Introduction

During the last decade, many businesses morphed into e-Businesses that offered their services to end-users over the World Wide Web. On top of advanced web infrastructures, enterprises hosted web sites and/or web applications that took their service offerings straight into the homes of customers! Vide these web-based windows, end-users could access any service – be it retail banking, shopping, or ticket booking – from the comfort of their couches, using just a browser and a mouse! Businesses too thrived, as the web significantly widened the reach of the services and increased revenues. The rise in popularity of web-based applications also bought about a paradigm shift in how application performance is monitored and managed.

Most web site/web applications overlay multiple inter-related IT infrastructure tiers – the web server front end, the business logic middle tier, the network tier, the backend storage tier, etc. For long, web site/ web application performance was measured only in terms of the availability and resource usage of the individual infrastructure tiers at work – is the web server available? is the application server sized right? does the database server have enough disk space? In recent times however, the emphasis on ‘service quality’ and the huge penalties attached to ‘service level’ violations, caused the performance management focus to shift from these “servers” to the “users”. But, why “users”? This is because, unlike IT administrators who are solely concerned with ‘server’ performance, users are the ones who provide a ‘service perspective to performance’. Typically, when problems occur in an IT infrastructure, an administrator will only want to know which ‘servers’ are affected and what kind of problems are affecting ‘server’ performance. These administrators are rarely ever concerned with how these ‘server’ problems impact the performance of the web site/web applications they support! End-users on the other hand, view any web site/web application they work with as a ‘service’ offered to them by an enterprise. When end-users run into issues while working with a web site/web application, they instantly tag them and flag them as ‘service’ issues. For instance, if say, a banking web site slows down, the corresponding user complaint will read as follows: the banking service is slow. In many cases, user may not even know that multiple ‘servers’ are in fact working together in the background to deliver the banking service. If IT is able to see what a user sees, it will be able to proactively identify a potential service outage, determine its root-cause, take rapid steps to avert it, and thus save millions by way of penalties! High service uptime also results in many happy users, which naturally translates into a wider customer base, increased revenues, and continued goodwill for the business. On the other hand, if poor user experience with a service goes undetected, user dissatisfaction with a service will persist, causing penalties to rise, reputation to be damaged, and revenues to dwindle. In short, the success/failure of a business today greatly depends on the user experience with its web services. It is hence imperative that IT administrators continuously monitor and efficiently manage user experience with their critical web services and detect the unavailability of the service and probable slowdowns in service delivery well before users do.