Organizations are migrating to the cloud in droves these days. The leaders of the pack are investing in an open and public cloud model. Others are slowly moving towards that by first testing the waters in a private cloud. This is fine as long as the vision remains clear and the private cloud model is viewed as a way to get to the open and public model. Realistically, a three-tier cloud model is a foregone conclusion, a commoditized element that all organizations have in their future.
The migration to cloud computing is complicated, to say the least. Besides the “tip of the spear” issues of capital expenditures (including amortization of the assets) versus operating expenditures (which will make anyone’s head spin), there are many other issues to consider. The following are the top seven issues to consider when moving to the cloud – in no particular order – with a simple explanation of why each matters (or doesn’t) in a cloud computing architecture.
- Architecture: A cloud computing architecture has to have three layers running on top of a public network, or else the benefits are lost. The three layers — infrastructure, platform, and software — roughly correspond to the main functions they perform: storage and connectivity, logic and processing, and presentation. If your provider or vendor offers anything other than three independent layers that can be interconnected or even replaced, you are not getting cloud computing components but most likely legacy SaaS or hosted components.
- Security: Much has been said about security in an open network, but not all of it is true. There is plenty of research that shows that cloud-based security is as secure as (if not more secure than) anything that can be done on-premises. If you hire a security provider, or your cloud provider does the security, you are getting the latest and greatest implementations of security managed by a team of experts that is bigger and more experienced that anything you could hire in-house. Further, they are invested in keeping things secure, versus being one of the functions your team performs. There is no matching the investment cloud providers can make on security while they distribute the costs.
- Software: This is probably the biggest problem you will face as you migrate into cloud computing. Vendors who have enterprise licenses already implemented have a hard time letting go of the security of knowing the licenses and maintenance fees will continue to roll in as long as the contract is renewed. In cloud computing, there are no certainties – it’s far easier to rip and replace if the vendor does not do a good job. The cloud computing architecture is based on the premise of replacing and extending by hiring one or more vendors that do the same function. As such, the fees are not paid for access to the application, but for usage; the model is based on value delivered. Planning, and your vendor, must conform to new models.
- Hardware: As mentioned above, amortization and customs of doing things a certain way play a major role in preventing, or rather delaying, migration to a cloud computing architecture. The hardware that has been purchased, maintained, optimized, and used for many years is now no longer under the control of the same people; cloud providers own it and rent it out as necessary. This changes a large number of traditions in the organization, and the people behind them are not going to be happy. Anticipate and plan for change management around them and the vendors behind the hardware.
- Integration: Organizations coming to the cloud from any other computing platform often have a hard time grasping the issue of integration. Integration in an on-premise, hosted or legacy SaaS application demands point-to-point work. You have to know where you are coming from and where you are going. You have to build an exclusive “pipe” between the two places, select the data you want to pass along, and then program it and hope nothing changes. In the cloud, it’s very different, since the metadata and services allow a platform to serve anyone who asks for the right information with the right security attached. If the person is authorized, the service runs and provides a result. If the data passed changes, the metadata handles that. If the call changes, the metadata handles that. In addition, a secure platform acts as an access point to any other secure and trusted platform, eliminating the need to do yet another integration point. This makes everything far easier and more powerful.
- Control: The control that your organization and the people in it have over the existing on-premises assets is going away. In spite of several industries and even some countries still holding on to the “location of the data” tales, the day is near where, as one CIO recently told me, “I don’t know where my data is, nor do I care. I know I can use it whenever I need it – that’s all I need to know.” It is not only the data, but the hardware, software configurations, management and maintenance of the assets, as well as the direct influence on how things get done. When a cloud provider delivers solutions to thousands of clients concurrently, there is not a lot of room for the users to derail the process. While we are starting to see more configuration and management options emerge via metadata, the amount of control that exists in the cloud is very different from on-premises. It isn’t worse, since it delivers more value. It’s just different.
- Business Transformation: This is one of the biggest items to consider when moving to the cloud. For digital transformation to occur, which is the next stage for business transformation in the upcoming decade, the cloud must be a commoditized, sunk-in investment that is considered to already exist and be implemented. Virtually none of the management and allocation of the data, knowledge, and content necessary for digital transformation to occur can be moved around without use of the cloud as the underlying infrastructure. Combined with the issues mentioned previously and all the benefits of cloud computing adoption, these things will make digital transformation happen.
Of course, each implementation is different, and these issues as well as others will vary from organization to organization. However, the past ten or more years of implementing cloud computing architectures have taught me these valuable lessons.
Do you have some others to add? Let me know in the comments!
disclaimer: Acumatica is a client, and this year the contributed to my research time exploring cloud. I cross-posted this on their site and they are also using the content in other manners – part of my sponsored research model. I do my research, I get paid to share some of the results, but I own the agenda, editorial rights, and distribution rights as well as intellectual property. I appreciate their support of my vices (food for my kids, a place to sleep) and hope to continue working with them on research on cloud topics. The thoughts and concepts in this post are totally mine.