Thanks to everyone that came out to hear my presentation tonight at the Utah QA group!
Here are the slides in PDF format. I'll be posting the video soon on YouTube.
http://www.riceconsulting.com/public_pdf/Principles%20Before%20Practice%20-%20SLC.pdf
Thanks,
Randy
Tuesday, October 22, 2013
Tuesday, October 01, 2013
Obamacare, Where Art Thou?
There are so many things that come to my mind about the Obamacare launch today.
1) That so many state exchanges that experienced problems make me wonder if none of these IT shops have heard about performance testing? Failover servers? Load balancing? In a way, this was almost engineered (my apologies to all engineers) to fail because it's the "everybody show up at the same scenario".
2) There were reportedly functional defects in some states that prevented people from even setting up a user account.
3) Once again, the idea prevailed that just because someone in government declared "Let there be a system for...", people assumed the resulting system would be on schedule, adequate quality, etc. There are no magic IT wands. But, on the other hand, how hard is it to build a web site that is just a directory to other sites? Of course, I'm just the consultant looking in from the outside. I've seen simple problems grow into complex monsters once vendors and government meet.
4) Then, of course, there are the flaws in the requirements concerning rate calculations.
I'm glad my state of Oklahoma opted out of building its own exchange.
I will be surprised if the problems are resolved quickly. I've seen these situations before and the more people try and fail to get access, the more they keep trying. It's a death spiral of performance.
Maybe, the people from United Airlines and the various state exchanges could get together and we could all have free insurance!
******* Update *******
10/6/13
In USA Today (http://www.usatoday.com/story/news/nation/2013/10/05/health-care-website-repairs/2927597/) we find the following:
"U.S. Chief Technology Officer Todd Park said the government expected HealthCare.gov to draw 50,000 to 60,000 simultaneous users, but instead it has drawn as many as 250,000 at a time since it launched Oct. 1." and "These bugs were functions of volume,'' Park said. "Take away the volume and it works.''
So, it appears that one contributing factor to the "bugs" (I would suggest this is system failure, not just a "bug") is that the performance targets were set way too low. This is like the infamous Victoria's Secret online fashion show failure at halftime of the Super Bowl a few years back. In performance testing of new launches, you have to take into account the curiosity factor. In the case of Obamacare, you tell 300 million people that a certain day is the day to check it out and expect only 60,000 people to show up? Come on, guys, you have to set your sights higher than that.
This is a great lesson in performance testing. You always go for high numbers for big launches (Like the Facebook IPO). Unless, of course, you want to go the public apology route.
1) That so many state exchanges that experienced problems make me wonder if none of these IT shops have heard about performance testing? Failover servers? Load balancing? In a way, this was almost engineered (my apologies to all engineers) to fail because it's the "everybody show up at the same scenario".
2) There were reportedly functional defects in some states that prevented people from even setting up a user account.
3) Once again, the idea prevailed that just because someone in government declared "Let there be a system for...", people assumed the resulting system would be on schedule, adequate quality, etc. There are no magic IT wands. But, on the other hand, how hard is it to build a web site that is just a directory to other sites? Of course, I'm just the consultant looking in from the outside. I've seen simple problems grow into complex monsters once vendors and government meet.
4) Then, of course, there are the flaws in the requirements concerning rate calculations.
I'm glad my state of Oklahoma opted out of building its own exchange.
I will be surprised if the problems are resolved quickly. I've seen these situations before and the more people try and fail to get access, the more they keep trying. It's a death spiral of performance.
Maybe, the people from United Airlines and the various state exchanges could get together and we could all have free insurance!
******* Update *******
10/6/13
In USA Today (http://www.usatoday.com/story/news/nation/2013/10/05/health-care-website-repairs/2927597/) we find the following:
"U.S. Chief Technology Officer Todd Park said the government expected HealthCare.gov to draw 50,000 to 60,000 simultaneous users, but instead it has drawn as many as 250,000 at a time since it launched Oct. 1." and "These bugs were functions of volume,'' Park said. "Take away the volume and it works.''
So, it appears that one contributing factor to the "bugs" (I would suggest this is system failure, not just a "bug") is that the performance targets were set way too low. This is like the infamous Victoria's Secret online fashion show failure at halftime of the Super Bowl a few years back. In performance testing of new launches, you have to take into account the curiosity factor. In the case of Obamacare, you tell 300 million people that a certain day is the day to check it out and expect only 60,000 people to show up? Come on, guys, you have to set your sights higher than that.
This is a great lesson in performance testing. You always go for high numbers for big launches (Like the Facebook IPO). Unless, of course, you want to go the public apology route.
Subscribe to:
Posts (Atom)