It happened with distributed architectures. It happened with storage networking. It happened with virtualization. And now it's happening with the cloud.
It seems like every time there is a shift in data center operations or infrastructure, the drive to deploy new technology always crowds out thoughts of managing it.
But because the cloud entails such broad interaction with a wide variety of platforms, establishing a truly effective management regime is proving to be elusive at best. A number of organizations have taken up the mantle of establishing some sort of cloud standardization, but it's difficult to imagine that there could be the kind of universal interactivity that we see on the Internet.
But that hasn't stopped people from trying. The latest move comes from the Object Management Group (OMG), which has gathered a number of organizations at its Cloud Standards Summit in Arlington, Va., this week with an eye toward coordinating their various cloud-related activities. The effort includes the Distributed Management Task Force (DMTF), the Open Grid Forum (OGF), the Storage Networking Industry Association (SNIA), the Open Cloud Consortium (OCC) and a host of others. The group is hoping to build on work by the Standards Development Organization Collaboration on Networked Resources Management (SCRM), which is focused on a wide range of coordinated management standards.
In the meantime, enterprises will have to contend with the numerous management solutions already in the channel, most of which are geared to specific vendor platforms. CA, for example, is targeting Cisco/VMware users by coming out with an upgraded management platform geared toward the vSphere 4 systems running on Nexus 1000V virtual switches. The package includes the Spectrum Infrastructure Manager, Automation Manager and eHealth Performance Manager modules, all of which are designed to extend management capabilities across physical, virtual and internal cloud infrastructures.
That sort of cross-environment capability has caught the imagination of other management firms as well. FastScale Technology, for one, just came out with a new version of its FastScale Composer Suite outfitted with what the company calls Application Blueprinting. The technology uses lightweight logical servers with "just enough operating system" (JeOS) to house customized Windows and Red Hat configurations complete with analytical and visualization tools, as well as web-based management tools, that can span physical, virtual and cloud environments. The company says it also can establish a software repository to oversee all your assets, servers and usage patterns across virtual environments like Vmware Infrastructure and cloud services like Amazon's EC2.
Yet another approach comes from RightScale, which recently added SpringSource's Hyperic HQ web application monitoring module to its RightScale Cloud Management Platform. The company's aim is to provide broad-based monitoring of numerous cloud-based software products, such as application servers, databases, messaging and authorization servers and even HTTP servers with an eye toward maintaining steady performance on public cloud platforms like EC2 or Google's App Engine. The system relies on monitoring agents embedded into workload images to extend management capabilities like application deployment, scaling and monitoring, into the cloud.
The problem in managing complex environments, particularly in the early stages, is that the technology is so fluid and the architectures so interdependent that simply mapping the relationships between all the various elements is an exercise in frustration. As cloud technology matures, however, there is every reason to believe that management capabilities will as well.
The unfortunate thing is that the time to lay down an overarching management scheme is before, or at least while, the infrastructures are being built, not after. For the early adopters, then, there could be some difficult times ahead. The rest, of course, will only gain from their troubles.