February 5, 2015 | Written by: Bruno Diegues
Share this post:
Cloud computing has radically changed the way information technology (IT) works, enabling a smarter business and a new infrastructure delivery model.
Organizations willing to run their resources on the cloud will eventually have to decide whether they want to use a managed or unmanaged infrastructure as a service (IaaS) delivery model. Each company’s choice depends on the requirements of the workloads they are planning to move to the cloud.
Managed-cloud service providers include a services layer on top of their infrastructure. IBM Cloud Managed Services, for example, provides virtual machine instances with services through the operating system (OS) layer and, optionally, to the middleware and database layers. The governance model is delivered on Information Technology Infrastructure Library (ITIL) methodology that includes change management process, patch management automation, and health checking (up to the OS layer).
This solution also includes enterprise-level monitoring tools, such as IBM Tivoli Monitoring, which can be integrated with the existing service management tooling system. Backup and restore activities are generally managed in this model, with the option for database backups.
In the unmanaged delivery model, infrastructure resources are provided to cloud consumers and availability levels do not go beyond that layer. IBM SoftLayer uses this as its standard delivery model. Provisioning support is provided through the service catalog portal. Organizations and business units can be self-sufficient, ordering the compute, network and storage resources they require.
With an unmanaged cloud, the cloud consumer has administrative access to manage the OS directly or use management services from a third-party provider. It is important to note, in this model the cloud IaaS provider is only responsible for the infrastructure availability, and OS, middleware or database management is not included.
Organizations tend to use the unmanaged model for development and test environments, especially for creating new applications that are able to scale resources up and down, such as virtual CPU, memory and storage, through the use of application programming interfaces (APIs).
In my opinion, the managed cloud delivery model is the best fit for mission-critical workloads, such as production systems and systems of record. These workloads in a cloud-enabled environment usually require scalable components in a virtualized infrastructure, along with automated lifecycle features.
Unmanaged cloud is the best fit for cloud native environments, also known as “born on the cloud,” which demand elastic infrastructure in a standardized infrastructure, such as systems of engagement or development and test environments.
What do you think about managed and unmanaged cloud delivery models? Share your ideas with me below!