Grid watch: Open standards architecture at the GGF
d4_sn0wm4n: The first article is found here
This installment of "Grid watch" provides a quick overview of OGSA, OGSI, and other architecture-related initiatives at GGF.
In the first installment of this column, I gave you a brief overview of the Global Grid Forum (GGF). Now I'll turn my attention to grid architecture, a topic I find extremely hard to talk about. It's not that I think architecture is boring or unnecessary. Quite the contrary. It's just a huge, rambling, complex topic, and my job here is to pick out what's important to the developer community without getting too lost in the weeds.
So what's so important about grid architecture?
Before we talk about grid architecture, I want to take a step back and ask, "What's in it for developers? Why is this important?"
Grid architecture is important for precisely the same reasons that real-world architecture is important. Like standing structures, grids can:
* Come in all sizes
* Consist of many different kinds of materials
* Be used for different kinds of purposes
* Have different levels of traffic (throughput) and capacity
* Require different levels of security
* Require different types and levels of infrastructure
Computational grids are very complex webs of computing resources. Systems in a grid can be running different operating systems on different hardware configurations, be owned by different organizations, contain different levels of available resources to share, and so forth.
A successful grid architecture can help grids manage resources across this heterogeneous environment. A good architecture should provide open, published interfaces that allow all the different grid components to interact. With any luck, a good architecture might even help deliver one of the holy grails of grid computing -- quality of service (QoS).
So that's what's in it for you: good architectures mean good grids.