PASS Summit 2012 Day 3

Up at the crack of dawn for Lynn Langit’s great session on Amazon’s cloud offerings, called AWS. They’ve got several managed service offerings of interest for DBAs, including:

• EC2: a VM service
• S3: a storage service
• Glacier: a backup service
• RDS: a managed database service

The most compelling for me was obviously RDS, the managed database service. According to Lynn, AWS has been in the market the longest, is usually the most cost-effective, and is most often used in production (highest level of confidence). That’s a compelling trifecta.

We recently had some problems with the HVAC in our server room, and it made us all take pause and wonder if we shouldn’t be moving things off premise and into the Cloud. So, although it doesn’t apply directly to my BI project, I would consider moving my OLTP environment there, including the Dev, Testing, and Production instances. I’d love to hear some stories of people who have migrated from on-premise to RDS, so if you’re out there, please share.

I also picked up a great resource from Microsoft, at their pavilion. It’s a one-page document that diagrams all of the components of the BI offerings. For context, it takes up about half of my bed here in the hotel room – that should give you an idea of how invested Microsoft is in BI. You can download it here, and then head over to Kinkos and have them print it out on a 5 x 3 ft poster for you.

Microsoft BI at a Glance

RT 90, Anyone?

Lynn also mentioned a NFP that she co-founded, which aims to teach kids programming. The programming language that they use is based on Logo, which is what they taught us in Grade 6. Anyone else remember this:

Who else remembers the turtle?

The other highlight of the day was the Birds of a Feather lunch with some other BI professionals. It’s always great to connect in person with professional confrères, and share stories.

(Home run buffet lunch, btw.)

A BI Sam Witwicky

As far as my OLAP Sprint goes, I leave here tomorrow with the following in tow:

• 1 SSTD MOLAP Solution
• 1 Fact Table
• 5 Dimensional Tables
• 1 Cube
• Roles, KPIs, Named Calculations, Translations, etc.
• 2 SSRS Reports
• 0 trips to the Space Needle, Pike St Market, etc.

Barring something terrible happening to my laptop, I’ll return to the office on Monday with a prototype cube on my laptop, which was my goal. I have to say, I feel a little bit like Sam Witwicky in Transformers at this point – ready to protect the cube at all costs.

Protect the Cube at all Costs

I’m very pleased with what I managed to accomplish this week.

I realize that I am lucky to have had the opportunity to execute my experiment, and I don’t think I could have done it in just 7 days without the following key elements:

• Focus (on one goal)
• Isolation (from regular work/life responsibilities)
• Immersion (in SQL Server)
• Access to Resources (Experts, Community Members, Vendors, Books, etc.)

The PASS Summit is undoubtedly the ideal environment for those elements to come together, and make the experiment work.

Please, Try This At Home

For anyone out there contemplating doing a similar experiment, I can’t recommend it enough. As IT professionals, we’re often responsible for planning and executing our own training, and I couldn’t be happier that I decided to take this on, and that I had executive support for it. It was intimidating at the outset, and frustrating at several points along the way (I had some Winnebego Man {NSFW} moments with attribute relationships and MDX), but boy was it worth it. If you’re thinking of doing something similar, find a way to make it work with the resources that you have – you won’t regret it.

I’m also completely exhausted. You could use the BIT datatype to count how many times I went to bed before midnight, but you’d need at least a TINYINT datatype to count the cups of coffee I’ve had in 7 days.

Yes, I know, that table would not scale well.

“Is E-mail Down?”

No VCs have emailed me yet about my buffet fly-over idea yet, so maybe our Exchange server was down for a while or something. Shouldn’t be too much longer. I’ll submit a helpdesk ticket for our sysadmin to make sure. He’ll like that.

Next Steps

The last chicken burrito is down the hatch (sniffle), which means that Chipotle should beat estimates when they next report earnings, but also that my time in Seattle has come to an end. The next steps for me involve a discussion with the executives about how and whether to implement this for real. Those discussions will require careful consideration of the costs involved, and how to scale our resources to build and support a BI platform.

All that can wait until Monday, though.