March 3, 2015 | Written by: Javier Barabas
Share this post:
When a company embraces the cloud, it must decide which adoption pattern best fits the workloads it plans to migrate to the cloud. That decision shouldn’t be made lightly—the more informed IT pros are going into the process, the easier it will be.
In this post, I’ll outline how you can find the right cloud adoption pattern for your workloads.
Cloud adoption patterns
An adoption pattern is simply the way a company consumes services from the cloud. There are different boundaries between different adoption patterns, and I will clarify each one here.
Each adoption pattern has its own characteristics, benefits and limitations.
In a traditional (or on premises) way of consuming services or resources, the owner of the infrastructure is responsible for managing every piece of hardware and software he or she uses (the black layers in the above chart). Normally, it takes some time for a user to access a new resource, but it can be configured exactly as needed.
The only limitation a user has when requesting a new resource is the availability and perhaps the quota his department has on the complete infrastructure. Once deployment takes place, the final user has the ability to configure and modify every aspect of the resource he now owns. Depending on the company, central services like patch management, security compliance with corporate rules, software upgrades, monitoring, backup and more can be managed in a centralized way. The costs (capital investments) of the resources are permanent and resources are often sub-utilized.
There are three main off-premises adoption patterns:
• Infrastructure as a service (IaaS). Using this pattern, a company can consume infrastructure from the cloud through a contract with a cloud service provider. The provider can supply any requirement of resources (typically CPU, memory and storage) and the company only has to pay per use.
Compared with the on-premises business model, the IaaS pattern provides a flexible way to acquire resources when needed and those services can be terminated any time. The management responsibility for the company starts with the operating system layer and the provider ensures the availability and reliability of the infrastructure provided. Hardware monitoring is usually done by the provider and is transparent to the client company.
Extended services like patch management and backup can be part of the contract or managed directly by the client company. IaaS services are starting to become a commodity and all cloud service providers are trying to provide services with extended capabilities like PaaS and SaaS (see the following paragraphs). For an example of this cloud adoption pattern, check out IBM SoftLayer.
• Platform as a service (PaaS). Using this adoption pattern, a client company can consume platform services while avoiding the need to manage the underlying infrastructure (see the above chart). The PaaS model is related to the “API Economy.” API means application programming interface and using this model, an application can be designed and deployed just by calling on available services instead of creating new ones.
Applications created using those services pay only for the time and services used , making transparent the management process needed to maintain those services as ready and up-to-date. Common services offered in this way are databases, development runtimes, security processes like single-sign-on and more. PaaS providers are growing fast enough to provide a wide range of services, facilities and development tools to make developers’ lives easier.
Small companies and startups find in this adoption pattern an excellent platform to test and develop applications without an initial investment and a way to “fail fast.” By failing fast, a new application can be tested and evaluated by users, and modified fast enough to avoid the initial implementation costs of on-premises projects. For an example of this cloud adoption pattern, check out IBM Bluemix.
• Software as a service (SaaS). With this cloud adoption pattern, a client company can gain access to an application ready to be used and only pays for the users for whom the service is contracted. The underlying infrastructure and software needed by the specific application will remain transparent and its management can be done by the provider. This business model is very elastic and grows as needed. Compared with an on-premises model, installation and initial configuration are not needed, nor is the hardware to run the application. There is just an onboarding process. Configuration and customization for SaaS applications are minimal compared with the on-premises equivalent. The application consumed as SaaS is always up-to-date and available, and its health is monitored by the provider. For an example of an actual SaaS marketplace, check out the IBM Cloud marketplace.
There is no one size fits all for cloud adoption. Companies should consider their own cost and benefit equation and then decide on the best model. Each application and process needed is a workload and a deep workload assessment is normally performed by companies that have decided to move to the cloud.
Where are you and your company in the cloud adoption process now?