Chapter 7: Data Collection

Question 1

What are the five types of data needed for SPE

Answer 1

1. Key Performance scenarios
2. Performance objectives
3. Execution Environment
4. Software Resource Requirements
5. Computer Resource Requirements

Question 2

What are heavy hitters?

Answer 2

Functions that take up the most resources

Question 3

What do we need to estimate in key performance scenarios?

Answer 3

• Workload intensity
• Arrival rate of requests
• Number of concurrent users

Question 4

What does the execution environment consist of?

Answer 4

1. Hardware and network
2. Operating system
3. Database usage
4. Other software that uses resources

Question 5

How do we estimate software resource requirements?

Answer 5

Can use work units

Question 6

What are the types of computer resource requirements?

Answer 6

1. CPU time for work units
2. Number of I/O
3. Message size

Question 7

How do we map software resource requirements into computer resource requirements?

Answer 7

Processing overhead matrix

Question 8

What is a walkthrough?

Answer 8

Informal review where member of development team discusses system

Question 9

What is a performance walkthrough?

Answer 9

Similar to walkthrough except goal is to gather the information needed to construct performance models.

Question 10

What is a user representative?

Answer 10

Provides data on how the system will be used, how often and typical usage conditions

Question 11

What is a system architect

Answer 11

Provides overview of system and its components along with use cases

Question 12

What is a performance specialist?

Answer 12

Someone who is familiar with measurements of target configuration and can either provide data or perform measurements to get data.

Question 13

When should performance walkthroughs be conducted?

Answer 13

1. Early stages, should be part of overall feasibility study to see if the project can continue.
2. When system evolves. The frequency of performing walkthroughs depends on performance risk.

Question 14

Are resource requirements the most difficult specifications to esimate?

Answer 14

True

Question 15

What are use measurements?

Answer 15

1. Start with an estimate
2. Use model results to study accuracy of model results to estimates
3. Modify it with accurate estimates as soon as possible

Continue to evaluate the system as it evovles

Question 16

What are study measurements?

Answer 16

Study similar systems to develop a frame of reference to estimate resource requirements

Question 17

What should you do if you haven’t participated in SPE studies before?

Answer 17

Use a mentor to assist and advise you on resource estimates in initial walkthroughs

Question 18

When uncertainty is an issue in measurements what should be used?

Answer 18

Use best-worse case estimates

1. First use best case as if that doesn’t work then no point in continuing
2. Use worst case estimates to find anomalies in system.

Question 19

What happens if worst-case estimates pass the system?

Answer 19

No point in developing more precise metrics.

Question 20

What should be estimated?

Answer 20

1. Dominant Resource Usage
2. I/O Requirements
3. Network messages

Question 21

What are dominant resources?

Answer 21

Resources that account for most of elapsed time in scenario

Question 22

What do we need to estimate I/O Requirements?

Answer 22

1. Number of physical I/O
2. I/O Service times
3. Random or sequential access

Question 23

How do we estimate number of physical I/O?

Answer 23

1. First method: Indirectly method of estimating I/O. Estimate the logical I/Os and then estimate hit ratio
2. Second method: Directly estimate I/O. Estimate which I/O requests might be stored in a cache/buffer, count only ones that would access physical I/O.

Question 24

How does read random access effect number of I/O?

Answer 24

For read requests that random access an indexed file or table the number of logical I/O is i+1,

where i is the number of files in file index and
1 is the I/O to read the data record

Question 25

How does write random access effect number of I/O?

Answer 25

For write requests that random access an indexed file or table the number of logical I/O is i+1, where i is the number of files in file index and 1 is the I/O to read the data record where 1 is best case as index is not effected and i+1 is worst case where all levels of index are effected

Question 26

How do we estimate network messages

Answer 26

1. Number of messages transferred 2. Number of characters transferred must account for the characters in message headers and the application messages

Question 27

What is the overhead for transferring a message?

Answer 27

1. CPU path length for preparing message 2. physical I/O for logging message 3. Time to transfer both the message and other extra required messages for protocol

Question 28

Where can we find overhead resource requirements?

Answer 28

Model libraries

Question 29

What happens if you cannot find overhead resource requirements?

Answer 29

Perform a performance measurement study

Question 30

What should you do after estimating computer resource requirements?

Answer 30

Make a library for others to use as well.

Question 31

How to have a successful performance walkthrough?

Answer 31

1. Partition long walkthroughs into smaller ones 2. Minimize action items 3. Include Solution Alternatives with results