Well something is working...

We had another test and this is much better. We peaked at 3.25MM recs/hour. Not too shabby. Still haven't looked at the mountain of data. I feel like a guy at NASA going through the Mars rover data (Thanks for the analogy, Dan L)

Website is up in full glory

It took some work, but something non-embarrassing is now up on www.istekonline.com. Now that I have somthing there, I need to redouble my efforts to get myself noticed.

Meanwhile, back in the lab... We were doing a full load test last week and the results were disasterous. The DASD is totally not performing. What's up with that? The vendors will be in next week. However, the new CIO wants this thing in this year, no matter what. So it should prove an interesting month.

This week, I will kick off a trial of the Uni-SPF product. This makes Linux look like a mainframe, and since most of our testers are mainframe people, this should make their life a lot easier, rather than learning vi.

Also, this week will try to get the new 2.6 linux kernel tested with our code to see if it makes any improvements.

Finally, need to take a look at the data we were processing during the huge test last week to see if it can be causing the huge performance problems.

Website is almost up

At http://www.istekonline.com


I don't have a link from the main page yet, but at least I am making some progress. I think I am going to resurrect my old copy of VisualPage and use that to edit my website. The doing it by hand method is ridiculous.

BTW, the whole HTML, CSS, JavaScript, etc... technology for building GUI's is total bullshit. This is the worst collection of hacks and garbage design I have ever encountered. Just find a nice website and do a View Source in your browser. HTML is meant for text!!!

Some day, all these people will wake up and say, "Oh my G-d! What were we thinking?", and I will be vindicated!

Meanwhile, I guess I better sharpen up my HTML coding skills so I don't look like a total embarassment when I put up my website.

Approaching the speed of light

I must be accelerating because time is slowing down. I still haven't updated my website with all the things I promised I'd do. Right now I am trying to find a web site creator package to make this a little simpler...

Last week we had a external design review with some guys from Big Blue to get their take on our architecture and whether it can scale. Got glowing reviews, so that is a feather in our caps. Took a lot of time to get ready for it though.

Trying to update our capacity model with the latest numbers. Also trying to cross validate the capacity model by building another one which is estimating from a different direction. The original model I designed was based on looking at per transaction costs and extrapolating, plus some % overhead of OS and monitors, etc... I want to cross validate by looking at overall machine utilization and dividing by number of records processed. By reconciling the two models we should find out if we are overestimating or underestimating the various overhead we put into the first model.

more content to come soon

I started writing up some stream of consciousness brain dumps of some of the technology issues I've been dealing with over the last 5 years on the mainframe Java project. I will post those, but I don't know if I am better off putting them on my website or in the blog. Right now, I think the website seems like a more appropriate place.

I registered this URL with both google and MSN search, but so far I still can't figure out how to get myself listed.