Building modern applications takes more than just moving to the cloud

Cloud computing has ushered in a new era for IT leaders, who have never been in a better position to deliver value directly to their customers. In the recent past, IT departments were widely considered to play a supporting role to "The Business." But that is no longer the case – technology is rapidly becoming the core of a company’s growth strategy. IT leaders suddenly have a new opportunity to lead huge transformations to drive massive business success, and most of those transformations are happening in the cloud.   

Building on a modern cloud platform – be it public, private, or hybrid – enables organizations of all sizes to deliver faster services internally to their organization and externally to their customers. Methodical capacity planning and manual infrastructure upgrades give way to auto-scaling and vendor-maintained infrastructure that free development and operations teams to focus on delivering on business objectives. Rigid vendor suites give way to modular, custom-tailored technology stacks built from best-of-breed hosted cloud services and open source software components. And slow, manual, error-prone development processes give way to automated build-test-deploy cycles enabled by an API-centric cloud universe.   

The cloud is not just a platform for building a better technological mousetrap - it is also where major shifts in culture and process are taking place. Cloud platforms are fertile ground for developing culture and agile development practices. Software-defined infrastructure and a platform that enables self-serve access empowers teams to experiment with new technologies, to adapt quickly to changing market conditions or customer preferences, and to deliver the multi-platform, always-on services that users have come to expect. 

With all the technological, cultural, and process changes sparked by the shift toward the cloud, many of the old ways of managing, monitoring, and maintaining IT projects no longer make sense. To harness all the benefits of a modern cloud infrastructure, it's necessary to adopt new monitoring and management practices to ensure that your infrastructure can keep pace with changing demand, that every component in a heterogeneous cloud stack is performing as expected, and that costs of cloud services don't spiral out of control.      

The great leveler 

For firms that have moved to the cloud, IT leaders’ next steps should be establishing measurements to track how effectively this shift has worked to impact the business. Moving to the cloud has become a catchphrase, but switching platforms is not the only transformation that needs to happen. Indeed, it is fully possible to engage in a “lift and shift”, where an application’s components are transferred from on-premise hardware to cloud-based servers, and the tools and processes used for the application remain the same. As a result, the advanced capabilities, and capacity to improve product delivery do not come to pass. Even if changes have been made, an IT leader should be assessing what else is possible to continue improvements – changes that allow for cycle times to shrink and more projects to take off can make a world of difference for the ultimate output of a company, and once a change is in place, this accelerated pace begins to compound improvements to a company’s offerings. 

Any IT leader should have the goal of putting their company in the position to quickly become the market leader in their category vis a vis their software capabilities. Because of the “leveling” force that the cloud provides, if a firm isn’t doing its utmost to move as fast as is physically possible, a company should constantly assume that other players, whether they be established competitors or upstarts that were attempting to change a business model, will build the capability to vastly accelerate go-to-market as a competitive advantage. The IT leader that rests on their laurels puts themselves in the position to be overtaken. While this was always true in business, the speed with which a company can be outrun by a competitor with the cloud has gone from years to weeks. 

Metrics to Gauge 

  • Adoption of Advanced Cloud Platform Capabilities 
  • Dev/Deploy Cycle Time Optimization 
  • Monitoring and Management Oversight 
  • Cost Oversight 
  • Staff Time Optimization 

New perspectives for a new platform 

Any cloud project that goes beyond a "lift and shift" of existing applications from an on-premises environment to a cloud-based infrastructure will introduce new technologies and potentially new complexities. One of the biggest challenges faced by large, established enterprises in particular is the shift from all-inclusive vendor suites to a more heterogeneous arrangement of components from the open source ecosystem, from cloud providers, and from other vendors. 

Managing a complex web of heterogeneous services is rarely possible with single-purpose legacy tools, opening the door for performance issues to arise undetected and causing technical staff to waste time trying to troubleshoot and resolve issues on an ad hoc basis. And while the major cloud vendors often provide powerful consoles and management tools for their hosted services, those tools often fail to integrate neatly with all of the open source software components that increasingly power businesses of all sizes. Therefore, selecting management and monitoring software that provides breadth of coverage and accommodates future change is paramount. 

Thankfully, the cloud revolution has ushered in new kinds of software for managing, monitoring, and understanding complex cloud environments. Infrastructure-as-code frameworks, such as Terraform, allow operators to provision, modify, and manage cloud resources in a safe and repeatable way. And because it is open source, extensible, and platform-agnostic, Terraform naturally accommodates evolving infrastructure, whether that evolution involves replacing an individual component, or moving an application from one cloud provider to another. 

Similarly, modern monitoring platforms have been designed expressly to provide visibility into dynamically scaling, rapidly evolving infrastructure. By integrating with a wide range of cloud services and infrastructure technologies, and enabling users to easily build custom integrations for niche technologies, these comprehensive platforms make it possible to monitor the performance and usage of every component in a modern cloud stack. 

Keeping a cap on costs 

The limitless scalability and self-service nature of the cloud enables developers and operations teams to build and deploy new services quickly, but it can also lead to an infrastructure that is littered with underused, redundant or abandoned components. Compared to the relatively fixed costs of running a data center, cloud services are much more open-ended in terms of potential expenses. Without some guardrails in place, the monthly service bill from a cloud provider can easily creep into the hundreds of thousands of dollars, or more, for very large environments.   

Several powerful tools for tracking, predicting, and optimizing cloud usage and costs have emerged in the past several years. Services like CloudHealth Technologies and CloudCheckr help companies understand and control the costs of cloud infrastructure, while also providing automated analysis and recommendations for optimization and right-sizing. 

Cloud, minus the opacity 

In recent years, we have witnessed a dramatic shift in cloud adoption, as companies in even the most risk-averse industries are joining cloud-native organizations and other early adopters in making significant cloud investments. These investments should produce substantial returns, especially for enterprises that have already attained great success with their existing product line and personnel. With the ability to innovate faster and respond much more quickly to the demands of the market, these incumbents can leverage recent technological developments to maintain or even extend their competitive advantage.   

But companies that are approaching the cloud for the first time should not expect all of their existing software for development, application support, management, and monitoring to make the journey with them.     

Alex Rosemblat, VP of Marketing at Datadog 

Image Credit: TZIDO SUN / Shutterstock