Powers Unfiltered

Greg Nawrocki, as usual, hits some good points today in his posting about Sun Grid.  His main point is that the key to grid adoption is available apps, and I could not agree more. 

This is exactly why we put so much energy and effort into our API, why we give away the Digipede Network Developer Edition, why I don’t leave the office without a few mini-CDs in my pocket to give to worthy developers.  (Apparently, a developer to whom I handed a CD last month posted about it today!)

My favorite comment, though, was Greg’s observation about a fundamental issue that keeps coming up about Sun Grid:

“…why would anyone want to contract for a service they generally already own, and if they don’t it can be had fairly cheap with the potential of a return on that investment[?]”

Why indeed?  Surely not for the reasons Jonathan Schwartz touted in his detailed posting back in March.  He hits all the right notes in describing how grid has been done to date, and then comes to exactly the opposite conclusion that we have.  If the problem is needless complexity and a lack of applications, his contention that it is somehow better to just hand over the whole problem to Sun may resonate with a few IT folks, but not many (as he pointed out himself).  More importantly, it will resonate with even fewer developers. 

It seems highly unlikely that most businesses will find it more economical to adopt Sun’s restrictions on application development than to deploy their own grid — especially as the options for deploying their own become easier and more flexible. 

Don’t take my word for it; see what the team at CRN Labs has to say about their experience adapting an application to Sun Grid, as compared to their experience adapting that same application to the Digipede Network:

The difference in ease of use in Sun Grid vs. Digipede was enormous. Because the Test Center designed an I/O-intensive test program to execute in parallel, the shell scripts required further development, so engineers did not complete the project originally intended to test the Sun Grid. The Sun job was completed after the Test Center contacted Sun’s engineers to help write scripts and compile the program.

Using the same test program with Digipede Team Edition, engineers compiled it using the DOS GNU compiler and were able to write the script in about a half-hour. The job wizard combines job creation with script development so developers need not to learn how to build scripts. Test Center engineers completed the Digipede job in less than one hour vs. five hours for Sun Grid.

Sun Grid will appeal to certain parts of the market, but Greg’s right — new infrastructure is just not very exciting.  New applications, and new ways to build those applications, and new ways to grid-enable those applications — that’s exciting.