http://feedly.com/e/RYDwMIkk
At the end of January we got a glimpse into yet another organization that was struggling with the financial burden of running entirely in the public cloud. Sarah Bird, the CEO of Moz, posted a year in review which highlighted a the significant financial impact that moving from AWS to a company-run private environment had on their financial picture. The short version: they are saving millions of dollars by diversifying how they acquire infrastructure. This isn’t the first time, and certainly won’t be the last time that we see this. (More coverage on SiliconANGLE and Geekwire)
At today’s prices, the value of public cloud providers like Amazon is primarily about the infrastructure agility and low barrier of entry offered by their services. For a few cents (or dollars) an hour, you can get massive amounts of compute power, advanced distributed services like DynamoDB, and highly automated systems like RDS in minutes with a simple credit card transaction. As far back as 2011, analysts were talking about the importance of this agility being under represented in the financial analysis of moving to public clouds.
The problem is that costs can add up, and add up quickly. Pretty soon, it starts looking much more financially attractive to bring some of your infrastructure in house (or perhaps even to another cloud provider).
So what exactly is one to do? Public clouds provide faster time to market, easy access to powerful technologies being offered “as a service”, and reduced risk from direct capital investments, but these benefits come at a premium (and the costs can add up fast). This is big dilemma for organizations.
The answer to this vexing dilemma is to start with the goal of infrastructure diversity in mind. Take advantage of public cloud services early on, but be prepared to make a move. Design your application to be infrastructure agnostic and easily deployable. If you are going to take advantage of more than just infrastructure from your provider (e.g.: RDS), have an alternative operational model in mind for other infrastructure environments.
Certainly data gravity will come into play with any actual move, but that can be overcome with specific design decisions that take the weight of your data into account from the start. You don’t *have* to move your data around, but be prepared for what it will take to make the move when you realize you need to.
CumuLogic’s platform may be part of the answer for your organization. Our software gives users many of the same services offered by the largest public clouds, but with the added benefit of being able to pick your infrastructure target. Want to push a database, queue broker or entire application onto AWS? No problem. Want to push that same application into a private cloud you have just stood up? No problem their either. In fact, application teams can use the CumuLogic controller to deploy to different cloud targets based on their organization’s specific financial and service level requirements.
-chip
--
via my feedly.com reader
No comments:
Post a Comment