Posted by Max Dunn
Sat, 07 Jan 2012 20:44:59 GMT | no comments
I have been working on the RealMealz site and am now looking for a suitable host server. Currently, RealMealz is in beta and has just light use, but we expect that once released the use will go up dramatically, so we want to be able to easily scale. I looked at Amazon EC2, Heroku and RackSpace Cloud Servers and the table below summarizes my findings:
Test: Login, hit Discover page 5 times Use a 5 second ramp up
Average Discover page load time (Apache + Passenger unless otherwise indicated)
Server
1 user
5 user
10 user
Cost/Mo
EC2, Micro
2s
20s
40s
$9
EC2, Micro, Thin
2s
23s
42s
$9
EC2, 3-Micros, balanced. Apache/Passenger
2s
6s
9s-18s (40s max)
$44
EC2, hi-cpu medium
2s
4s
6s
$75
EC2, hi-cpu medium, Thin
2s
5s
10s
$75
BlueBox
2s
2s
5s
$79
Heroku, 1 dyno, Thin
2s
3s
6s
$15
Heroku, 2 dynos, Thin
2s
3s
4s
$50
Heroku, 3 dynos, Thin
2s
2s
6s (20 users!)
$86
Rackspace
2s
7s (max 66s)
40s+
$22
Rackspace, Thin
2s
4s
9s
$22
Rackspace, Unicorn
2s
4s
8s
$22
Notes.
1. Users are concurrent users
2. I selected the Discover page since this is our most CPU intensive page. This test is not that realistic in actual use, but serves as a worse case example.
3. The Rackspace server had only 512Mb of memory and for the 10 user case for Apache it started swapping out. If I was using their next size up with 1Gb of memory, the cost would have been $44 per month but wouldn’t have suffered this performance issues as drastically.
4. The Amazon Micro instances provide bursts up to 2 ECUs, but if used at a high level for more than a few seconds, get throttled back greatly for about 15 seconds. This is why the 10 user case was so high, even for the 3 instance, load balanced case.
4a. The Amazon servers are a little underpowered. Even the hi-cpu instance that provides 5 ECUs is not all that impressive. My BlueBox server which has 4 VPSes on the server performed better.
5. It wasn’t clear how the Rackspace Cloud Servers could be easily scaled since RackSpace doesn’t seem to provide a load balancer for them.
6. Amazon’s EC2 instances are a bit of a pain to scale. You can’t add an extra instance to an existing load balancer, so you need to delete the current one then create a new one. There is an autoscaling feature, but this only works from the command line and I didn’t test it.
7. Heroku offered pretty amazing performance and was super easy to scale by just typing “heroku scale web=3”. The prices include $15/month for the larger database, although under 5Mbs is free.
8. With 3 dynos, Heroku was able to handle 20 concurrent users where the other tests were only with 10.
9. A disadvantage of Heroku is that it doesn’t offer shell access and there is no way to access the database natively. We do have some table manipulation built into our program, but if there are major changes or corruptions, we will need to export the data, fix it, and then re-import.
In the end, we are going with Heroku because it has great performance, easy to scale and is low cost.
Posted by Max Dunn
Wed, 16 Sep 2009 23:22:53 GMT | no comments
Many business are built entirely around extrinsic motivators – carrots and sticks. That’s actually fine for many kinds of 20th century mechanistic tasks. But for 21st century cognitive tasks, monetary rewards just don’t work.
This has been confirmed over and over again. Dan Ariely found that when tasks called for even rudimentary cognitive skill, a larger reward led to poorer performance. The London School of Economics reports: “We find that financial incentives can result in a negative impact on overall performance.”
Posted by Max Dunn
Mon, 07 Sep 2009 18:07:11 GMT | no comments
John Mackey, the founder and CEO of Whole Foods, has a great definition of what a business should be:
“Business has noble purposes: to provide goods and services that improve its customers’ lives, to provide jobs and meaningful work for employees, to create wealth and prosperity for its investors, and to be a responsible and caring citizen.”
Posted by Max Dunn
Sat, 21 Feb 2009 19:39:13 GMT | 1 comment
With the recent financial breakdown, it was refreshing to hear of a new type of bank at the VLAB event Upside of the Downturn on January 20th, 2008.
Renaud Laplanche, Founder & CEO of the Lending Club described how normal banks pay 2-3% but collect 16%. Lending Club lenders on the other hand, get 10.29% while borrowers pay 13% on average. The Lending Club narrows the spread rates by transacting over the Internet and having people lending to other people. As a lender, you can choose which borrower to lend to which creates a connection and social responsibility among users. Borrowers are also carefully screened and 85% are declined, so only prime borrowers are accepted. This leads to a low default rate of 2.7% and a late rate of 3.8%. The Lending Club is experiencing high growth rates of 50-100%, and in December of 2008, $2.5M was placed.
Maybe this is a new model for more sustainable banking!
Posted by Max Dunn
Fri, 04 Apr 2008 17:26:58 GMT | no comments
The other day, I was talking to my friend Bill Keller about the best way to hire people. He mentioned that it was important to have a defined company culture and to hire people that fit in with that culture. I thought this was a good idea. We also agreed that the normal process of interviewing people across a desk was almost worthless and only proved whether people had interviewing skills, but not whether they had the necessary job skills.
Another thing we agreed on was that most people will only hire people that are not as good as they are. Generally this is a self confidence issue because people worry that the person they hire could take over their job.
I certainly have fallen into this, in particular in a job a had right before I graduated from college. I had to hire a programmer to continue my work while I went home to fulfill a summer obligation. I was very conscious to hire someone that was good, but not too good so that I would have a job when I got back!
So if this is how most people behave, what happens when you allow a large group of people to decide who to hire? You end up with someone that is less capable than everyone making the decision!
So while it is good to involve other people in the hiring process and get their feedback, the worst thing to do when hiring someone is to do it based on the consensus of a group.
Posted by Max Dunn
Tue, 01 Apr 2008 22:30:28 GMT | no comments
The Surfline report said this morning about 38th Street:
Pretty much flat this morning
and about Pleasure Point
Maybe you could get up on a longboard but its pretty desperate.
Well Mark and I went anyways and were rewarded with a fantastic day of surf. For the first 40 minutes, we were the only two out at 38th Street. Then a few other people came, but it never got crowded. The waves weren’t big, lots of 2 and 3 foot waves, but they had nice shape and Mark and I caught a ton of waves and had a great time practicing walking on the board. The water was still a little cold, but the sun was out and the weather was nice and warm. I am not sure if Surfline meant this as an April’s Fool joke, but the bad report sure kept away the crowds and made it a fantastabulous day for us!
Posted by Max Dunn
Tue, 25 Mar 2008 22:48:20 GMT | 1 comment
It is amazing how distorted a story can be when you hear only rumors about it. For instance, it was my impression that the recent trouble in Tibet was caused by the Chinese cracking down on the Tibetans, and this was also the impression of several people I talked to. But then I read “Trashing the Beijing Road” in the March 22nd, 2008 edition of The Economist and got quite a different picture.
It turns out that it wasn’t the Chinese security forces that started the trouble – the Tibetans started rioting after hearing rumors that the police beat a couple of Buddhist monks. The Tibetans smashed into non-Tibetans shops, pulling merchandise from the stores and lighting it on fire. Many Han Chinese were quick to flee, otherwise the death toll might have been much higher than the 13 people the government reported were killed by rioters, mostly in fires. Shops owned by Tibetans were marked with traditional white scarves – almost every other one was wrecked. During the night, fire trucks were sent in, backed by armored personnel carriers with riot police, to put out the biggest fires. But the police did not immediately move into the alleys where rioting continued for a second day.
Overall, the efforts of the security forces appeared relatively measured. Rather than going in with guns blazing, which was the tactic used in the last outbreak of anti-Chinese unrest in Lhasa in 1989, they chose to let the rioters vent their anger, then gradually go in several days later. The government’s decision not to declare martial law showed its concern about the Olympics, since the Olympic flame is due to arrive in Lhasa on June 20th.
While the rioting was started by the Tibetans, the Chinese government is not completely blameless. The surge of spending in Tibet and the region’s high growth rate has been seen as benefitting mostly the Han Chinese. Tibetans also resent the hardline tactics of Tibet’s party chief – a Han – who is seen as harassing the Tibetans and stepping up the official criticism of the Dalai Lama. Nonetheless, whether or not it was true the the monks were beaten, it was the Tibetans that started the riots and the security forces that showed restraint.
So here is an example where the real facts of a story were quite different from the initial impression. We should all keep this in mind when hearing rumors and wait for the full story before making any judgments.
Posted by Max Dunn
Thu, 13 Mar 2008 15:02:17 GMT | no comments
Most people assume that teachers make very little money. In some parts of the country, this is very true. However, in the Cupertino Union School District in California (CUSD), teachers are paid some of the highest salaries in the country.
For instance, a first year teacher starts at $51,071 per year. After 7 years, they can make $62,321. After 15 years, it jumps to $80,836. (These last two salary brackets require additional semester units be completed.)
What about work hours? Their contract requires teachers to work 6.75 hours per day (7.25 hours with a 30 minute lunch) for 187 day s a year. However, teachers will often spend more time preparing lesson plans, having parent conferences, coaching a team or working on a school play. So let’s assume most teachers work a 40-hour week. But teachers only work for 187 days a years, less than the average of 225 days that most Americans work.
This means that teachers work about 20% less days. Factoring this into their salary gives the equivalent (using the figures above) of about $61,000, $74,000 and $97,000 respectively.
Posted by Max Dunn
Mon, 03 Mar 2008 16:08:56 GMT | 1 comment
Even while we continue to work on and define what Web 2.0 really is, many people have started to ask what is next; what is Web 3.0? Here is the answer: Web 3.0 is when all the Web 2.0 applications can start working together.