System Architecture; Resources Running The Platespin Orchestrate Agent; Jobs; Typical Use Of The Grid - Novell PLATESPIN ORCHESTRATE 2.0.2 - GETTING STARTED REFERENCE 06-17-2009 Getting Started

2.1.3 Resources Running the PlateSpin Orchestrate Agent

For PlateSpin Orchestrate, a resource is a computing node somewhere in your network (that is, your
grid or data center network), which has a PlateSpin Orchestrate Agent installed on it. When the
Orchestrate Agent is installed on a resource, communication between the agent and the PlateSpin
Orchestrate Server is established and the computing resource can be discovered and can begin
performing jobs that are assigned to it by the Orchestrate Server.

2.1.4 Jobs

For PlateSpin Orchestrate, a job is remotely executable logic that can run on some or all of the
resources in the Orchestrate grid. Job logic is written in Python, and can include instructions and
policies that dictate how, when, and where that job runs. In addition, the job can embed instructions
that dictate processes or applications that the resource needs to launch.
A Job Developer uses Python in a prepackaged Job Definition Language (JDL) editor to create or
modify a job. When the job is ready, the administrator of the PlateSpin Orchestrate system uses the
Orchestrate Server functionality to deploy it and to allocate appropriate resources to run it, based on
its parameters and associated policies.

2.1.5 Typical Use of the Grid

In general, the everyday use of the grid and its components follows this sequence:
1. Jobs are created using JDL. A job might also optionally reference one or more policies (defined
in XML).
2. The Orchestrate Server is started. It discovers all the available computing resources.
3. The administrator logs in to the Orchestrate Server and deploys jobs that users can run. The
administrator might also create user logins.
4. Users log in to the Orchestrate Server run jobs.
5. When a user selects a job to run, he or she runs the job based on selected options; for example,
he or she might select when to run the job, how many computing resources the job should run
on, the type of computing resources to be used, and so on.
6. Users monitor and control their own jobs by using certain functions, such as canceling a job,
pausing a job, or even changing the priority of a job if they have rights to do so.
7. Administrators can monitor and control all of the running jobs. They can dynamically change
how a job runs, they can change job priorities, and so on.
8. Steps 5, 6, and 7 can be repeated as long as the Orchestrate Server is running.

2.2 System Architecture

PlateSpin Orchestrate is built as an enterprise server with capabilities to support extremely large
computing environments. It includes all the services required to provide an enhanced grid
computing application server environment.
A simple view of the PlateSpin Orchestrate system architecture is shown below, showing the
PlateSpin Orchestrate software components:
16 PlateSpin Orchestrate 2.0 Getting Started Reference