The key to mainframe DevOps success is in quickly identifying and removing major bottlenecks in the application delivery lifecycle. Major challenges include collaboration between mainframe and distributed teams, lack of visibility into the impact of software changes, and limited resource flexibility with scaling out necessary testing initiatives. Now let's take a closer look at some of these key challenges and how IT departments can address them ...
In the last few months a half dozen startups in the cloud application migration space have been snapped up by larger vendors and cloud providers (e.g. Cliqr acquired by
Cisco, ITapp acquired by ServiceNow, ElasticBox acquired by CenturyLink). This sudden interest in application migration is perhaps surprising given all the buzz around containers. Aren't containers going to make application migration a problem of the past? Clearly the answer is "No!" Let's examine what makes application migration so important and difficult.
Application migration is increasing in importance for a number of reasons. One key reason is the growth of Hybrid clouds. While almost all enterprises are using the public cloud, almost none of them are using public cloud exclusively. This means that some of their applications are migrating to the cloud. This step also necessitates that these apps be capable of running in more than one public cloud because most large organizations prefer not to create a dependency on a single cloud provider. At the same time, some applications are not moving to the public cloud, which means that they are moving to a private cloud environment. This means that application migration is important to almost all organizations, regardless of where they are migrating to.
At a time when software development groups are trying to move to DevOps, they are faced with developing applications for legacy, as well as these new and different environments. DevOps drivers cause organizations to look for ways to automate the software development, test and deployment processes. Whatever these processes are, they need to work in all of the cloud and datacenter environments that applications are targeted for.
The combination of hybrid cloud and DevOps requirements has led to an increased interest in application migration. The companies that have been acquired shared a common capability for creating an application template that could be deployed in different infrastructure.
A core tenant of application deployment that is agnostic to infrastructure is the need to create an application template that carries meta data about the application's requirements that can be interpreted for deployment to multiple cloud or virtual environments. This template carries information that is not available with containers. Containers do help to make applications portable so that the application image can run in many different cloud environments, but they do not carry the information about which cloud to target, application unique requirements, and post-deployment configuration.
Smarter firms provide application templating as part of its cloud sandboxing software. Sandboxes are used to create a complete replica of a production environment, from network to applications and hybrid cloud configuration. As part of this replica, a template for each application type that allows the application to be deployed in different sandboxes with different virtualization or cloud configurations using the same template should be included. This template should be designed to allow complete flexibility of deployment since it is used in development and testing and is incorporated into an automated DevOps process. In addition, as part of sandboxes, applications are automatically connected to networks so the application template needs to provide enough information to enable network connections to be orchestrated automatically.
These types of application templates follow a modeling approach. The application template includes:
■ A data model for the application that allows the users to describe a structure for the application.
■ A set of attributes that are used to provide additional information about application requirements such as CPU, memory, etc.
■ Deployment path that includes the list of virtualization or cloud environments on which this app can be deployed.
■ The template includes a pointer to the location and type of image for each deployment path.
■ Driver that enables automated control of the app after it has been deployed.
Application templates are key to describing applications in a way that is cloud agnostic and also enables automation for DevOps. The software that creates these templates has become a critical component for many cloud providers and DevOps solutions, whether for development and test, like sandboxing software, or for production deployment.
Joan Wrabetz is CTO of Quali.