The New Business of IT Infrastructure
by Steve McDonald
These days, software, websites and mobile devices are driving business plans. That’s a big shift from days of old when IT was a background service supporting internal productivity. Business leaders are starting to talk about infrastructure in terms of how it supports revenue streams. On-demand cloud models are helping business leaders be more agile and respond to marketplace trends when the time is right. That’s exciting, but how can CIOs design, monitor and measure this new business-oriented infrastructure?
One struggle is that IT organizations haven’t been structured to view and monitor infrastructure for business value. Moving away from the utility-based infrastructure model, when the primary purpose of managing servers, storage and networking was to make sure computers were up and running, requires cultural shifts, different skills and new toolsets. It was only a few years ago when CIOs began to make real inroads into lowering the costs of infrastructure management, made possible by virtualization and automation. The next step is all about delivering business results from IT infrastructure. Here are some guidelines for this transformation:
- The Silicon Valley mindset: The app economy is now a global phenomenon, not just one confined to tech centers like Silicon Valley, Atlanta and Seattle. Consumer brands like Nike and Coca-Cola have realized that to maintain profits, they must move beyond products and traditional customer relationships. Senior executives are being tasked to deliver technology-based services that bring new value to customers. For Nike, that was the entry into wearables; for Coca-Cola, it was the launch of Freestyle, an engaging new soda fountain experience enabling customized drinks along with complementary mobile apps. Progressive CIOs will determine how to shift the role of IT from utility provider to broker of innovative services for employees and customers.
- Delivering services not technology: Intrinsic to this new role is the need to guide business units to the best resources for executing customer-facing initiatives. Increasingly, that means evaluating, selecting and governing the use of third-party service providers such as co-location, private cloud, public cloud and managed services. IT must be responsible for overseeing providers for security, uptime and reliability, delivering hybrid IT through integrating on-premise and off-premise applications and data and providing the governance framework for everything. This service-oriented business model must be able to adapt quickly to changing expectations of end-users while consistently delivering speed, performance and innovation.
- The speed challenge: Never has time-to-market been more imperative, with upstart competitors not just appearing in your own region, but around the world. Traditional IT organizations are struggling to keep pace. First there is the need to move from Waterfall development to Agile and DevOps practices. That entails frequent releases, frequent testing, continual feedback and high levels of automation. It also requires collaboration in ways that IT employees have never had to do and may not be excited about, if they are independent workers by nature. Yet there is no other way to be more responsive to the business and adapt to changing workplace conditions. A related challenge is that enterprise toolsets haven’t kept up with changes in development and IT operations practices. Fragmentation is common, with small groups of employees operating independently and using their own tools. Bringing people and technologies together into a cohesive, standardized framework is imperative. It may require hiring new talent, if they can be found. The Weather Company exemplifies the rapid development framework. As a service provider that delivers timely data to thousands of websites and apps, the company’s engineers are continually recognized for leadership in the development of new techniques while actively participating in open source communities. They are a data centric organization and have a strong DevOps team. Rapid iteration and standardization is how other organizations can also get ahead.
- Orchestration is the name of the game: Deploying more complex applications in a hybrid cloud world is more complicated. An application infrastructure today is not just one virtual machine but four or five, and those machines need to be deployed in specific service tiers that are template-driven not manufactured on demand. The old way was to take weeks to manufacture lots of VMs. An IT manager needs to need to click once and deploy the entire host of virtual machines in a template driven manner with predefined SLAs. The business requires the orchestration of all these resources.
- New measurements: When moving to a business-driven infrastructure model, an important step is changing metrics. It’s no longer enough to monitor server utilization and uptime. Here’s what else to consider:
- First, companies should begin to separate the infrastructure that runs the business from the infrastructure that runs new initiatives. This requires highly standardized environments in which utilization can be clearly delineated between utility services and revenue-generating workloads. Then, you need to associate revenue to those systems, not just show IT performance.
- The concept of chargebacks is also important. While not a new idea, most organizations struggle to shift away from paying a fixed cost for a level of service. Cloud IaaS and SaaS apps make chargeback more palatable, but most companies will have to evolve to this level of detail and shared cost. An IT leader that is trying to move from a dictator to a consultant who guides the business on the best combination of internal and external systems needs to show what the business is consuming in terms of costs.
- Another metric is speed. How quickly can a company move from idea to value creation, supported by the infrastructure? Older infrastructure models are limited in speed so it does sometimes require investing in new infrastructure technology.
- Finally there’s quality of service. This is no longer measured by the infrastructure as a whole. The business unit manager doesn’t care about that but about the specific applications in her area. IT must be able to show the quality of service and SLAs for that application. This is an emerging space, which will be supported by better tools down the road.
Steve McDonald is the vice president of product at StrataCloud.