Wednesday, October 22, 2008

Technical Economizing I (Capacity 2)

_

When working on technical projects it is often times necessary to economize based on the available alternatives, that is, as a result of scarcity, resources must be allocated in such a fashion that a minimum acceptable standard is met. The process by which economizing generally takes place is as follows:

1. Investigate alternatives
2. Identify minimum standard of acceptability
3. Choose alternative that at a minimum meets minimum standard of acceptability
4. Define rationale for choice of selected alternative
5. Analyze effectiveness of selected alternative (future)

Once again, the project capacity presents an excellent opportunity to showcase the issue at hand, technical economizing. The first step in the technical economizing process is to investigate alternatives for generating capacity requirements. The diagram below displays the high level component breakdown of ETS:






The technical economizing process will be applied to each component in the system and low level assumptions will be generated as a result:
Assumptions about assumptions: These are the assumptions from which all others will be based.
ASSUMPTION 1: Assumptions are generated based on the best case alternative. While best case alternative may not always be the most likely, it does provide the simplest model from which to work and from which to generate a working model.
ASSUMPTION 2: Best case alternatives will be updated as information becomes available. It is understood that in many instances the best case alternative cannot realistically be selected for the final system. Cases in which the decision to select an alternative based on best case logic are denoted as part of the alternative definition.
INVESTIGATE ALTERNATIVES
1. Operations – The operational component layer is defined by the operational limitations imposed on the system including:
a. Availability of technicians to work the system
i. Known – No limitations regarding technician availability or required expertise have been specified.
ii. Alternatives
1. ETS Technicians will never be available – Starting with this assumption gets nowhere fast…
2. ETS Technicians will always be available – This is the best case alternative and a good point of start from which to work.
3. ETS Technicians will be available during normal work hours – While this is a possible alternative, the case of constant availability is simpler just to assume technicians will be available.
4. ETS Technicians will not be qualified to run the system – This is probably a safe assumption as technicians will not know how to use the system to start with.
5. As Technicians will not be qualified to start with, training will be necessary
6. There is a minimum amount of education \ experience required to operate the ETS system.
7. ETS operator should be selected based on their education and experience.
iii. Assumptions:
1. ETS Technicians will always be available to operate the system - even if the system should require operating at odd hours of the day and night.
2. ETS Technicians will be qualified to run the system.
3. ETS Technicians will be available for training as needed.
4. ETS Technicians will not be qualified to run the system initially.
5. ETS Technicians will require training.
iv. Questions:
1. When will technicians be available to operate system?
2. How many technicians will be available?
3. How much training will technicians require?
4. How much education \ experience will be required of ETS technicians?
b. Execution time for technician to process each test operation
i. Load Carrier
1. Physical load of carrier:
a. Known – Currently, the system provides the technicians one hour to load a carrier.
b. Alternatives:
i. Technicians will load carriers as soon as required by the scheduler.
ii. Some amount of time must be allocated by the scheduler to provide for carrier insertion by technicians.
iii. Technicians will load carriers immediately; no time is required – although best case, not a realistic alternative.
iv. Scheduling will take place before physical carrier insertion – So technician knows where to place carrier.
v. The technician will be allowed one hour to insert carriers following scheduling and prior to first command of test. Some amount of time is required. Why one hour? Because it’s the way we do it now and why not?
vi. The technicians will be both physically fit enough and cognitively developed enough to handle the insertion of carriers into the system.
vii. The docking station and carrier will be designed in such a fashion that a trained technician who is sharp of mind and sound of body will be able to make the insertion.

c. Assumptions:
1. Scheduling will take place before physical carrier insertion.
2. The technician will be allowed one hour to insert carriers following scheduling and prior to first command of test.
3. Some amount of time must be allocated by the scheduler to provide for carrier insertion by technicians.
4. The technicians will be both physically fit enough and cognitively developed enough to handle the insertion of carriers into the system.
5. The docking station and carrier will be designed in such a fashion that a trained technician who is sharp of mind and sound of body will be able to make the insertion.

d. Questions:
1. Does scheduling need to take place prior to carrier insertion?
2. How long should the technician be given to insert a carrier?
3. Does the time given to a technician need to be the same across recipes?
4. What are the physical and intellectual requirements of making a carrier insertion?
2. Virtual load of carrier
a. Known – Depending on the type of system, the current system requires a carrier insert to be clicked following carrier insertion.
b. Alternatives:
i. Technicians must work with in conjunction with the system to provide system notifications of carrier insertions.
ii. Iodine PXI’s do not send out a carrier insert message upon carrier insertion.
iii. Pulse PXI’s send out a carrier insertion upon carrier insertion.
iv. Technicians will be required to take action to produce carrier insertions when working with Iodine systems.
v. Technicians will not be required to take action to produce carrier insertions when working with Pulse systems.

c. Assumptions:
1. The technician should not have to click a button following physical insertion of the carrier.
2. Iodine PXI’s do not send out a carrier insert message upon carrier insertion.
3. Pulse PXI’s send out a carrier insertion upon carrier insertion.
4. Technicians must work with in conjunction with the system to provide system notifications of carrier insertions.
5. Iodine PXI’s do not send out a carrier insert message upon carrier insertion.
6. Pulse PXI’s send out a carrier insertion upon carrier insertion.
7. Technicians will be required to take action to produce carrier insertions when working with Iodine systems.
8. Technicians will not be required to take action to produce carrier insertions when working with Pulse systems.

d. Questions:
1. Is it necessary to mandate clicking a button following carrier insert?
2. Why is there a difference between PXI behaviors based on type?

ii. Remove Carrier
1. Physical removal of carrier
a. Known – Currently, the system provides the technicians one hour to remove a carrier.
b. Alternatives:
i. Technicians will remove carriers as required by the scheduler. But right now there are no scheduled remove requirements.
ii. Some amount of time must be allocated by the scheduler to provide for carrier removal by technicians.
iii. Technicians will remove carriers immediately; no time is required – although best case, not a realistic alternative.
iv. The technician will be allowed one hour to remove carriers following scheduling and prior to first command of test. Some amount of time is required.
v. The technicians will be both physically fit enough and cognitively developed enough to handle the removal of carriers into the system.
vi. The docking station and carrier will be designed in such a fashion that a trained technician who is sharp of mind and sound of body will be able to make the removal.

c. Assumptions:
1. The technician will be allowed one hour to remove carriers following scheduling and prior to first command of test.
2. Some amount of time must be allocated by the scheduler to provide for carrier removal by technicians.
3. The technicians will be both physically fit enough and cognitively developed enough to handle the removal of carriers into the system.
4. The docking station and carrier will be designed in such a fashion that a trained technician who is sharp of mind and sound of body will be able to make the removal.

d. Questions:
1. Will scheduling be dependent on carrier removal within one hour?
2. How long should the technician be given to insert a carrier?
3. Does the time given to a technician need to remove a carrier need to be the same across recipes?
4. What are the physical and intellectual requirements of making a carrier removal?

2. Virtual removal of carrier
a. Known – Depending on the type of system, the current system requires a carrier remove to be clicked following carrier insertion.
b. Alternatives:
i. Technicians must work with in conjunction with the system to provide system notifications of carrier removals.
ii. Iodine PXI’s do not send out a carrier remove message upon carrier insertion.
iii. Pulse PXI’s send out a carrier remove upon carrier removal.
iv. Technicians will be required to take action to produce carrier removed when working with Iodine systems.
v. Technicians will not be required to take action to produce carrier removals when working with Pulse systems.

c. Assumptions:
1. The technician should not have to click a button following physical removal of a carrier.
2. Iodine PXI’s do not send out a carrier remove message upon carrier removal.
3. Pulse PXI’s send out a carrier remove upon carrier removal.
4. Technicians must work with in conjunction with the system to provide system notifications of carrier removals.
5. Iodine PXI’s do not send out a carrier insert message upon carrier removal.
6. Pulse PXI’s send out a carrier remove upon carrier removal.
7. Technicians will be required to take action to produce carrier removed when working with Iodine systems.
8. Technicians will not be required to take action to produce carrier removed when working with Pulse systems.

d. Questions:
1. Is it necessary to mandate clicking a button following carrier removal?
2. Why is there a difference between PXI behaviors based on type?

c. Timing of required system (PXI) maintenance \ Calibration
a. Known – The PXI will require maintenance
b. Alternatives:
i. There will be a fixed period of time down for maintenance.
ii. There will be no maintenance. Chosen for simplicity of working model.
iii. There will be no calibration. Chosen for simplicity of working model.
iv. There will be continuous maintenance.
v. The system will be required to take docking stations offline for maintenance.
vi. The system will be required to take PXIs offline for maintenance.
vii. The PXI will always be up and running 24 hours a day and 7 days a week.
viii. The PXI will be in a continuous state of maintenance.
ix. The PXI will be in a continuous state of calibration.
x. There will exist a calibration recipe.

c. Assumptions:
1. There will be one 24 hour maintenance window each month dedicated to maintenance and calibration.
2. The system will be required to take docking stations offline for maintenance.
3. The system will be required to take PXIs offline for maintenance.
4. The system will be required to take the CPC offline for maintenance.
5. The system will be required to take PXIs offline for calibration.
6. The PXI will always be up and running 24 hours a day and 7 days a week.
d. Questions:
1. What are the requirements for maintenance and calibration?

d. Network timing and availability
a. Known – The will be a server used to facilitate testing of the PXI. The server will require a network connection and maintenance. There will also be a database used to maintain system state and store test results.
b. Alternatives:
i. The network will always be available.
ii. The network will never be available.
iii. The system will need an active connection to the database to start a test.
iv. The system will not need a connection to the database to complete a test once started.
v. The system will use Velocity technology to support the need to complete a test without a database.
vi. Network maintenance will not affect system performance. Best case assumption made for working model simplicity.
vii. The network will always be up and running 24 hours a day and 7 days a week. Best case assumption made for working model simplicity.
c. Assumptions:
1. The system will need an active connection to the database to start a test.
2. The system will not need a connection to the database to complete a test once started.
3. The system will use Velocity technology to support the need to complete a test without a database.
4. Network maintenance will not affect system performance.
5. The network will always be up and running 24 hours a day and 7 days a week.
d. Questions:
1. Is velocity the optimal solution to handling the requirement of no database?
2. What are the system requirements for calibration?
3. What are the system requirements for maintenance?
4. How does the current system handle maintenance and calibration?

e. PXI Operational Considerations
a. Known – The PXI box will be necessary to run a recipe.
b. Alternatives:
i. The PXI will run 24 hours a day and 7 days a week
ii. The CPC will be available during recipe execution
iii. The PXI will be available during recipe execution
c. Assumptions:
1. The PXI will run 24 hours a day and 7 days a week
2. The CPC will be available during recipe execution
3. The PXI will be available during recipe execution
d. Questions:

2. ETS Software – The ETS software layer is defined by the limitations imposed on the system by software run on the control PC (CPC) including:
a. Time to send each command in the recipe
a. Known – As part of the recipe, there will be a maximum amount of time for the system to process each command. Detecting and sending each test command is one portion of the total system processing time.
b. Alternatives:
i. The time it takes for the CPC to detect and send a command is one part of the whole time allowed by the CPC as part of total command processing time.
ii. The time it takes for the CPC to detect and send a command will be less than one second.
iii. The time it takes for the CPC to detect and send a command will have such a small value it will not be relevant to the general estimate for total command processing time.
c. Assumptions:
1. The time it takes for the CPC to detect and send a command will be less than one second.
2. The time it takes for the CPC to detect and send a command will have such a small value it will not be relevant to the general estimate for total command processing time.

d. Questions:
1. Is the time it takes the CPC to detect and send a command significant to consider as part of the general time estimation for the allowed time to process a command?

b. Time to store recipe results
a. Known – As part of the recipe, there will be a maximum amount of time for the system to process each command. Storing recipe results is one portion of the total system processing time.
b. Alternatives:
i. The time it takes for the CPC to store recipe results is one part of the whole time allowed by the CPC as part of total command processing time.
ii. The time it takes for the CPC to store background measurements will be less than one second.
iii. The time it takes for the CPC to store background measurements will have such a small value it will not be relevant to the general estimate for total command processing time.
iv. The storing of pulse train results will take place on a secondary thread so as not to effect total command processing time.

c. Assumptions:
1. The time it takes for the CPC to store background measurements will be less than one second.
2. The time it takes for the CPC to store background measurements will have such a small value it will not be relevant to the estimate for total command processing time.
3. The storing of pulse train results will take place on a secondary thread so as not to effect total command processing time.
d. Questions:
1. Is the time it takes the CPC store recipe results significant enough to consider as part of the general time estimation for the allowed time to process a command?

c. Time to display system status
a. Known – The system status display is used to issue carrier insert and carrier remove messages to the system.
b. Alternatives:
1. The technician will perform actions required to facilitate testing in a timely manner.
2. Interaction with the system status display will not affect capacity calculation.
c. Assumptions:
1. As directed by the status display screen, the technician will perform actions required to facilitate testing in a timely manner.
2. Interaction with the system status display will not affect capacity calculation.
d. Questions:
1. Does the system status display effect capacity?

d. Requirements associated with error detection and recovery
a. Known – The system will have some means of detecting a system error.
b. Alternatives:
i. System performance may be estimated using normal flow of execution.
ii. Error detection and correction will not affect system performance.
c. Assumptions:
1. System performance may be estimated using normal flow of execution.
2. Error detection and correction will not affect system performance.
d. Questions:
i. What kind of error correction and detection will the system support?

3. Scheduling - The scheduling layer is defined by the limitations imposed on the system by the scheduling algorithm run within ETS software including:
a. Acceptable time into the future the scheduling algorithm will look relative to the total test time
a. Known – The current software will look as far into the future as necessary to make a reservation.
b. Alternatives:
1. The software will be bound by a finite future reservation window.
2. The software will be bound by a finite present reservation window.
3. The software will not be bound and look as far into the future as necessary to make a reservation. Selected for simplicity.
4. The software will not be bound by start time and look as early into the present as necessary to make a reservation. Selected for simplicity.
c. Assumptions:
1. The software will not be bound by a future window and look as far into the future as necessary to make a reservation.
2. The software will not be bound by lower reservation bound and look as early into the present as necessary to make a reservation.
d. Questions:
1. Would it be worthwhile to define constraints associated with when a reservation may take place with respect to the current time?
2. Would it be worthwhile to maintain a grid of availability in the database? I personally think it would…

ii. Second level scheduling
a. Known – The system does not currently support layered scheduling. To be scheduled against, a docking station must be open and available for the duration of the test.
b. Alternatives:
i. The software will not support layered scheduling – scheduling against docking stations that are full or reserved. Best case selection.
ii. The software will support scheduling into the future, even against docking stations that are currently occupied or reserved.
c. Assumptions:
1. The software will not support layered scheduling – scheduling against docking stations that are full or reserved.
d. Questions:
1. Is there a need for the system to allow scheduling into the future?

b. Allowable start time of carriers within a group
a. Known – Carriers will be scheduled in groups.
b. Alternatives:
i. Carriers will be scheduled with no consideration for of what group they are a part
ii. The scheduled start time of carriers within the same group will be within one hour of each other
c. Assumptions:
1. Carriers will be scheduled with no consideration for of what group they are a part
d. Questions:
1. Are groups considered when scheduling a particular carrier for test?

4. Recipe - The recipe layer is defined by the limitations imposed on the system by the scheduling algorithm run within ETS software including:

a. Time between each command
a. Known – There may be a constant time between commands contained in the recipe.
b. Alternatives:
i. There will be 30 minutes between the scheduled start times of each command.
ii. There will be a variance between the scheduled start times of each command.
c. Assumptions:
1. There will be 30 minutes between the scheduled start times of each command.
d. Questions:
1. Will the time between commands be constant?

b. Total length of the recipe
a. Known – The PXI will run recipes to completion.
b. Alternatives:
i. Recipes run on the PXI will be of variable lengths.
ii. The total execution time of a recipe will not exceed 86 hours in length. Used for SAT recipe. Chosen for simplicity.

c. Assumptions:
1. The total execution time of a recipe will not exceed 86 hours in length.
d. Questions:


5. The PXI command execution layer is defined by the limitations imposed on the system by the time it takes commands to execute software including:
a. Maximum time allotted for apply load command to run
a. Known – The apply load command will be sent to the PXI as part of the recipe run.
b. Alternatives:
i. The total time to execute an apply load command will not exceed 20 seconds. Apply load contains 2 background measurements. Using 10 seconds for execution time of background measurement, 2 * (10) seconds = 20 seconds. Chosen for simplicity.
ii. There will be a variance between time allowed to execute an apply load command.
c. Assumptions:
1. The total time to execute an apply load command will not exceed 20 seconds.
d. Questions:

b. Maximum time allotted for remove load command to run
a. Known – The remove load command will be sent to the PXI as part of the recipe run.
b. Alternatives:
i. The total time to execute a remove load command will not exceed 20 seconds. Remove load contains 2 background measurements. Using 10 seconds for execution time of background measurement, 2 * (10) seconds = 20 seconds. Chosen for simplicity.
ii. There will be a variance between time allowed to execute an apply load command.

c. Assumptions:
1. The total time to execute a remove load command will not exceed 20 seconds.

d. Questions:

c. Maximum time allotted for change load command to run
a. Known – The change load command will be sent to the PXI as part of the recipe run.
b. Alternatives:
i. The total time to execute a change load command will not exceed 30 seconds. Change load contains 3 background measurements. Using 10 seconds for execution time of background measurement, 3 * (10) seconds = 30 seconds. Chosen for simplicity.
ii. There will be a variance between time allowed to execute an apply load command.
c.
d. Assumptions:
1. The total time to execute a change load command will not exceed 30 seconds.
e. Questions:

d. Maximum time allotted for background measurement command to run
a. Known – The background measurement command will be sent to the PXI as part of the recipe run.
b. Alternatives:
i. The total time to execute a background measurement command will not exceed 10 seconds. Chosen for simplicity.
ii. There will be a variance between time allowed to execute an apply load command.

c. Assumptions:
1. The total time to execute background measurement command will not exceed 10 seconds.

d. Questions:

e. Maximum time allotted for pulse command to run
a. Known – The pulse train command will be sent to the PXI as part of the recipe run.
b. Alternatives:
i. The total time to execute a pulse train command will not exceed 2 minutes. Chosen for simplicity.
ii. There will be a variance between times allowed to execute a pulse train command.

c. Assumptions:
1. The total time to execute a pulse train command will not exceed 2 minutes.
d. Questions:

f. Number of commands PXI can simultaneously process
a. Known – The PXI can process one command at a time.
b. Alternatives:
i. PXI can only process one command at a time
ii. PXI can process multiple commands at a time
c. Assumptions:
1. PXI can only process one command at a time
d. Questions:


The next step will be to explicitly define a minimum standard of acceptability, weigh assumptions against each other to eliminate redundancy effectively “normalizing” our list of assumptions and provide rationale for each assumption included in the final list.

No comments: