Last week, ezine GRIDtoday published my contributed opinion on the creation of the Open Grid Forum (OGF). I remain enthusiastic about the merger that created OGF. However, I also remain skeptical that this merger will have a major impact. (More on that below.) According to GRIDtoday’s stats, mine was last week’s second most-read item. I drew second to a Q&A piece with OGF president and CEO Mark Linesch. I’m good with this 🙂
Not entirely to my surprise, none of the feedback I’ve received privately or publicly has indicated that I was way off base in this item. Nonetheless, there was feedback to the original blog entry on which the GRIDtoday item was based.
Virtualization and Grid Computing aren’t necessarily mutually exclusive. I confess I painted them as such in the GRIDtoday item. I equated virtualization with vmWare and the like. I did this for two reasons. Firstly, to effectively argue my point. And secondly, to represent the non-Grid perception that is pervasive elsewhere in IT. As I presented and wrote numerous times while with Platform Computing, Grid technology can be used to create virtualization solutions. And vendors like Platform can also use virtualization technology to create grids. This reciprocity provides a excellent segue to my next point.
Grid Computing means everything and nothing. In other words, the definition remains contentious. This is true despite the existence of a three-point checklist from Ian Foster. Foster, along with collaborator Carl Kesselman, is responsible for coining the The Grid. In addition to this and other technical definitions, there are marketing and other definitions of Grid Computing. The number of times Grid-related panel sessions devolved into heated discussions on the definition of Grid is also testimony to this difficulty. How much effort is spent debating the definition of Web services or virtualization? How much confusion is created when such definitions are provided? Not a lot in the case of Web services or virtualization. And in these cases, most people have some idea as to the context. Can the same be said for Grid Computing? Not in my estimation. I believe that all of this underlies the everything-and-nothing nature of the beast. It’s not a great start when uncertainty principles apply to something as basic as a definition.
Customers just need to get things done. If what they use is standards-compliant that’s great. If it isn’t, but allows them to make progress, then that’s fine too. This applies to some customers, but not all. (I wrote about the value of standards to customers elsewhere. Now as a customer, I do what I can to make standards-wise choices.) Needless to say, standards-cavalier attitudes don’t help the fortunes of organizations like OGF.
I know it sounds cliche, but this is a critical time for OGF and for Grid Computing. I still believe that “Grid computing sorely needs to deliver definitive outcomes that really matter.”
In a future entry, I’ll try to share some suggestions for such outcomes and Grid Computing in general. If you have any suggestions, feel free to comment on this entry or drop me a line via email to ian DOT lumb AT rogers DOT com. I look forward to hearing from you.
Pingback: Ian Lumb’s Blog » Blog Archive » GRIDtoday Article on the Open Grid Forum: Post at Grid Meter
Pingback: Grid: NSF Seeks Definition « Ian Lumb’s Blog