The cloud has proven itself as an effective means to scale up infrastructure and resources — and not a moment too soon considering the Big Data and rich media challenges confronting the enterprise.
But while I hate to use the term Cloud 2.0, it nevertheless seems we are on the cusp of a major expansion of the cloud’s role in the data architecture as it evolves into a full-fledged application platform. For the enterprise to take advantage, however, it will have to deploy new generations of application management and orchestration software specifically optimized for the distributed, dynamic environments found that are common in the cloud.
Companies like Lyatiss are already looking to define the new market for “application-defined networking” (ADN), which uses a mixture of automation and performance enhancement techniques to integrate application-layer systems with network resources to boost the availability and performance of cloud-based apps. The company’s CloudWeaver platform provides mapping, deployment, operations and other modules to ensure that applications and services function smoothly across dispersed, heterogeneous architectures. Unlike SDN, which operates on the control and forwarding planes on the network, AND targets application flows in such a way that applications can more readily adapt to individual network topographies.
Meanwhile, Europe’s OpenNebula organization is pursuing an open source strategy in an effort to ensure broad compatibility as applications encounter various cloud platforms in their travels. The Apache-based OpenNebulaApps toolkit sits atop the group’s standard infrastructure management system, providing functions like automated installation and configuration of virtual machines, multi-tiered application execution and virtual build and deployment environments that allow for sharing of virtual appliances across multiple instances. The idea is to provide an expansive environment for enterprises that want to use the cloud as an extension of the data center but don’t want to be locked into proprietary cloud platforms.
Enterprises looking to push application management into the cloud need to familiarize themselves with some of the key differences between a traditional data center system and one geared for the cloud, says Compuware’s Michael Kopp. For example, there needs to be 24-hour transaction monitoring in the cloud, as well as an ingrained ability to keep up with dynamic resource and service configurations. And on a more fundamental level, the focus of the management stack has to shift away from underlying resources and infrastructure in favor of application speed and availability.
App management in the cloud also needs to recognize that business units, not IT, are increasingly responsible for provisioning and employing cloud services and infrastructure. So any management stack needs to address the business needs of the organization, not just the technical ones, according to Oracle’s Jay Parekh. That’s why the company has adopted a value-based approach that stresses business-centered modeling, maturity assessment and component compilation aimed at improving competitiveness and achieving business objectives.
This change in focus represents a larger shift in the enterprise’s responsibilities. With infrastructure shifting over to the cloud, IT will concern itself less with the workings of servers, storage and networking and more on the performance of critical data environments sitting atop dynamic virtual architectures.
IT will still be tasked with making sure things get done, but the day-to-day responsibilities will fall to someone else.