The ability of the system to change the size in terms of
storage, RAM according to the need of the user that is the dynamic allocation
of resources and data. Using the resources without any memory wastage constitute
the full use of resources. For eg, 100 mb of storage is allocated for two
processes, if memory is fixed as 50mb for process A and 50mb for process B
where the size is fixed but process A requires 60mb during run time and process
consumed only 40mb, 10mb of memory is unused. In this scenario, the resources
are not used wisely and hence dynamic allocation(scalability) is required to
use all the resources equally.
All the resources available in the
cluster are centralized and are controlled by a single node. The computer
cluster appears like a single system to the user. It includes single user
interface, process and I/O space and entry point.
The internal and external resources
needed by the computer cluster should be available whenever want by the user.
For example, If process A needs an I/O devices like printer to complete the job
but that resource is being used by another process B, A needs to wait until B
is over ie, A will be in waiting stage and the I/O device will be in blocked
stage which results in a deadlock state. Hence, scheduling must be done to
manage the availablity of the resources using necessary algorithms like first
in first out (FIFO), round robin, last in first out (LIFO), least recently used
or any suitable algorithms.
of cluster jobs include providing the scalability and availablity to the user.
Allocation of resources ,executing parallel jobs or serial jobs when required.
For example, if two jobs in the cluster required same resource, the job can be
run parallely at the same time instead of serializing the job. Likewise based
on the situation, the jobs are managed accordingly in the cluster based on the
requirement of the resources.
communication is the communication between the nodes in th clustering
environment . This is required when a node wants to interact with another node
in order to share inputs from node to node or to pass some paramater values to
another node present in the cluster. This is also needed to execute when one job is completed, the succeeding jobs
are signaled or acknowledged to start its execution. Only through the internode
communication, availability can take place orderly.
the system results in error or failure, data loss should be prevented and the
system should be recovered. Abnormal state of a device is called as a fault and
so the system should be capable of handling the state. Recovery is the process
of restoring an erroneous stae to error free environment. This can be prevented
by using logs to maintain the state information and recovery points to retain
the data. Necessary system models are designed to prevent the nodes from fault
and provide error free environment.
Computers that are combined together either
tightly or loosely coupled can be viewed as a single system. It is used to
execute a large single job. These type of clusters cannot handle multiple
operations like databases.
Internode communication are established using
computational clusters. It contains small number of nodes to perform. A large
job split into multiple nodes that runs in parallel and return the results to
the main computer.
The highly needed cluster in the network that
is used frequently to identify errors or failures. This cluster is used to
maintain all the other cluster by preventing from failure.
Giving equal chances for each node to execute
its job based on the requirement or priority is the load balancing clusters.
This cluster is to provide high performance and utilization of the resources.