Information Technology Reference
In-Depth Information
Moreover, all available monitoring components are notified about events, allow-
ing them to only store a partial set of the events enabling distributed storage.
Two
workflow clients
have been implemented which use the access compo-
nent to interact with the system. The first is a standard workflow management
client with a user interface enabling a workflow participant to process work
items, monitor the system and perform administrative tasks, provided they pos-
sess the necessary privileges. The second client is an automated client which is
used to test process models by repeated execution of test scenarios [4]. A third,
web-based client is currently being developed. All of the clients use the service
discovery of the infrastructure to find available workflow management systems
and allow the user to choose any system available within the enterprise cloud.
As shown, both the workflow management system and the available client soft-
ware make extensive use of the functionality of the infrastructure. This enables
them to distribute the workload across multiple nodes, dynamically replace parts
of the system and access remote functionality. Since the infrastructure manages
the technical details, the implementation can be relatively simple, requiring little
effort to enable a complex application to be distributed.
5 Related Work
The main rationale of the assessment of related approaches, stemming from grid
and cloud PAAS, consists in evaluating them with respect to the five challenges
from Section 1. In Fig. 7 an overview of the analyzed approaches is given.
The original intent of grid computing approaches is exploiting computing
power of other nodes e.g. for high performance computing. This motivation has
led to client/server approaches with load being distributed from the server to
clients. They perform the assigned tasks and send back the results to the server.
Fig. 7.
Related approaches from cloud and grid computing
