Simplifying design, satisfying the business and putting assets to work – this is Modern IT.
There are many systems in data centers needed for critical functions, but some appear to be doing little or nothing a lot of the time. Three of these that tend to have very unpredictable workloads are the development, testing and disaster recovery environments.
What if we used ITaaS techniques to satisfy the demand for these three workloads with the same, shared infrastructure with a smaller subset of the resources, managed in a different way than that of traditional standalone environments?
The development and testing environments are now starting to be managed under the auspices of DevOps. Development teams have become one of IT’s most demanding clients as their success and ability to deliver tested applications has a direct impact on overall business health. To satisfy this, most companies (with or without IT knowledge) are leveraging cloud resources to augment infrastructure provided by “traditional” IT departments in order to get on with their work. On the other side of the IT workload spectrum, a less dynamic environment lies in wait for disaster recovery (DR) – an event that may never happen.
Although the mission and the role of these systems are very different, the stakeholders of each of them would tend to agree that if there were standard practices to build, consume and manage these systems, it would be advantageous in many ways. Having standard approaches to development, testing and DR can allow us to reap the benefits of economies of scale with greater quality, predictable performance, increased flexibility and lower overall costs.
As we define our overall ITaaS transformation strategy, design of infrastructure, orchestration, management, workflow and processes are key to the successful delivery of effective systems in modern IT.
One of the most overlooked features of effective orchestration platforms is how to define, track and reassign attributes of the infrastructure for use in the repurposed environment. It is exactly this functionality that would enable a DR site to act as a standardized sand box with a self-service portal and a clearly defined, repeatable, change-control process for IT to run, and at the same time, enable fully functional DR site strategy with aggressive RPO/RTO objectives to co-exist.
Typical workflow for enabling service platform consists of a number of tasks usually identified by one of four common IT sub-structures: networking team, compute/ server team, storage team and virtualization team. What is apparent in these workflows is that most of the time the tasks are performed in sequence, with manual notifications that a particular step has been completed and the next can be initiated, even though many are clearly identified within the same group. Development of these workflows is a key step in enabling stateless compute infrastructure to address DR and auxiliary needs, as these repeatable tasks can be translated into automation/orchestration objects to be used in the orchestration platform. It must be re-emphasized that the orchestration platform needs to be integrated with the change control process to ensure that the current state is managed and the DR environment can be recreated as needed.
Utilizing automation and orchestration tools to manage stateless compute resources, a converged network platform, SAN and virtualized workload/applications, we can simplify DR site design while enabling ITaaS transformation to take roots within IT organizations.
The concepts outlined in this article drive value to the business in different ways … doing so in ways that exploit technology innovation that were not practical or achievable in the past, but are now quite attainable with the current ITaaS enablers.