Evolution of Cloud Consumption Models
Evolution of Cloud Consumption Models
There is a myriad of Cloud consumption models to choose from today, and the decision is not just about cost or location, but about many other factors. This post delves into some of these factors that help determine the right consumption model for your organization.
Join the DZone community and get the full member experience.Join For Free
Learn how to migrate and modernize stateless applications and run them in a Kubernetes cluster.
Given how new the cloud is, we can’t even say 'This is not your father’s cloud.' The rate of change in cloud technology is moving at such breakneck speed and acceleration that it is becoming almost impossible for many organizations to keep up. The ever-evolving nature of cloud technology has also resulted in significant evolution in the consumption models of cloud. No longer is choosing how you want to consume cloud as simple as choosing between 'Private' and 'Public', like it was in the good not-so-old-days. There is a myriad of options to choose from today, and the decision is not just about cost or location, but about many other factors. This post delves into some of these factors that help determine the right consumption model for your organization.
'Private' vs 'Public'
This is probably the first significant change in how organizations consume the cloud. When Cloud began, there was only the 'Public' cloud—located in the vendor’s data center, managed by the vendor, and multi-tenant. Even today, the general assumption is that when one mentions the word Cloud, one means a Public cloud. As Cloud technology evolved, Cloud management software became available that allowed organizations to stand up a Cloud in their own data centers. This was a self-managed Cloud, located within the organization's firewall, in their own datacenter—on-premises (on-prem), running on their own hardware. It was, of course, single tenant. 'Private' was always thought of as 'on-prem'. These became the only two choices available. A decision on which one to choose was based on many factors—cost, data location, single or multi-tenant, ability to support and self-manage an in-house cloud, etc.
Enter 'Dedicated' and 'Local'
The latest iteration of consumption models has disrupted the very concept of 'Private', as being just 'on-prem.' Cloud technology has evolved today to a state where now Cloud vendors are able to provide single-tenant, managed cloud in their data centers. At IBM, we refer to this model as 'Dedicated' Cloud. So, now we have a Cloud which is dedicated to your organization—vendor managed, in a vendor's Datacenter (off-prem), on the vendor's hardware, but single-tenant. It is hence, a 'Private' Cloud, but located off-prem, and vendor managed. Think of 'Dedicated' as 'Private, managed, off-prem Cloud'-as-a-Service.
The next evolution of how organizations can consume cloud is even more disruptive. Organizations have always wanted the luxury and comfort of having the cloud be on-prem. (Note, I don’t say security. Whether an organization’s datacenter is always more secure than a Cloud Service Providers is debatable.) With an on-prem cloud, there are no issues with where the client data is. No issues with managing compliance-related to running critical applications in someone else’s datacenter. No issues with having to 'tunnel' into another network to access applications and infrastructure. However, the challenge with having an on-prem cloud has always been the skills needed and risk associated with self-managing the cloud. We are talking about having IT organizations that have excelled in managing hardware infrastructure, now manage the cloud and the constantly evolving technology stacks that needed to be managed and maintained. We are talking of running and managing an efficient cloud with all the relevant Cloud services with their own SLAs, that are required to have an on-prem cloud deliver the promise of a Public Cloud.
Such a consumption model delivering the best of both worlds is now available by what is known as a 'Local' Cloud. This is a cloud that is in a client’s datacenter (on-prem); it is, however, managed by the vendor who is delivering the Cloud; and, it is, of course, single-tenant. In such a model, the organization consuming the Cloud continues to just manage the hardware infrastructure on which the Cloud is delivered. (IBM also offers an option to bring its hardware into the organization’s datacenter.) The Cloud itself is managed by the vendor. This is achieved by a 'tether' that allows the vendor to remotely initially deploy the Cloud itself and continue to periodically monitor and update the Cloud software stack, as and when needed. (At IBM we call our tether technology 'Relay.') The client organization coordinates the timeframe when updates will be done, allowing them to control outage windows, if any. It is hence, a vendor managed 'Private' Cloud, which is, however, on-prem. Think of 'Local' as being 'Private, Managed on-prem Cloud'-as-a-Service. IBM offers 'Local' Cloud for its Blue Box OpenStack IaaS Cloud and Bluemix PaaS Cloud.
Choosing the Right Consumption Model
The decision of choosing the right consumption model is not trivial. Consuming Public Cloud is by definition 'non-sticky.' It is (relatively) easy-on/easy-off from a public cloud. 'Dedicated' is also the same way, as you are essentially leveraging a 'Public' cloud, that has now been 'dedicated' to you. You are renting a stand-alone building, not an apartment in a multi-tenant building. 'Local' on the other hand is different. Understanding the division of duties between your organization and the vendor are critical. Understanding the change management processes is essential. You are forming a partnership with the vendor that is more complex, and in turn, much more valuable than using a Public cloud.
The first decision point to be made is the most critical. Are you looking for a Managed Cloud or will you self-manage? If the latter, you need a traditional, on-prem 'Private' Cloud. You buy the Cloud Platform technology from a vendor, train your people, prep your infrastructure, and go to town. If the former, then you can move on to the on-prem versus off-prem decision. This decision of Managed vs self-managed Cloud has to be made first. It impacts not just which Cloud you pick, but also impacts your IT staffing. If you are challenged by your ability to manage a Cloud with your own IT staff, due to skills or headcount; or are unsure of your ability and desire to be able to deliver the services and associated SLA’s your clients are expecting from a Cloud, then you should consider a Managed Cloud. This decision was a not an option until now as all Managed Cloud options required an off-prem Cloud. With the advent of IBM’s Local Cloud offerings, both for IaaS and PaaS, that is no longer the case. You can get an on-prem Cloud, as a managed service. Have your cake and eat it too, and let IBM do the baking.
The rest of the set of decisions remain the same—IaaS vs PaaS? Which Cloud Services will you need? What are your scalability and elasticity needs? What Cloud platform standards are you looking at? Are you concerned about vendor lock-in? Do you want to have a Hybrid Cloud that extends across on-prem and off-prem platforms? More on those decisions in future blog posts. To know more about IBM Bluemix click here.
This article was originally written by Sanjeev Sharma.
Opinions expressed by DZone contributors are their own.