BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Bimodal IT: A New Buzzword For Old Concepts Presents Teachable Moment

This article is more than 8 years old.

IT analysts are like fashion designers, always searching for something new, even if that means recycling age-old concepts in new terminology. But like successful fashionistas, IT leaders need to stay on top of the latest trends if only to be prepared when the CEO comes back from an industry conference full of questions. One of the latest buzzwords circulating the industry is Bimodal IT: another Gartner creation like Hype Cycle and Magic Quadrant. As Gartner defines it, Bimodal IT is an organizational model that segments services into two categories based on application requirements, maturity and criticality. "Mode 1 is traditional, emphasizing scalability, efficiency, safety and accuracy. Mode 2 is nonsequential, emphasizing agility and speed." Seems logical enough and hardly controversial, but also not new. Yet it offers a lesson in how cloud services and engineering practices can improve enterprise IT.

Whether you call it legacy versus emergent systems, Brownfield versus Greenfield deployments or sustaining versus disruptive technologies, the dichotomy between old and new or maintenance and development has been around since the dawn of IT. Each category has always required a different set of investment, management and governance techniques. The difference now is the pace at which new products are developed and refined and a concomitant decrease in useful half-life of mature services.

The bimodal concept is deeply entwined with the DevOps and Agile Development philosophies revolutionizing IT. What started in the engineering culture of Bay Area startups, the notion of coupling rapid, evolutionary application and service development with IT operations and maintenance to deliver products on a rapid release cycle, is resonating with vendors and enterprises. Look no further than the thousands of attendees at the recent Cloud Expo/DevOps Summit which validates Gartner's estimate that a quarter of large enterprise IT shops will be using DevOps by next year.

There's a logical rationale for the bimodal decoupling: legacy systems are responsible for an organization's most critical business processes and sensitive data. Reliability, availability, serviceability (RAS) and security are the most valued attributes. These systems live by the maxim: "if it ain't broke, don't fix it" You don't want to jeopardize order processing and accounts payable on some risky new software approach.

In contrast, it's nearly impossible to experiment with innovative application development tools and cloud services on Brownfield environments: the IT equivalent of spaghetti code, a byzantine tangle of systems, software and support personnel. Instead, the bimodal approach identifies areas where IT has less to lose, but large potential gains by using iterative development, new software tools and public cloud services. Many times, even the business model for an emergent IT service isn't well understood and subject to rapid, market- and user-driven change. Here, it's far better to have development and IT operations processes with the smallest possible overhead. While the tortoise wins the slow, steady marathons, the hare is far better on an unpredictable course full of twists and turns.

Source: Wikipedia

But, the bimodal dichotomy is more than just an operations versus innovation distinction. A chief way software achieves agility is through modularity: cloud-scale, Mode 2 services are much less monolithic than traditional Mode 1 enterprise software. Without decomposing software functions and business processes into the smallest logical units, so-called microservices, it's virtually impossible to both rapidly iterate features and scale capacity. Modularity also well suits cloud deployment, where common, standardized pieces of the overall system are best run as SaaS, while other components use a mix of IaaS products.

One dilemma bimodal adopters will eventually face is the eventual transition of successful disruptive (Mode 2) projects into core enterprise processes and services. One approach that's tailored to the world of virtualized cloud services is so-called blue/green deployment, aka continuous delivery. It's a variant of the common partitioning of application code into stable, beta and development branches in which you run a next-generation version of the service (the green environment) in parallel with the stable release (blue environment). Using virtual resources allows live testing by quickly shifting traffic from the stable environment to the development branch, and back again should problems arise. This blog from an AWS startup highlights the benefits:

If the validation tests of the green environment fail before you start sending real traffic, then you can dispose of the environment without ever having affected the live blue production environment. AWS on-demand resources allow you to stop paying for the failed green environment resources and simply release those resources.

If you start switching a small proportion of traffic from blue to green and encounter serious issues, then you can quickly switch back to blue and restore service within minutes. This process is sometimes called canary analysis: you test that the new deployment works in a real-world scenario with a small set of your users.

The beauty of using cloud services is their ethereal nature, where arbitrarily complex infrastructure can be easily created, tweaked, scaled and disposed.

Bimodal IT may be old wine in new bottles, but it draws needed attention to divergent service and development requirements for different phases of the IT lifecycle and highlights how cloud services don't just disrupt IT operations, but when used creatively, can enable highly adaptive services and efficient processes.