January 16, 2015 | Written by: Manish Gupta
Share this post:
Today, enterprises can have an infrastructure as a service (IaaS) cloud on premises, off premises or both—the combination is known as a hybrid cloud environment. There are many choices for off-premises IaaS cloud choices, such as IBM SoftLayer, Rackspace, Amazon Web Services Elastic Cloud 2 and Google Compute Engine.
An organization’s information technology (IT) department is tasked with making decisions regarding which IaaS service to subscribe to or unsubscribe from. In order to meet business requirements, such as increasing quality and security and reducing cost of operation, an enterprise’s IT department may subscribe to multiple clouds at the same time or may switch between clouds. Enabling seamless mobility of workloads from one cloud to another cloud can spur greater contestability of cloud services.
Technologies are now emerging that can be integrated in cloud broker software to allow migration or mobility of a workload from one cloud to another. Companies like RackWare, CloudVelox, and HotLink are examples of these technologies. In particular, RackWare and CloudVelox allow you to replicate the source servers to the target cloud and keep the target environment in sync with the source environment in terms of operating system (OS), binaries, configurations, application stacks and data.
Let’s discuss in a little bit more detail how mobility can be enabled.
The migration technologies from these companies typically model or abstract the workload and cloud image formats, as well as capture or discover the compute, storage, network and security configurations from the source environment. This allows them to transform from the workload and image formats of the source cloud to that of the target cloud where we want the workload to migrate to.
They also employ cloud application programming interfaces (APIs) to perform discovery and cloning operations and configurations for compute, storage and network. Once the target environment is set up they will typically perform differential snapshots to keep the target environment in sync with source.
One key function that these technologies have to perform is to discover the networking and security configurations at the source environment, including (but not limited to) the IP, domain name system, gateway, subnet, firewall rules and VLANs. This information is then replicated at the target environment. The automation of the entire migration process helps to achieve this in a seamless manner with minimal human intervention.
Once workload migration between pairs of clouds is enabled, service contestability can be as easy as clicking a button on a graphical user interface. An IT department can choose to move from cloud A to cloud B to achieve higher availability or lower cost of operation. Once that decision is made, all the workloads can be moved from A to B using the above migration technologies.
I hope this has been a helpful overview. To discuss further details of how workload mobility can generate greater contestability, please leave a comment below.