Issue link: https://insights.oneneck.com/i/1293607
CHAPTER 3 The Impact of Containers 23 These materials are © 2017 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited. With the right tools, developers have true infrastructure power right at their fingertips, alongside their operations counterparts. In this software-driven data center model, all the servers and storage and networking equipment that drives the environment are simply software elements that can be controlled by code. But not every organization will require the exact same balance of shared infrastructure management between operations and development. With tools such as Docker Enterprise Edition, oper- ations will still be able to maintain the control they require, while granting developers the access they need to get their work done more quickly. In fact, there is a real-world example that best serves to illustrate this, and it's found in the nearby sidebar, "HPE IT: OneView for centralized control." HPE IT: ONEVIEW FOR CENTRALIZED CONTROL HPE internal IT has contributed a ton of work to the open-source community and continues to do so. In addition, HPE has extended its OneView product with a Docker plug-in than enables automated deployment and management of Docker containers. According to HPE Pointnext architects: Prior to completing our work, the only thing we were really miss- ing [in order] to integrate it with native Docker tools was a soft- ware development kit or SDK. So we invested in creating that SDK for OneView and we first coded for OneView 2.0 and we imple- mented a plug-in for Docker Machine, which we open-sourced. That's available at GitHub.com. That gave us the first building blocks of creating an SDK that we then applied to more scaled-out solutions for provisioning. The impact has been very good. For example, one of the teams of developers that manage[s] our infrastructure wanted to learn how to use our playbooks. We did some training with them on inventory management, after which we made the decision to deploy a series of secondary production nodes with the latest inventory files. They spent basically 35 minutes, and were able to deploy a full production cluster within that time. Prior to Docker, this would have been a weeks- or months-long project.