New mgmt. standards needed

When we made the paradigm shift from the glass house to client/server architectures, the distributed and heterogeneous nature of the new model presented major management challenges. Management software adapted, with tools like agents, connectors and in some cases proprietary management technology. SNMP became a de facto standard, which allowed management tools to have some level of integration.

As we embark on another paradigm shift to grids, Web services and other emerging technologies, the challenges are even greater. These environments will be even more distributed and require the integration of an even wider variety of heterogeneous components – and in some cases that integration will be done on the fly.

Web services could be offered for general consumption (for example, a credit card validation service), and users could select and use these services as a part of their normal business processes. So you could have a buyer-and-seller “trading floor” for services, with performance, service levels, cost and functionality as some of the determining factors for the selection of services.

This is vastly different from today’s environment, because in the visionary world of Web services, you may be using services from companies that you don’t necessarily have a “trusted relationship” with, like you do with your business partners today. The span of future infrastructures, including Web service providers’ infrastructures, will extend well beyond the largest infrastructures that are managed today. So, integration challenges will be much greater by orders of magnitude.

While these scenarios will not happen next month, the management of them must be planned for now. To that end, standards bodies – such as OGSA, OASIS, and DMTF, to name a few – are developing standards.

The newly formed DCML Organization – headed by Computer Associates, EDS and Opsware – recently threw its hat into the ring. The management vendors will develop a schema based on XML to allow diverse systems to exchange operational management data for operations in the data center. This group intends to develop Data Center Markup Language (DCML) and then turn it over to a standards body. However, the notable absence of a few big players in the DCML organization (most notably, IBM, HP, Microsoft and Sun) raises some concerns.

These moves prompt several questions. Will the different groups develop their standards in synch, or will we see competing standards? Will the DCML organization be successful at attracting the other large players to play in their pond, or will those companies develop competing standards through other standards bodies?

The answers to these questions will play themselves out as the DCML organization gets going, and as it works to build a constituency. What’s clear is that these standards bodies must be coordinated in their efforts, in their focus and in their eventual output.

The development of standards is a two-edged sword. The good news is that standards are being developed, and the bad news is that standards are being developed.

I’m an advocate for the development of standards; it’s the only way that IT infrastructure management can evolve to handle future demands. However, development of standards can be a long process, and it’s still ongoing. Sometimes, standards initiatives have a way of developing a life of their own, sometimes they are effective and are adopted by the industry, and sometimes they are a bust and fall by the wayside.

I do hope that these standards initiatives are successful – they could change the face of management as we know it today.