According to a recent SolarWinds survey, almost all organizations have migrated at least some infrastructure to the cloud. In fact, just 9% of IT departments have not migrated anything. Furthermore, databases rank in the top three for infrastructure already migrated to the cloud and infrastructure with the highest priority for future migration.
If you’re one of the few who have not migrated any workloads to the cloud, hopefully, this helps you realize that the cloud is in your future. And if you’re one of the many who have already migrated some of your workloads offsite, know that more cloud is in your future (though, surely not all of your infrastructure will find its way to the cloud—hello hybrid IT).
The key question is really how to embrace the cloud—what are the options to transform applications to take advantage of cloud services. Let’s explore the seven ways you can progress down the cloud pathway.
First up—do nothing. Even if you “do nothing,” you will still likely have a hybrid IT environment, either now or in the very near future. The reality is that nearly every organization uses software as a service applications already. In fact, most marketing departments alone use 30 to 70 cloud-based applications, and most of these applications require integrating to data sources hosted on-premises in some way.
Second, most organizations will eventually transform their internal IT departments to look awfully similar to a private cloud—an elastic, software- provisioned and measured service. Some companies will go the virtualization route, while others will implement an actual private cloud platform such as OpenStack.
Third, do your testing and development in the cloud. Test/dev environments are notorious for their infrastructure requirements and ephemeral nature, which make them a natural fit for the cloud.
Fourth, “cloudify” existing applications. This entails moving components of existing applications or database workloads to the cloud. For example, most web applications should be storing images, large files, and videos in the cloud where they can enjoy the benefits of a content delivery network (CDN) and get the load off of web servers. The primary databases may still be hosted on-premises, but you could store backups or archives in the cloud.
Fifth, “lift and shift” your existing applications to the cloud. Once an application has been developed and tested in the cloud, and perhaps its disaster recovery site also hosted, it might make sense to move the application itself to the cloud, too. The term “lift and shift” implies that there are no major architectural changes and that the cloud is being used almost as an extension of hosting. During this process, it’s not uncommon to also cloudify some components of a workload, using a CDN, as explained before, or using a database as a service (DBaaS) instead of a full database instance on a dedicated cloud server.
Sixth, when replacing an application with an altogether new one, consider starting the new application in the cloud. Let’s say an IT department is replacing an internal portal and collaboration tool with the latest version of SharePoint. It would be smart to evaluate how such a workload could run in the cloud. These projects are usually line-of-business applications that also maintain their traditional architecture, aside from perhaps taking advantage of DBaaS or database mirroring.
Finally, develop a cloud-first application. Cloud-first applications are developed specifically to run in the cloud and architected to take advantage of cloud services. They embrace service orientation, APIs, software-provisioned hardware, and built-in redundancy. Everything is software—databases, encryption, mirroring, and backups are a service. They often use “modern” languages and frameworks, such as JavaScript/HTML5, NGINX, Perl, and open source components, including NoSQL databases, such as Cassandra or MongoDB.
The reality is that your IT department will likely end up with a mix of these seven cloud deployment models. To ensure success with any of them, it’s important to develop a cloud/hybrid IT mindset: Build a cloud adoption roadmap based on a workload-by-workload evaluation that considers requirements, potential upside, costs, and urgency. And lastly, implement a monitoring and management toolset that surfaces a “single point of truth” across on-premises and cloud workloads. This normalization of metrics, alerts and other collected data from applications, regardless of their location, will enable a more efficient approach to remediation, troubleshooting, and optimization.