Fighting the Internet

Last week our main Internet connection went down. The office didn’t even notice because we are hosting Ankhos in the basement and not in “The cloud”.  It happens.

Despite this outage, we have been working on getting an architecture set up for hosting Ankhos from Amazon.  There are lots of things to consider:

Heroku or Amazon or Xyz?: We determined that for our usage, Amazon would be more cost effective, and we would have great compatibility with S3 (The Amazon storage engine) and even Amazon Glacier(The Amazon long-term storage engine). Other suggestions for hosting platforms are welcome!

Amazon RDS or our own replicated Postgres instances? RDS means switching to MySQL which feels gross, but may make our lives a whole lot easier. Our ultimate decision was to go with RDS. There are too many moving parts in the Postgres approach that we would waste time solving a problem that didn’t need to be solved.

Deployment: We still don’t have this one quite ironed out. We would like to have a deployment similar to Heroku where a Git Push updates the code on the servers, then manually cycle the instances behind a load balancer to get the code running. Or even have the ec2 instances detect changes and cycle themselves!


We have already learned so much with this project:  A cursory vocabulary of chemotherapy, Knowledge of the oncology workflow, Unix administration, HL7, Django, ‘Meaningful Use’,… tons of stuff. We have also learned squishier things like Project Management, Resource/Requirements gathering and creating positive client relationships.

Now we get to learn even more with cloud hosting for all of our future customers 🙂


PS: If you know of a good Amazon veteran, send them our way!

Tags: , ,

2 Responses to “Fighting the Internet”

  1. Mark Jones Says:

    I realize you have decided on Amazon, but we’ve been running on Linode for quite some time and I think they deserve some acknowledgement. I’d stick with Postgres, (better django migrations) spin up in multiple data centers and let them sync with each other.

    The price is very nice, and so far we are running at 100% uptime in excess of 1 year (almost 2). The only time our system has been down has been for update reboots that we performed.

    I must mention that one time we did lose connectivity from our offices to linode (intermediate outage), but the customer didn’t go thru the same intermediate and never witnessed anything.

  2. Nick Orlowski Says:

    We haven’t rolled anything out yet, so we will definitely take a look at Linode. Thanks for the commendation!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: