When many think cloud, they think Amazon AWS. Some even think Microsoft Azure. However, there is a growing trend to want more out of a cloud than those clouds can deliver—well, deliver at a cost people can afford. I am not talking about “white glove” treatment, or any other approach that wins smaller clouds business, but rather the basics. There are certain realities that IT is required to deliver, and unfortunately, with the cloud, those realities do not disappear. For some, the costs of using AWS or Azure are acceptable; for others, they are not. What AWS, specifically, has done is required IT to do things differently. What is different? Why is this an issue? It is different enough that MSPs and other clouds will pick up the slack!
There are several big differences when you use AWS or Azure. They can be summed up as follows:
- High availability is now an application issue!
- Data protection options become more limited
- Monitoring and audit are more limited
- Data management is now paramount
- Costs change from CapEx to OpEx, but not 100%
- Processes for security, audit, and many aspects of the business must change
Any other cloud (CSP) or managed service provider (MSP) that can do those things better will challenge AWS and Azure for seats. It may not get the quantity of seats that AWS and Azure have, but it will be successful if priced appropriately. The last two items in the list above have nearly nothing to do with technology. Those items have a lot to do with people, and people are where the real problems begin.
There are many naysayers, haters, misinterpretaters, and full steam ahead–type folks to make those last two items crash and burn any expansion to the hybrid cloud. The last two may even deny there is a hybrid cloud; while everyone in the office is using Google for something, they might use Amazon for various things and let Azure host their email. There is still the feeling out there that cloud is not in use at specific organizations. I dare them to prove it. Look at the bills, and tell me if there is not even one service in use. That aside, there are many reasons the last two get in the way.
The solution is a consultancy or MSP that can help folks get over this barrier. Many vendors have such consultancies now, as they need them to sell their services. Companies that do this include Dell EMC and Velostrata, among others. The list is nearly endless, if you think about all of the companies that sell services into a cloud. The goal is to transform your environment to be cloud friendly.
Once you get over the people and process issues, other realizations occur. Clouds are actually quite limited, and they need to be. Why? Because no one likes to build snowflakes such that every customer has something different. Instead, every customer gets a piece of what is already there. New services may come out, but new hardware mechanisms, ways of doing high availability, and data protection are limited by nature. The best use of AWS and Azure is to not put anything there you cannot afford to lose, but also to realize that many of the apparent infrastructure elements, such as high availability and data protection, move up the stack. They move up to the application code, not the underlying infrastructure.
This is a crucial realization that some seem not to grasp. If your application is not designed to do HA within the application or to provide data protection within the application or even within the operating system running, you have hamstrung your cloud deployment. You haven’t done it immensely, but it is nevertheless without protection. One way clouds get away with this is by using multiple zones. However, transfer of data and keeping things in sync are not the cloud’s responsibility. They are the application’s responsibility, or even yours.
If they are your responsibility, you also need to ensure that you know where all your data is. Further, you need to know who did what, where, when, and how to your data. You need auditing that is limited within stock clouds. You need to be able to adapt your processes to use new tools, to use new techniques to manage your data better, to get your audit logs, to basically run your business.
It is about process and people, but also about technology. What have been some of your stumbling blocks? How did you solve these issues?