provisions
incoming application workload across available EC2 instances while the Auto
Scaling service can be used to dynamically scale-in or scale-out the number of EC2
instances for handling changes in service demand patterns. Finally the CloudWatch
service can be integrated with the above services for strategic decision making based
on collected real-time information.
Eucalyptus is an open source cloud computing platform. It is composed of
three controllers. Among the controllers, the
cluster controller
is a key component
that supports application service provisioning and load balancing. Each cluster
controller is hosted on the
head node
of a cluster to interconnect the outer public
networks and inner private networks together. By monitoring the state information
of instances in the pool of server controllers, the cluster controller can select any
available service/server for provisioning incoming requests. However, as compared
to Amazon services, Eucalyptus still lacks some of the critical functionalities, such
as auto scaling for its built-in provisioner.
Fundamentally, Microsoft Windows Azure
fabric
has a weave-like structure,
which is composed of node including servers and load balancers, and edges
including power and Ethernet. The
fabric controller
manages a
service node
through
a built-in service, named Azure Fabric Controller Agent, running in the background,
tracking the state of the server, and reporting these metrics to the controller. If a
fault state is reported, the controller can manage a reboot of the server or a
migration of services from the current server to other healthy servers. Moreover,
the controller also supports service provisioning by matching the VMs that meet
required demands.
GoGrid Cloud Hosting offers developers the F5 Load Balancer [
18
] for distribut-
ing application service traffic across servers, as long as IPs and specific ports of
these servers are attached. The load balancer provides the round robin algorithm
and least connect algorithm for routing application service requests. Additionally,
the load balancer is able to detect the occurrence of a server crash, redirecting further
requests to other available servers. But currently, GoGrid only gives developers a
programmatic set of APIs to implement their custom auto-scaling service.
Unlike other cloud platforms, Google App Engine offers developers a scalable
platform in which applications can run, rather than providing direct access to a
customized virtual machine. Therefore, access to the underlying operating system
is restricted in App Engine where load-balancing strategies, service provisioning,
and auto scaling are all automatically managed by the system behind the scenes
where the implementation is largely unknown. Chohan et al. [
105
] have presented
initial efforts of building App Engine-like framework,
AppScale
, on top of Amazon
EC2 and Eucalyptus. Their offering consists of multiple components that auto-
mate deployment, management, scaling, and fault tolerance of an App Engine
application. In their design and implementation, a single
AppLoadBalancer
exists
in AppScale for distributing initial requests of users to the
AppServer
sofApp
Engine applications. The users initially contact AppLoaderBalancer to request a
login to an App Engine application. The AppLoadBalander then authenticates the
login and redirects request to a randomly selected AppServer. Once the request is
