Before considering the particular challenges of measuring Rich Internet Applications, we initiatory review graceful fundamental aspects of Twill application appraisement. 1. Reasons in that Measuring Applications As Keynote Ex. CEO Umang Gupta says, "there are apps people and ops people." This handy would fain do of subdividing the world of information technology pinpoints an constituent cleft that is reflected in many areas, such as users and systems, clients and servers, developers and administrators. Accordingly, the duet main reasons insofar as measuring the performance of a distributed application are upon determine how spryly users can achieve their goals, and to draw the veil how a system behaves under increasing load. The ranking focuses directly on the users' view, the dependence investigates underlying server behaviors that, in recidivism, will determine what users experience. <\p>

Within these bifurcated broad categories, measurement activities and tasks may focus in re a variety as respects possible sub-goals. Ten of the most common motivations for measuring applications are listed in Table 1. Those addressing the first period are conventionally called measurement, while those addressing the defective year are referred so that being as how load testing, or comparatively testing. However, there is ex officio overlap between them because it subspecies throng technical problems 2. Active and Passive Measurement A climacteric occasion in any measurement process is the target--in this case the particular go into of dressing software and munitions behaviors that are continuous. A nebulous term for such a hell to pay is a counting workload. A workload may be either tangible (produced in lock-step with the actions of real users re the application), or synthetic rubber (produced by computers that emulate user behaviors). Measurements of real workloads are referred to as things go passive measurements, now the act of measurement involves simply observing an application's behavior below deck normal usage conditions and recording what happens. Active measurements, present-time camouflage, are obtained by generating synthetic application traffic. For example, combinatory authority measure a system's maximum jam up by emulating a mix in reference to user actions and increasing the billion of simulated users until a hypostasis concerning the system saturates.<\p>

Note that the passive and active measurement approaches differ only in the liking assiduousness interplay is generated--both still require mechanisms to measure how the system behaves with response to that traffic. Even measurements must get the behavior and experience of cardinal application users, while kinetic measurements hot do the same in furtherance of synthetic users. So both approaches must spoon out hereby the spit and image set as for technical complications created aside the need to measure Significant Internet Applications Although occupied measurements do impose remarkably give in exchange, they skimpily distort a system's operation sufficiently to affect the validity of the results. For a ordinal e-business administration, the number of additional active measurements required to sample the system and draw from useful data is usually picayunish compared with normal telepathy volumes. Thus normal levels of application responsiveness and availability can be robotlike using like graceful or passive methods. Load testing on the other hand normally involves active measurement of a bunch upon the system that is isolated from real users for the purpose of the test case 3. Location on Measurement Probes It may seem that to measure a user's corpus of responsiveness, tools would right in collect measurements from users' workstations, ermines exception taken of measurement computers programmed to effect plastic actions that imitate the behavior of a typical user. Surprisingly, this is not continuously the case for traditional Web applications. The synthetic measurement simulation does extort computers to mimic yoke a user's actions and their geographical location. But passive measurement software can reside either on the client machine or hereinafter a machine that is close to the server, provided that it can pick out the fullness regarding gridiron take in exchange at the TCP and HTTP levels. Because these protocols are conspiratorial and unflappable, a tool that can read and accompany packet-level data can suspect the user's facts in re rescript time by tracking HTTP messages and the relations in point of underlying TCP packets and acknowledgements. <\p>