Why I’m at Amazon Architect Training This Week

People bring me in when they’re having data problems.  They can’t store data fast enough, they can’t make it reliable enough, they can’t hire people to manage it, etc.  When I’m in the conference room, it’s because there’s a fire in the disco datacenter.

You know how it is as a DBA, though – DBA means Default Blame Acceptor.  Everybody thinks the fire started in the database, but often it’s the SAN, VMware, crappy code, bad third party apps, or any number of combustible materials.  The company gets more and more concerned about the growing smoke, and they ask the DBAs, “Who’s a SQL Server expert you can call to put this fire out?”  The DBA thinks about my crazy blog posts and blurts my name out – mostly because he wants to find out if I’ll show up in the Richard Simmons costume.  (That costs extra.)

Budget Fire Extinguisher

Now put yourself in my shoes: I show up in a conference room or on a WebEx, and there’s a huge problem somewhere in the infrastructure.  Everybody involved is pointing fingers at each other, and they’re all armed with volumes of reports proving that it’s not their problem.  In a matter of 3-4 days, I need to:

  • Find the real root cause of the problem
  • Prove it to everyone involved using their own language
  • Show a few possible solutions and recommend the right one
  • Convince them to implement it as quickly as possible

SQL Server isn’t the only fire extinguisher, and I have to know how to put out data fires with other tools.  Amazon’s got a ridiculous list of services that are easy to get started with, including:

  • Relational Database Service – point, click, and deploy Microsoft SQL Server, Oracle, and MySQL instances. Amazon manages the backups, patching, and security. The MySQL ones even support readable replicas and replication to multiple datacenters.
  • DynamoDB – super-fast NoSQL database hosted on SSDs.  You pick how fast you want it to go, and Amazon makes it happen.
  • Glacier – store your backups in the cloud for $.01 per gigabyte per month with no cost for incoming data.
  • Import/Export – ship them a USB drive, and they’ll hook it up to the cloud.  For folks with slow upload links, this is the fastest way to move your data online.

That’s why I’m in Dallas, Texas for a few days attending Amazon Web Services Architect Training.  It’s a three-day design session that covers how to design solutions with their services.  It’s not going to make me a Certified Master of Cloudiness across their broad range of tools, but that’s not the point.  Clients don’t usually want me to do the work myself: they want me to find the right answer fast, get the staff on the right page, and let the staff knock out the work together.

If you’re a data professional, and you’re frustrated when people keep saying it’s a database problem when it’s not, what are you doing to bridge the gap?  Are you frustrated that The Other Guy doesn’t know anything about SQL Server?  Or are you reaching out to learn The Other Guy’s technology to help him to see where the smoke is coming from?

Previous Post
When Your DBA Job Sucks: The Phases of Burnout
Next Post
Email Templates for Recruiters and Questions

2 Comments. Leave new

  • I agree with that. Now I am DBA but I started my career as Windows support analyst and my Windows knowledger help many times to “solve” DB problems.
    Thinking about being a data professional, I will take Cloud and Big Data courses, but some network and SQN training are on my to-do list. Maybe I should put Amazon web services on my courses list.
    We never know if poiting-finger professionals are telling us DBA the truth. It is better to know their tecnologies to find out where the real cause of DB problems are.

  • Stanley Johns K
    September 13, 2012 10:53 pm

    Thanks for the info on Dynamo DB (NoSQL on SSDs, Dynao). Looks like a hybrid architecture is the way to go for high performance with relational logic.


Leave a Reply

Your email address will not be published.

Fill out this field
Fill out this field
Please enter a valid email address.