Usage Scenarios
From JPPFWiki
basboletoli
| Main Page > Usage Scenarios |
These usage scenarios describes where does JPPF can be applied (or planned to be)
Planning
Scenarios dedicated to planning activity in organizations
Building scenarios for job scheduling in job or flow shop
the user want to build many scenarios to decision support. Each scenario is created using a different algorithm that minimize one issue, like makespan, lateness, and so on. The data used by the user is: machine description, jobs description, and some helper data, so there is not too much data involved in this scenarios
Demand forecast and modelling
To better understand the market where the company is competing, marketing staff need to cluster the customer, create buying profiles, forecast the demand for short and middle range time ahead. So the user need to run a bunch of algorithms over a huge amount of data of sales and customer details. Some algorithms can run in parallel others no.
Supply Chain Planning
The user need to create many scenarios of material production, based on available capacity, demand forecasted, suppliers plan (and capacity), cash and logistics contraints. So the user need use some data from many different sources, and probably systems in different domains.
Software in Value Chain
Scenarios dedicated to software in value chain of organization
Billing in telco industry
Call processing requires the manipulation of huge ascii files, the processing of each file (in average) can be done in parallel, and often requires much computation processing to rate it, aggregate it. The non disclosure of call details is required by law, so this is a very sensitive data.
Bill processing normaly requires the integration with tax calculation and aggregation.
Fraud detection
Users must train neural networks with huge amount of data, and requires the evaluation of each event in very tiny time.
Scientific Computing
Applications that require millions of processing nodes for huge scientific computations
Wheather Prediction
Need we say any more? Wheater prediction is one of the largest most complex scientific challenges. It requires the combination of millions of parameters with predictive heuristics based on no less huge historical data.
Protein Folding
Used in genetic research to better undertand how our genotype and phenotype are tied, so as to track down and cure many genetic diseases.
Astronomical Data Analysis
The sheer amount of data to process here is daunting. Among potential applications: signal enhancement, pattern searching, EM spectrum analysis, etc...
