Look At The Nitty Griddy

There are two ways to look at the formation of the new Grid Consortium, headed by Argonne National Laboratory and Globus veteran Greg Nawrocki. It could be that things in the grid world are moving ahead briskly and the consortium's task -- to promote the commercial use of the Globus Toolkit, an open standards building block for enterprise-level Grid implementations -- is the natural next step. Or it could be that, despite widespread academic use, grid has yet to be understood, much less accepted, among businesses.

What confounds the matter more are the obvious gaps in the consortium: Microsoft and Oracle, both of which have their own grid objectives, are not members. Bob Thome, Group Product Manager for Oracle's Distributed Database Product, will participate in the upcoming GlobusWorld 2005 conference in Boston (Feb. 7-11). He and others will coach attendees on how to lobby for grid in their organization's budgetary discussions. But Oracle itself is not joining Hewlett-Packard, IBM, Intel, and Sun Microsystems in the consortium.

Neither is Microsoft Research. An early sponsor of Globus, Microsoft Research has been working to develop grid applications using .Net with only partial success. It is also reportedly working on a commercial grid project, called BigTop. Yet, it, too, has not been named as a member of the consortium. It's not clear why.

Meanwhile, statements from Globus Consortium founding members must be taken in context. For example, the statement by Richard Wirt, vice president, Intel senior fellow, general manager, Software and Solutions Group, that "[g]rid computing is going mainstream," might be seen as slight hyperbole or at least extreme foresight. About a dozen high-profile companies, including Oracle, have pursued Grid strategies based on the Globus Toolkit since 2000. But there is no evidence of widespread adoption of grid computing by enterprise end users. So, too, the statement by Ken King, vice president, Grid computing, IBM, that "[b]usinesses are adopting grid computing at a quickening pace" is a bit vague.

Clearly, the Globus Consortium will serve a useful purpose. But for the commercial adoption of grid to be truly widespread, one of its tasks should be to shed light on actual use of grid computing by enterprises not connected to Globus; the other task should be to explain the absence of Microsoft and Oracle.