Notes from the July 2014 ONETUG Meeting

Posted on by SQLAndy
  • Rained as I left (early) for the meeting, confirming again why I always leave early – stuff happens (in Orlando, rain = commute time * 1.5 or more)
  • New location for ONETUG in the Orlando Public Library downtown, public parking garage directly across the street ($2/hour)
  • Meeting was on the 2nd floor in the new Melrose Center. Gorgeous space. Only downside was no whiteboard/flip chart (a miss on my part to not ask ahead of time). Note that they wouldn’t let me in early, so I found a table in the library to go over my presentation until they opened up the meeting space. I also stopped by the used book store and spent a few dollars.
  • About 30 people attended, I think that’s pretty good for a stormy evening
  • I forgot the SQLSaturday flyers! But mentioned it several times as did Esteban (and I’ll send a follow up today).
  • My plan was to get a sense of the audience and then take them through some demos of the performance tools, but I also had in mind to adapt to how they responded to the first few minutes. I ended up just explaining things – no demo, no supporting deck – for the hour. Lots of great questions as I took them through why things work (or don’t) with simple scenarios they could visualize. Lots of fun to just engage directly.
  • Speakers, here’s a phrase I use that can connect you to developers: SQL Server was written by developers. One of my goals is to get them to understand that SQL is a tremendously complex product and getting them to think about the “how” engages them. You can see them light up when I talk about being smart about read ahead reads, flushing out least use plans, etc, etc, not just because it’s interesting but because they get the work involved.
  • Sign in for the meeting was on a web page, an experiment that worked ok, but not great
  • No name badges and not much real networking, I think they could do more there
  • July seems like a great month for someone from oPASS/MagicPASS to visit ONETUG (as lead-in to SQLSaturday) and we should look at the corollary for Code Camp
  • I had a great time. Thanks to Esteban and team for inviting me!

2 thoughts on “Notes from the July 2014 ONETUG Meeting”

  1. Andy,

    I was in attendance for the talk and enjoyed the topics. While you were describing the issues caused by random guids and the sequential guid functions in sql server, I asked if you had heard of the “comb algorithm”. Here’s the algorithm implemented from a dba perspective: http://www.informit.com/articles/article.aspx?p=25862

    I believe is solves the fragmentation problem and the security issues from the other methods. Of course you still have the 4byte vs 16byte issue.

    Patrick

    Like

    1. Patrick, thanks for attending and the follow up. I was just reading http://en.wikipedia.org/wiki/Globally_unique_identifier and it looks like that is pretty close to what we have in the newsequentialid(). It will certainly help reduce fragmentation, the only other variable is how many machines are generating them, you’d wind up with a sequential range from each machine (client or web/app server, depending on the architecture) which in some cases might even be an advantage.

      Like

Comments are closed.