Username
Password
Remember Me
This little quiz is intended to give you a flavor of the breadth and depth of questions that can be answered by scientific, cross-platform techniques.
Select the most appropriate answer. A batch program processes 10,000 records/hr. When run in isolation the average utilization is 33%. Business forecasts that load will go up 3 times. The admin thinks a hardware upgrade will help, but is not sure by how much. Your thoughts? No hardware upgrade is required as the existing hardware will scale to support the increase in workload with no other changes. Remember batch program can take up 100% utilization Hardware upgrade may be required as the hardware utilization will be close to saturation with the increase in workload, and this is not desirable No question of hardware upgrade, the application has to be examined to fix existing bottlenecks Upgrade capacity by a factor of 1.23 times, taking the cut-off utilization as 80% An architect is asked to upgrade a server that currently handles 100 requests per second, and is expecting 100% growth. What should her deployment strategy for this Intranet system be to ensure the average response times do not exceed 0.5s? Propose another server box with the same capacity and place them in parallel. This will ensure the throughput handled by them jointly is 200 requests/second. The response time target will be met Propose a server box that has double the capacity of the existing server box ( use benchmark data to get this). This will ensure the throughput handled by the box is 200 requests/second, if the application scales well. The response time target will be met It is impossible to guarantee response times in this scenario An application is performance tested on two different boxes - L (smaller) and H (bigger). On both of these, the testing is carried out with exactly the same number of virtual users and think times. What do you expect to see? Throughput with H will be higher than with L. Since utilization increases with increase in throughput, utilization with H will also be higher The throughputs with both L and H will be the same. However utilization with H will come down by a factor that is equal to the capacity factor between H and L The throughput with H will be higher, if the application is scalable. The increase in throughput with H as compared to L will depend on the think time value. The greater the think time, the smaller will be this increase An intranet app is hosted on box L (smaller) and is observed to support 99 requests/s with an average response time of 1s. What average response time can you expect if this app is hosted on box H for the same workload if H is twice the capacity of L? About 0.01 seconds About 0.05 seconds About 0.09 seconds About 0.5 seconds Your site supports 8 million registered users and 1 million logged-in users, who generate 22 hits per second. The target is to achieve sub-second response times. Each vUser license costs $30. How much do you need to spend on vUser licenses for the test? $330 $660 $30 million $240 million An application is load-tested with different combinations of virtual users and think times. The test are such that the same steady-state throughput is achieved each time. What would be the average response time in each case? Response time will be higher for the case where the number of users is higher, as the overall queue lengths will be greater Response time will be higher for case where think time is lower, as contentions will be higher Each case will have the same response time Impossible to estimate with this minimal input
