DSXchange

I have just a general question about Parallel Extender. I have never used it myself. We have Windows Servers here. I have heard it is available for Windows DataStage, but that pretty much everyone that uses Parallel Extender uses it on Unix Servers. I am wondering (from a useage point, not a saleperson perspective) if any of you have any thoughts on using Parallel Extender with Windows vs. Unix.

The reason I ask is that it seems like something we'd definitely like to use here. However, upgrading to Unix servers is probably NOT in the cards for quite a long time.

Any thoughts and opinions on this would be greatly appreciated. Thanks!

You gain performance because of a massively parallel distributed processing model. You gain processing performance because you link multiple servers together into a "grid" like processing model.

For a single Windoze server with limited cpu count, you'll probably not gain the distributed processing improvement, but you're on Windoze because you're not dealing with billions of rows of data, right? So, you'll be looking at PX for the platform it provides for the other aspects of the tool suite as well as some of the newer features offered on PX only.

It's kind of like putting a racecar on a track that limits the maximum speed of any car you put on it. It works, but it will not showcase the power of the tool.

Although Windows has come a long way in the business environment it is still perceived by many (including myself) to be a consideration only for small and medium sized businesses.

PX is such an expensive product that I tend to think that the real performance gains are only visible on a large system - by "visible" I mean as a measurable business quantity related to a budget amount for the package against a performance boost measurable in dollars.

If the cost of the PX software (including consulting/development) is a high proportion of the system cost then the benefits would need to be huge. This could be the case with a Windows system. On the other hand a large UNIX system might be so pricey that the price of PX would be a much smaller percentage and thus the cost justification might not need to be as high a value.

Of course there are some huge Windows farms out there and the development of PX in a grid computing environment might change that calculation drastically in the near future.

PX can, when handled correctly, process amazing amounts of data in short order and is a powerful tool in both a UNIX and a Windows environment.