When somebody walks into a dark room and flips a switch, the room turns bright instantaneously. That is what users expect when they click or touch their business or personal application (App) tabs – instant responses. We may never quite get there for all user applications – but Google’s pre-rendering during web search is almost there. Application Performance Management (APM) techniques are what we need to help corporations approach such lofty goals.
APM plays an important role when corporations consolidate their data centers. Virtualization made centralizing the data centers, into one or a handful, even more attractive from cost savings and operations point of view. With Cloud computing in the horizon, corporations can benefit from a variety of choices from public to private Cloud and Infrastructure-as-a service to Application-as-a service.
Corporations following best practices tend to incorporate APM issues right in the planning stages of data center or Cloud migration. A tiny fraction of the millions of dollars spent on migration itself provides good assurance that the end-users (employees and customers) get fast response times for their application transactions no matter where they are physically located – users want interactions at the speed of lightening not the delayed response of a thunder.
The end-users that are most affected are the ones that are in the same facility as the old data center. Because of very high bandwidth and almost-zero latency of the LANs they are used to zippy response times for years. In contrast, the latencies across a WAN or the Internet could be in the range of tens of milliseconds to hundreds of milliseconds; there could be choke points along the path with low bandwidth.
Now with the Cloud migration and the intervening WAN or Internet, they are likely to get stuck with slow responding applications resulting in loss of productivity and customer dissatisfaction – unless their company is pre-planned in managing the end-user expectations as an integral part of the migration project.
The main steps involved in addressing the APM issues during a Cloud migration are – identifying and bench-marking critical applications, predicting how they would behave when the migration is completed, and taking remediating actions for those applications that are predicted to have unacceptable response times.
The remediation steps typically involve providing additional bandwidth at the affected locations, using WAN Optimization, incorporating Content Delivery Network (CDN) solutions, converting the application into a thin-client one such as Citrix, or modifying the application to reduce chattiness. In rare cases, such as a mission-critical legacy application, a corporation may decide not to move the application into the Cloud.
An APM project during a Cloud migration not only requires sophisticated tools from tool vendors such as OPNET but also consultants who can wield these tools with minimal disruption to the current business operations.
With the right APM team as a part of the broader Cloud migration endeavor, performance issues can be addressed head-on. Such forethought can ensure that Cloud migration is a happy one for the end-users thus preserving productivity and customer satisfaction. Expensive fixes after migration are avoided and ROI on Cloud migration in-tact.
APM best practices have been in place for a while and benefitted situations such as a new application roll out – this was well before IT became Cloudy!