When deploying the Moab family of products or Maui Cluster Scheduler, proper requirements assessment, design and deployment can mean the difference between a fully optimized and easily managed high performance center or an unorganized and burdensome project that never fully materializes.

For years, Cluster Resources has successfully worked with hundreds of customers with a wide variety of needs and backgrounds, and as a result have developed an in-depth knowledge and experience of how the full potential of their resources can be unleashed. One constant in these relationships has been an over-arching drive to understand our customer’s true needs; it is this drive to understand their needs that has led to achieving the full potential of our customers’ assets.

Whether your needs consist of any of the following or others not listed, we have the experience and knowledge that will aid you in successfully meeting your business objectives:

• Requirements Assessment
• Design
• Emulation
• Installation
• Configuration
• Algorithm Specification and Development
• Optimization
• Maintenance

Further, to assist you in acquiring the expertise needed to manage your own resources or provide services for others, we provide the following classes:

Basic Scheduling for Scientific Computing

• Purposes of batch systems
• Tools to promote a single system image
• Tools for traffic control
• Enforcing site mission policies
• Basic resource management and optimizations
• Installing and configuring Maui
• Informational commands
• Control commands
• Commands for end users. (For Administrators)

How to Optimize - Tips and Techniques for Maximizing System Performance

• Sources of resource waste
• 'Behind the Scenes' Scheduling
• How to design policies
• How to tune systems for performance
• Pros and cons of various optimization techniques
• Evaluating statistics and profiling usage
• Using simulation to predict the performance impact of changes in scheduling environment
• Enabling user feedback through scheduler commands and job analysis. (For Administrators)

Trouble Shooting in the Realm of Resource Management

• Sources and types of failures on batch systems
• Pursuing High availability on clusters
• Using built-in diagnostic tools to analyze behavior and failures
• Log management to isolate and track issues
• Reproducing failures with simulation
• Minimizing risk through simulation and test mode evaluation
• Avoiding common configuration mistakes. (For Administrators)

Advanced Scheduling Configuration and Tuning

• Determining who the 'customers' are and what they really want
• Translating legacy policies into correct scheduling constraints
• Imposing fairness through throttling policies
• Managing cycle distribution with fairshare
• Controlling service levels with the QoS feature
• Implementing flexible policies with advance reservations. (For Administrators)

Grid Scheduling with Silver

• Benefits of grid scheduling
• False assumptions from the local cluster
• High level design of a batch grid environment
• Handling credentials in a multi-administrative domain environment
• Managing data in grids
• Tracking allocations and enforcing grid-wide policies. (For Administrators)

Getting Along in the World of Batch

• How to create batch jobs
• Steps to get the most out of available resources
• How to 'train' your admin
• Determining Resource Usage
• Preventing over-specification
• Tuning your jobs to minimize response time. (For End Users)

Each class provides between 4 and 5 hours of training and includes a lab where appropriate. Consulting contracts are set up on a custom project basis. Services can either be provided on-site or remotely. Please contact our sales department for details.