Wayne Fenton, eBay, Inc.

The structure of applications is getting more complex in many regards. SOA, EDA, mash-ups, Ajax/RIA, all imply more dynamic use of loosely coupled sub-systems in a potentially massive scale.

Architectures that focus on the "ilities" (scalability, reliability, etc.), too often ignore or give short shrift to the requirements imposed by the operations staff and processes that govern application management in production environments. Grid computing is an attempt to standardize components and interfaces to allow tools and systems to manage complex multi-tier applications and associated resource pools in a consistent, scalable way. However, application architects and designers need to shift their thinking so that systems will function well in that environment.

This talk will cover the ways in which we as software architects can design systems for much-improved efficiency and reliability from an operational perspective. This includes topics such as configuration discovery and management, server/application provisioning, application monitoring and alerting, HW/power optimization, and disaster planning.