Interaction of Operations and Application Development Teams


One of the obstacles to effective applications management is not a lack of technology; rather, it is often a structural problem within IT organizations. The network operations staff, application managers, and developers are often unaware of the impacts of their decisions on the overall service quality. Although most groups have experts with deep technical knowledge of their own area, barriers exist that inhibit communications with other experts who can make service delivery more effective. This isolation also results in situations where a decision made by one group negatively impacts overall service because of an unanticipated interaction with technologies that are managed by a different group.

The Effect of Organizational Structures

One of the most difficult issues with application management is the gap in knowledge and perspectives between application developers and the operations teams. Of course, both have their specialized functions, but often application performance problems hinge on poor application design or operational choices that hobble performance.

Organizational structures often accentuate this gap. Many IT groups are divided into teams by technical specialty, with independent network, systems, and applications teams; at other times, there is a division between the operations and development groups. I'd characterize many of these relationships as ranging from aloof to adversarial. The finger pointing that typically occurs with a service quality failure is an unpleasant, but, unfortunately, all too common occurrence within many IT groups. This frequently leads to poor decision-making on either side with impacts that affect overall service quality and economics.

Sometimes, lack of a common vocabulary keeps communications from being clear. At other times, the network and applications people don't understand basic assumptions about the other's work and perspectives. Often what's required is collecting and organizing objective information and design guidelines that both sides can use. It is easier to resolve and prevent performance problems when both sides have information they share and trust, and when both sides have learned to collaborate on design issues.

The Need to Understand the Operational Environment

I was recently speaking with some developers visiting a large e-business software firm. The developers were quite excited about incorporating more object request broker technology into their new applications. They cited faster implementation times and simpler programming because the object request broker handles all kinds of messy details. An application simply requests an object and the object request broker takes over the task of locating it, accessing the object, and performing any necessary data and formatting transformations.

Granted, this does simplify the application development process, but there is the impact of not knowing where your objects (content) are actually located. Content location can have a significant performance impact due to long distance (propagation delay), restricted bandwidth, or overloaded servers at a given location. In response, the teams felt that any unacceptable access delays could be fixed by adding more bandwidth. However, this answer masks confusion between bandwidth and propagation delay. In this instance, the stated solution would undoubtedly contribute to poor application design decisions that cannot always be fixed later by throwing resources (money) at the problem. This is especially true when wide-area networks (WANs) spanning long distances are involved.

To enhance the understanding of network impact on application behavior, tools such as Compuware's Application Expert can be useful. Application Expert is used during the development phase to quickly test different application scenarios. It enables developers to see the effect of network delays and bandwidth on application performance. This information is fresh enough that it has an impact during the development process rather than after the fact; this leads to better implementation. In the deployment phase, Application Expert can be used to monitor the actual performance and identify further opportunities for improvements.

Time Lines Are Shorter

Another consideration is that the market is unforgiving. In the past, applications could be introduced and then tuned and shaken down over a period of time. Today, applications must deliver peak performance and functionality immediately. Many customers won't give you a second chance. If the application is sluggish, if links are broken, if there are bugs, you've lost your chance to attract or keep those customers. It's extremely easy for your customer to shop a competitor's web sitepossibly while waiting for your slow site to respond. If your customer, having been inspired to shop around by your site's poor performance, then discovers that your competitor has better prices, or better products, or even the same prices and products in an easier-to-use, more reliable web siteyou're in trouble!




Practical Service Level Management. Delivering High-Quality Web-Based Services
Practical Service Level Management: Delivering High-Quality Web-Based Services
ISBN: 158705079X
EAN: 2147483647
Year: 2003
Pages: 128

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net