Aaron Delp, Director of Technology Solutions, SolidFire
The last four years have seen an amazing amount of change in our industry. We are in the middle of an architecture shift of IT Operations not seen since the days of the transition from mainframe to client/server computing. Just as the mainframe to client/server shift required both new software applications and hardware infrastructure, so does this shift, commonly referred to as Bimodal IT.
Bimodal IT is the concept of the operations of two distinct environments based on entirely different applications, underlying architectures and operation models; much like operating a mainframe and client/server environments simultaneously. In this case one environment runs traditional virtualization applications in a highly available and redundant infrastructure. The other half of Bimodal IT is agile micro-service applications loosely coupled together in an environment that does not require high availability or redundancy from the underlying infrastructure.
Just as your organization would not take a mainframe application and easily convert it to run in a client/server environment, the same thinking applies to Bimodal IT as well. Traditional virtualized applications are not designed to run on an agile infrastructure and doing so would often lead to disastrous results.
Five disruptive trends have developed as a result of Bimodal IT that will change the way we both design and operate IT moving forward:
1. Open Source Software—While always a viable alternative in the past, Open Source Software (OSS) has become the keystone to agile IT based workloads. OSS allows for easier and open collaboration and development between vendors and customers. The capitalized expense of closed source software is moving to an operating expense of staff qualified to design and operate Open Source Software.
2. Public Cloud—The arrival of mature public cloud offerings has radically changed the pace at which IT operations/development plan and fulfill critical business needs. Public cloud allows for quick and effortless procurement of infrastructure hardware while also changing infrastructure from a capitalized expense to a pay as you go metered operating expense. In addition to this decreased deployment time, new operating methods become available such as rapid development and short term testing of applications at massive scale. Early detractors of public cloud often mention security as a concern, but this problem has been solved through the development and maturity of Virtual Private Clouds (VPCs).
3. Linux Containers—Linux Containers have existed for years but only recently have received attention as the agile IT model has matured. Prior to containers, each application component required an individual operating system per virtual machine. We had a one-to-one relationship of application components to operating systems. Containers allow us to isolate the application from the operating system allowing for a one-to-many relationship of application components to operating systems. By stacking multiple application containers onto one operating system, we achieve an operating benefit of fewer systems to maintain, as well as faster recovery time of a single applicationcontainer vs. recovery of an application and the operating system in the event of maintenance or a failure scenario.
4. DevOps—Historically, IT has consisted of two distinct departments with different responsibilities: Development and Operations. Development creates and modifies software to business requirements and Operations is responsible for the installation and maintenance of all facets of the data center. These exclusive silos have led to extended time for deployments and sluggishness of a business to respond to the ever-increasing rate of change and agility required to stay competitive. DevOps is the concept of moving away from these traditional silos and embracing a new delivery philosophy that decreases friction in the deployment and execution of both development and operations. By increasing efficiency we lower overall operating costs as well as increase the ability to serve line of business needs more quickly.
5. Shifting IT Skillsets—The previous four trends culminate in the agile micro-services model. This model requires a new skillset in Open Source and DevOps that must be developed in-house to form the transition from exclusively a traditional virtualized application model into Bimodal IT. This model of operating both traditional virtualized and agile micro-service applications must be maintained for the foreseeable future and presents a new and unique management challenge to today’s IT leadership. New talent must be developed and fostered to combine the above trends into a new consumption model that is flexible and dynamic. Skills to look for include DevOps workflows experience including the concepts of Continuous Integration & Delivery (CI/CD). Additionally, a background in Linux and Open Source software tools and the ability to automate and manipulate infrastructure through programming interfaces will be requirements to building an agile micro-services based model.
The next generation of applications will change the way we approach IT operations. We will shift from a capitalization-focused model of software and infrastructure procurement into a more agile model of software and infrastructure that will evolve and change rapidly to maintain pace with the market. Long gone will be the days of three to seven year procurement cycles of capitalized expenses. To match this change in operations of assets, both the skill sets as well as the workflow models to develop and maintain critical line of business functions must also change to meet this new demand. It is a noteworthy time to be in IT operations, and I look forward to seeing the efficiency and productivity gains offered by this new model.