PILOT SPIN
Spin Zone => Spin Zone => Topic started by: President-Elect Bob Noel on April 08, 2020, 05:57:41 AM
-
A $349 billion dollar loan program for small businesses.
The initial wave of applications apparently have crippled the SBA computers.
(there's an AP story on this)
And, I'm sure nobody could have anticipating this.
Now, in fairness, there is a huge difference between the processing needs of google, amazon, etc and the nominal computing needs of the SBA.
ps - we can bet large sums of OPM that President Trump will be blamed
-
In any competent private company a capacity study would have been done. That used to be one of my jobs.
But in government, powerful people order it done. Peons carry out their orders. The project fails. Nobody is at fault but the solution is more money. Nobody is at fault, but a scapegoat will be selected to throw under the bus. In this case, you are correct. That will be Trump.
-
In any competent private company a capacity study would have been done. That used to be one of my jobs.
sadly, it wouldn't take particularly long to get a rough idea. right?
-
sadly, it wouldn't take particularly long to get a rough idea. right?
Only if you want a rough answer, which is "we need more computing power"
I deal with capacity planning too and it's a precise art and inexact science at best.
-
Only if you want a rough answer, which is "we need more computing power"
I deal with capacity planning too and it's a precise art and inexact science at best.
"Order of Magnitude" stuff.
-
Only if you want a rough answer, which is "we need more computing power"
I deal with capacity planning too and it's a precise art and inexact science at best.
Too much computing power is never enough!
But with the various scalable compute clouds available it should be pretty easy. The need for onsite massive server rooms, especially for short-term use, is just not there. Elastic capacity: scale up automatically as needed, scale back down as workload drops. Management and oversight is much more difficult, but only if the company is not used to management and oversight.
-
Only if you want a rough answer, which is "we need more computing power"
I don't put much stock in someone who always replies "we need more computing power", there is a lack of credibility.
I dealt with a test organization that always replied "we need to fly the test jet" before knowing what actually needed to be tested.
edit: if my SME always said "we need more computing power", then I wouldn't need the SME. The top-notch SME will tell me how much more is needed and how it could work with less than optimum computing power.
-
I don't put much stock in someone who always replies "we need more computing power", there is a lack of credibility.
I dealt with a test organization that always replied "we need to fly the test jet" before knowing what actually needed to be tested.
That's an interesting quote, but I don't quite understand it. Can you go into a little more detail? But my expertise is in software. Boot the new software and see what happens is a good way to get the scope of catastrophe before settling down to the unit and integration testing. Probably not so much fun with aircraft.
-
That's an interesting quote, but I don't quite understand it. Can you go into a little more detail? But my expertise is in software. Boot the new software and see what happens is a good way to get the scope of catastrophe before settling down to the unit and integration testing. Probably not so much fun with aircraft.
I'm sorry, it would be inappropriate to go into more details.
But think system testing, not software integration testing. Think flight pay, even if it's in the back of the tube.
-
I'm sorry, it would be inappropriate to go into more details.
But think system testing, not software integration testing. Think flight pay, even if it's in the back of the tube.
Got it! Thanks!
-
I don't put much stock in someone who always replies "we need more computing power", there is a lack of credibility.
I dealt with a test organization that always replied "we need to fly the test jet" before knowing what actually needed to be tested.
edit: if my SME always said "we need more computing power", then I wouldn't need the SME. The top-notch SME will tell me how much more is needed and how it could work with less than optimum computing power.
The question was asked in the most raw form. I want to process more loans. Ok, you probably need more computing power.
I want my jet to go faster. What's the answer?
-
I want my jet to go faster. What's the answer?
Spend more money.
-
The question was asked in the most raw form. I want to process more loans. Ok, you probably need more computing power.
I wouldn't need a SME that wouldn't seek a clarification. And "I want to process more loans" wouldn't be the question. It would be "I want to process a surge of loan applications". The good SME would try to bound how many more applications could be processed with the existing capacity (or the capacity that could be added by the time the applications start hitting the systems).
I want my jet to go faster. What's the answer?
push the nose over and dive
-
push the nose over and dive
LOL!! ;D
-
I don't put much stock in someone who always replies "we need more computing power", there is a lack of credibility.
I dealt with a test organization that always replied "we need to fly the test jet" before knowing what actually needed to be tested.
edit: if my SME always said "we need more computing power", then I wouldn't need the SME. The top-notch SME will tell me how much more is needed and how it could work with less than optimum computing power.
I'm glad I never worked for you.
Proper capacity planning is not just about analyzing a new application and determining how much additional processing power you need.
Proper capacity planning begins in designing a scalable system that is sufficient for current and foreseeable loads, but can add hardware without major reconfiguration. Older mainframe technologies sucked at this, and many of the government computer systems are still older mainframe technology that are difficult not only to maintain but to scale up.
As to your statement, I disagree. It is silly to put a system in place that has way more power than might be needed for a few years. It is best to add hardware and scale up as needed with newer equipment.
-
I'm glad I never worked for you.
Proper capacity planning is not just about analyzing a new application and determining how much additional processing power you need.
Proper capacity planning begins in designing a scalable system that is sufficient for current and foreseeable loads, but can add hardware without major reconfiguration. Older mainframe technologies sucked at this, and many of the government computer systems are still older mainframe technology that are difficult not only to maintain but to scale up.
As to your statement, I disagree. It is silly to put a system in place that has way more power than might be needed for a few years. It is best to add hardware and scale up as needed with newer equipment.
I agree, to a point. The problem with old mainframe systems is that it's impossible to add capacity that's compatible with the old systems. It may be better to buy excess capacity now than hope that what's available in the future is compatible with what you have. Newer equipment does not always play nice with the old stuff.
-
I'm glad I never worked for you.
Proper capacity planning is not just about analyzing a new application and determining how much additional processing power you need.
Proper capacity planning begins in designing a scalable system that is sufficient for current and foreseeable loads, but can add hardware without major reconfiguration. Older mainframe technologies sucked at this, and many of the government computer systems are still older mainframe technology that are difficult not only to maintain but to scale up.
We agree more than you realize
As to your statement, I disagree. It is silly to put a system in place that has way more power than might be needed for a few years. It is best to add hardware and scale up as needed with newer equipment.
Well, there are situations where the fielding of additional capacity can be prohibitively expensive.
-
I agree, to a point. The problem with old mainframe systems is that it's impossible to add capacity that's compatible with the old systems. It may be better to buy excess capacity now than hope that what's available in the future is compatible with what you have. Newer equipment does not always play nice with the old stuff.
fyi, I meant to put a smiley face after my jab about being glad I didn't work for you. I'm sure you would have been a much better boss than many of my old bosses were. Their problem was largely that I didn't bend when I thought they were wrong.