glen

In my 30 years of IT and counting, I have seen a shift in companies wanting to migrate their on premise servers to AWS or Azure. The driving factor appears to be saving costs in the long run and not having to worry about the underlying hardware that hosts the services. But little do they know that without proper management of the resources deployed and a deep understanding of the services provided, things can get out of control quite quickly. In fact, you will actually start spending more than you initially anticipated. With on premise, the detriment of going out of control is less impactful because the ability to scale up is fixed. While with Cloud, there is no limit for the Administrators who have no knowledge of the costs associated with each services being utilized. In AWS for example, a simple pull of information via it's API like getting instance details will incur a cost. Another example is enabling an Elastic IP without associating it to an instance will incur a 10USD per hour cost. AWS has since reduced this to a relatively insignificant amount. However it is still worth noting. There are also many default settings in AWS that incur costs without your knowing. You have to go in and turn them off or disable them during the wizard creation. As for not having to worry about hardware, yes that is true but how often do you need to upgrade your hardware for on premise? I hear this every 5 years from management going around. Not sure who created that standard but nowadays with SSD storage there are hardly any moving parts except maybe for the cooling fans that need replacing every 10 years. Or if you have a dusty environment, do some vacuuming on them and they are as good as new. Hardware failures that impact the company is usually the result of negligence.

This blog isn't about me bashing Cloud services. I still believe they are a good alternative if you don't have the technical know how or don't want to host an on premise infrastructure but you have to really keep an eye on the costs and start using it in a way unlike you've been accustomed to while your servers were on premise.

In summary, I feel going PAAS, SAAS or IAAS is doable but only if you keep a close watch on the costs and review your monthly bills to make sure you are paying for what you subscribed to. Hire admins who not only know how to deploy services correctly but also know how to keep costs to the bare minimum without sacrificing performance. This applies to both Cloud and on premise. As for Disaster Recovery plan, I find this is just an exercise that will never be implemented. You only need your data backed up somewhere that you can rebuild from should it come to that situation. However, if there was actually a real natural disaster going on, the last thing in my mind is about failing over to the DR site.

Peace! mofos.