Hacker News new | past | comments | ask | show | jobs | submit login

> running in a legacy data center costing millions, and used it as a showcase as to how you can take a legacy application and move it to the public cloud to both improve its usability and reduce the cost by a factor of 10.

Hahaha ... what? You reduced the cost by moving into one of the most overpriced hosting options? Or are you telling me that the government is getting massive discounts?




Yeah, there were two major factors. First its not cheap to run a professional data center, this isn't even a government thing almost all big companies are making this calculation and moving towards the big cloud providers. Second when you run your own data center you need to plan for the surge capacity. In this case they had well over 10x the amount of capacity they needed provisioned so they could handle a surge of traffic. In AWS we just set up an auto scaling group with a ansible baked AMI and only pay for what we needed.

Its expensive, but in this case the savings were massive ( saved almost 20m per year )


Can you explain exactly how this cost savings project is "changing the world"? You did a cost savings job on the cheap for a large government and didn't get paid for it is all I hear. They could afford to spend $20 million per year on a datacenter but can't afford to pay their workers in this program market rate salary??


That application we moved over was used to schedule doctors appointments. Before we got involved it was being used to schedule about 100 appointments a month. The failure rate of people trying to use it was over 95%. By moving to the cloud, using modern tools and better design we increased usage to 10,000 per week.

Veterans could get appointments to see their doctors. The number one kind of appointment was mental health. That is how it was changing the world. What we did with that one application can now be done thousands of times over for their other projects. We wrote the playbook.


Good for you, you did your job and fixed a broken system. Note that this does not require anything other than average technical competency. It's expected that people know how to do their jobs. And my original question still stands -- how are they able to afford spending $20 million on a broken system, but not able to pay their employees a fair, market rate salary?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: