Performance Evaluation Operating Systems Fall 2002 OS Fall

Performance Evaluation Operating Systems Fall 2002 OS Fall’ 02

Performance evaluation § There are several approaches for implementing the same OS functionality Different scheduling algorithms Different memory management schemes § Performance evaluation deals with the question how to compare wellness of different approaches Metrics, methods for evaluating metrics OS Fall’ 02

Performance Metrics § What is the performance metric for the sorting algorithms? § Is something wrong with the following statement: The complexity of my OS is O(n)? § This statement is inherently flawed The reason: OS is a reactive program OS Fall’ 02

Performance metrics § Response time § Throughput § Utilization § Other metrics: Mean Time Between Failures (MTBF) Supportable load OS Fall’ 02

Response time § The time interval between a user’s request and the system response Response time, reaction time, turnaround time, etc. § Small response time is good: For the user: waiting less For the system: free to do other things OS Fall’ 02

Throughput § Number of work units done per time unit Applications being run, files transferred, etc. § High throughput is good For the system: was able to serve many clients For the user: might imply worse service OS Fall’ 02

Utilization § Percentage of time the system is busy servicing clients Important for expensive shared system Less important (if at all) w for single user systems, for real time systems § Utilization and response time are interrelated Very high utilization may negatively affect response time OS Fall’ 02

Performance evaluation methods § Mathematical analysis Based on a rigorous mathematical model § Simulation Simulate the system operation (usually only small parts thereof) § Measurement Implement the system in full and measure its performance directly OS Fall’ 02

Analysis: Pros and Cons +Provides the best insight into the effects of different parameters and their interaction Is it better to configure the system with one fast disk or with two slow disks? +Can be done before the system is built and takes a short time - Rarely accurate Depends on host of simplifying assumptions OS Fall’ 02

Simulation: Pros and Cons 1. Flexibility: full control of Simulation model, parameters, Level of detail w Disk: average seek time vs. acceleration and stabilization of the head 2. Can be done before the system is built 3. Simulation of a full system is infeasible 4. Simulation of the system parts does not take everything into account OS Fall’ 02

Measurements: Pros and Cons +The most convincing - Effects of varying parameter values cannot (if at all) be easily isolated Often confused with random changes in the environment - High cost: Implement the system in full, buy hardware OS Fall’ 02

The bottom line § Simulation is the most widely used technique § Combination of techniques Never trust the results produced by the single method w Validate with another one w E. g. , simulation + analysis, simulation + measurements, etc. OS Fall’ 02

Workload § Workload is the sequence of things to do Sequence of jobs submitted to the system w Arrival time, resources needed File system: Sequence of I/O operations w Number of bytes to access § Workload is the input of the reactive system The system performance depends on the workload OS Fall’ 02

Workload analysis § Workload modeling Use past measurements to create a model w E. g. , fit them into a distribution Analysis, simulation, measurement § Recorded workload Use past workload directly to drive evaluation Simulation, measurement OS Fall’ 02

Statistical characterization § Every workload item is sampled at random from the distribution of some random variable Workload is characterized by a distribution E. g. , take all possible job times and fit them to a distribution OS Fall’ 02

The Exponential Distribution § A lot of low values and a few high values § The distributions of salaries, lifetimes, and waiting times are often fit the exponential distribution § The distribution of Job runtimes Job inter-arrival times File sizes OS Fall’ 02

Exponential Probability Density Function (pdf) § If X has an exponential distribution with parameter , then its probability density function is given by: where OS Fall’ 02

The Exponential Distribution OS Fall’ 02

Mean and Variance § If X~Exponential( ), then its mean and variance are given by: OS Fall’ 02

Exponential Probabilities f(x) 0 a x OS Fall’ 02 0 a x

Cumulative Distribution Function § The cumulative distribution function (cdf), F(x), is defined as § For the Exponential distribution: OS Fall’ 02

The Exponential CDF OS Fall’ 02

Memoryless Property § The exponential is the only distribution with the property that § For modeling runtimes: the probability to run for additional b time units is the same regardless of how much the process has been running already in average OS Fall’ 02

Fat-tailed distribution § The real life workloads frequently do not fit the exponential distribution § Fat-tailed distributions: OS Fall’ 02

Pareto Distribution § The more you wait, the more additional time you should expect to wait § The longer a job has been running, the longer additional time it is expected to run OS Fall’ 02

Pareto dist. CDF OS Fall’ 02

Exponential vs Pareto OS Fall’ 02

Queuing Systems queue Disk A Disk B new jobs queue CPU queue finished jobs § Computing system can be viewed as a network of queues and servers § Shared queues are also possible OS Fall’ 02

The role of randomness § Arrival (departure) are random processes Deviations from the average are possible The deviation probabilities depend on the inter-arrival time distribution § Randomness makes you wait in queue Each job takes exactly 100 ms to complete If jobs arrive each 100 ms exactly, utilization is 100% But what if both these values are on average? OS Fall’ 02

Queuing analysis arriving jobs queue server departing jobs OS Fall’ 02

Little’s Law OS Fall’ 02

M/M/1 queue arriving jobs queue server departing jobs Both interarrival time and service time are exponentially distributed M stands for memoryless OS Fall’ 02

How average response time depends on utilization? § The job arrival and departure are § § approximated by Poisson processes The distribution of the number of jobs in the system in the steady state is unique Use queuing analysis to determine this distribution Once it is known, can be found Use the Little law to determine OS Fall’ 02

M/M/1 queue analysis 0 1 2 OS Fall’ 02 3

OS Fall’ 02

Response time (utilization( OS Fall’ 02

A Bank or a Supermarket? arriving jobs shared queue departing CPU 1 jobs CPU 1 departing jobs CPU 2 CPU 3 departing jobs arriving jobs departing CPU 4 jobs M/M/4 CPU 3 CPU 4 4 x M/M/1 OS Fall’ 02

It is a Bank! 4 x M/M/1 M/M/4 OS Fall’ 02

Summary § What are three main performance § § § evaluation metrics? What are three main performance evaluation techniques? What is the most important thing for performance evaluation? Which workload models do you know? What does make you to wait in queue? How response time depends on utilization? OS Fall’ 02

To read more § Notes § Stallings, Appendix A § Raj Jain, The Art of Computer Performance Analysis OS Fall’ 02

Next: § Processes OS Fall’ 02