Windows Azure Hybrid Cloud

8 Optimizing for Performance

IN THIS CHAPTER:

• Weighing the cloud application performance fundamentals
• Storage considerations
• Network optimization considerations
• Using affinity groups

In this and the next chapter we're going to focus on concepts that are closely related: performance optimization, and operations and management. One of the fundamental trends we see going forward is that of planning for management and operations from the application's inception. This is imperative. Several new methodologies have emerged to close the gaps in the life cycle of an application — from inception to design and development on one side, to operating and managing (O&M) and finally deprecation on the other. Historically, the disparity between development and post-deployment has been systemic, a result of the way in which the organization is structured and made worse by the culture of the two IT camps: application development and infrastructure teams.

Developers tend to see infrastructure as plumbing; boxes and wires, nothing too complex and nothing to worry about too much. If post-deployment issues arise, the infrastructure people can handle it. The people involved on the infrastructure side, however, tend to view developers as not understanding the implications of proper planning and proper operations from the beginning. The infrastructure teams tend to get caught in a firefighting mode without having a clear understanding of how the application was designed and therefore how to detect and ...