Modernizing the Application Stack Paves the Way for a Multi-Cloud Strategy

Cloud computing has transformed how IT professionals need to think about computing resources and application development. Historically, we have thought about things in an “infrastructure up” manner, focusing more on compute, storage, and network resources than on the application. But in truth, everyone consumes IT from some type of application. As described in the Intel IT Annual Performance Report, “Driving the Digital Enterprise Transformation,” we are transforming our cloud strategy to be better aligned with business needs by taking an “application down” approach—enabling us to maximize the business value of the cloud.

A Healthy Cloud Strategy Includes Application Management and Modernization

We have developed a process for application rationalization—also referred to as application portfolio management—that examines each application from a business need perspective, as well as from a workload placement perspective. We use an internally developed application profiler to inventory all applications in use and determine where the application is in the application lifecycle.

First, the application owner uses a simple decision tree to see if the application is still needed (some applications may be duplicative, or no longer provide business value). If it’s not, we EOL that application. If the application is still needed, we determine if it is hosted in the “right place.” If the app is non-differentiating from an Intel business perspective, we migrate that application to a public cloud software-as -a-service (SaaS) model. For applications that are strategic, we re-architect them to take advantage of platform-as-a-service (PaaS) and container-as-a-service (CaaS) capabilities. In certain cases, some applications are not compatible with a cloud environment and must remain where they are.

Application Migration Rationalization Process

For apps that are indeed in the right place, application management primarily consists of maintenance. For apps that are in the “wrong place,” we develop a plan that states where the app is today, what we’re doing to get it to the right place, and when and where we’ll move it. Regular reports from the application profiler track progress for each app.

It is important to realize that the application rationalization and workload placement decisions are not a “one-and-done” process—we repeat these processes on a regular cycle because the right or wrong place is dynamic. If a new SaaS application enters the market, and we make a decision to use it, any app that is then duplicative or unneeded is immediately designated to be in the wrong place even if was in the right place before.

The application management process I have just described empowers application modernization—the process of re-architecting an application to be cloud-native. A cloud-native app is abstracted from the infrastructure. One often-overlooked benefit from this approach to app development is the elimination of infrastructure dependency. In the traditional app environment, if you put an app on a server, the application runs only as long as that server and the associated network remains running. As a result, you build in redundancy (fail-over servers and networks). But a cloud-native app moves the redundancy to the application layer, using containers that can transfer from one data center to another, private or public, as necessary.

Modernizing the application can, in fact, change the infrastructure itself. Just recently, we experienced a network outage, and every legacy application took noticeable time to fail over to another server and some just simply crashed. But one containerized app was not affected by the outage—it seamlessly transferred to another data center and the outage was invisible to that app’s users. This illustrates the benefit of building resiliency into the application to reduce the reliance on the infrastructure.

Developer Outreach Increases Cloud Strategy Success

It is human nature to resist change, especially if you’ve been doing something successfully for ten or twenty years. We find that, because many of our IT app developers don’t understand or haven’t been exposed to modern cloud application development, we have to shift much of our attention from technology to training. We initiated programs to educate developers about what cloud-native app development means and what the advantages are. When application developers understand the value, then they embrace it.

Based on this experience, we believe it is important to remember that “cloud strategy” and “cloud transformation” aren’t so much about moving everything to a private or public cloud environment. To take true advantage of cloud capabilities, you have to completely change how you develop applications. And that requires re-educating the people who are developing the applications.

Cloud Strategy Three-Year Roadmap

We first used our application rationalization and workload placement processes on IT-owned applications (that took about a year). Now, we are using the knowledge gained to refine our processes and turn our attention to applications across the enterprise (another year). Phase 3 will encompass ongoing application and workload rationalization to optimize our use of public and private cloud services – leading to a true multi-cloud environment.

Read More about Intel IT’s Cloud Strategy

To learn more about how Intel IT’s multi-cloud strategy is contributing to Intel’s digital transformation, read the 2018-2019 Intel IT Annual Performance Report, “Driving the Digital Enterprise Transformation.”

Published on Categories Cloud ComputingTags , , , ,
Chris Sellers

About Chris Sellers

Director, IT Software Defined Infrastructure (SDI) | Information Technology | INTEL CORPORATION Chris Sellers is the Director of the Software Defined Infrastructure (SDI) group within Intel's Information Technology. In this role he leads Intel’s enterprise hosting and Lab Computing capabilities responsible for all Intel data centers, infrastructure as a service (IaaS) and cloud capabilities, Lab Computing and managed hosting solutions. As a 25+ year IT executive, Chris has held a number of roles including GM for IT Information Security, Director of IT Service Operations, GM for IT M&A, Director of Supply Network Operations and many other operations and engineering roles across IT. Chris has been with Intel for 21 years and is the recipient of two Intel Achievement Awards and also led the largest divestiture in Intel’s history.