Maplewave's Journey to the Cloud - Part 2

bank of computer servers in the sky


This is our second post in the series that documents Maplewave’s journey to transition our IT infrastructure to the public cloud. You can find part 1 here.


Ok, we know where we want to go.

In my last blog post, I chronicled our journey from server-based hosting to a virtualized co-location environment. We have the benefits of a quasi-private cloud, but we still own all the hardware and (all the headaches) that this entails. We do all the firmware updates, all the hardware refresh work – and we pay for power and cooling.

So, intuitively we believe that it makes sense for us to move to an IaaS or public cloud environment, but how do prove it with numbers?

And if it is the right decision financially, what are the other benefits that we expect to help drive our company forward? What are the next steps to getting us closer to our goal?

Budgeting for the Cloud

To begin with, we need to take a much closer look at what our hard costs are for our current co-location environment.

We are lucky in that we have done a good job in the last few years of operationalizing our large capital expenditures, so we have a very good idea of our spend in that area on an annual, or even monthly basis.

We also know that we have a server refresh that would need to happen in one of our datacenters in the next 12-24 months. As we are using co-location services, we know what we are paying for rack space, power and network (Internet) connectivity across our data centers.

About the only other things we need to consider will be licensing and support for the various OS, OS tools and management tools that we use to keep things running on a day to day basis – with our biggest cost being for the hypervisor and hypervisor management.

The Bottom Line

The challenge lies with determining what cloud services we’ll be taking advantage of – and how much they will cost us. To keep it simple, we have decided to just start with standard compute resources, along with managed databases. We will only look at moving customer inbound workloads to the public cloud until our existing hardware needs to be refreshed. We know that we’ll be adding services as we go, so we’ll need to be diligent with our cost management process to evaluate this on an ongoing basis.

Based on our calculations and supporting just what we have today, we expect to have significant cost savings within 18 months as we move resources from our co-location facilities to the public cloud.

So just from a pure financial model, we will be able to optimize our infrastructure expenditures within a couple of years, while at the same time, adding capabilities around scalability, redundancy and <gasp> even security.

Next Time

In my next post, I’ll explain more about the added advantages that cloud technologies can bring, as well as outline our guiding principles as we start to design our environments.

You Might Also Like.