Development Update
0 Comments | published by Scott Szarapka on Monday, April 30th, 2012
I should begin by apologizing for our lack of blog updates over the last few months! We're sorry, LemonStand Now is a huge priority of ours right now and we get so caught up in development that we forget about this blog. I promise we'll be better moving forward.
The last two months have been an exciting time for LemonStand Now, some big changes, more lines of code than you can shake a stick at, and enough testing to make any sane person go slightly crazy. There is one change however, that is bigger than all other combined: we migrated to a new primary provider, Amazon. I could write several blog posts about why we've changed and I may touch on infrastructure planning, requirements and forecasting more if anyone is interested. The short of it is Amazon is a better choice for us to bring you a cloud based e-commerce platform that is fast, secure and reliable.
Thankfully it's been a easy switch as we ensured early on that we built a platform that was as host agnostic as possible. There were some critical changes that have been made, notably the removal of the router from our stack. We gained a huge boost in speed with this change, directly routing requests on our shared platform was something we really wanted early on but wasn't possible, it's now possible on Amazon and we're thrilled with the performance so far. There are countless other changes that we've made with respect to the Amazon migration (and we'll detail them in future blog posts), and we're excited at the possibilities that their wide breadth of services opens up.
We've also introduced a new tool into our stack to further automate management tasks and to allow for better provisioning, Chef. We've deployed the open source version in our pre-production environment and we've seen a solid 10% increase in provisioning efficiency. Chef also allows us to remove some of the overhead on our manager which originally handled provisioning using a custom service and scripts we developed, this change should allow us to reduce the amount of manager slave instances we need to run as we grow. Three cheers for lower operating costs!
Hopefully that gives you some insight into what's been going on here, and if there's specifics that you'd like to know about, please let us know in the comments and we'll try to put together a post covering them!
There is one last thing we'd like to announce, our Development Roadmap is now live, it's updated in real-time so check back often to see what we're currently working on as we get ready to launch the first phase of our beta!
