Playing at the Playground!

Playing at the Playground!

Jun 11, 2012 by in Data Quality

One of my clients has the concept of a Playground on their database platform.  The definition of this environment (for them) is as follows:

  1. Used for prototyping a new application – build it yourself and see if it works!
  2. Used for a proof of concept of software that would require a database or repository.  I keep wondering who has time for this type of fun!
  3. The platform can be used for NO more than 3 months for one application, and then it must deleted from the Playground or put in production.

What has happened is:

  1. The business users are building real applications out there using data (based on views) from the production database.
  2. Most of the applications, in the Playground, have been there more than 30 days.
  3. None of these projects are ever planned for the production environment.  Why — because they work in the Playground.

What the Playground has caused:

  1. Because the users were given the freedom to build anything they wanted, with no thought of data quality or data integration – the practice continues.
  2. Since all work is project based, the project just needs more money to fund more people.  More people to keep the equipment on the Playground in tip top shape.

I think the Playground needs audited, managed properly, and then monitored on an on-going basis.  Maybe the Playground should not be backed up every night – would that push people into production?  What about reducing access instead of opening it up to anyone?   I am pretty sure if we ‘delete they will come’!  I do love a good delete!

Read more of Joyce’s posts on the Data Roundtable.

Tags:

2 Responses to “Playing at the Playground!”

  1. Paul Boal

    Jun 11, 2012

    It seems to me that the reason for playgrounds is to create an “easier way” for people to get things done without the “overhead” of production controls or processes. I’m in support of always creating easier ways to be productive, so “yeah” for playgrounds. That doesn’t mean the way into production can continue to be seen as “too hard.” Seems that this sort of situation is inevitable until your client finds a way to add value for users as they go from sandbox to production. I’d ask that client how often they reach out to sandbox users and say “Hey, I noticed you were running a 4 hour query out there. Can I help you tune that and get it into production where it’ll perform better?” Or “Looks like you’ve got a monthly process out there that you run during month-end-close. Would it be helpful to automate that and send over the results automatically?”

    I think that kind of value add will always go over better than a 3-month-or-I-delete-it stick.

    Reply to this comment

  2. Vish Agashe

    Jun 11, 2012

    I think Playground concept is really good, as long as the rules are adhered to! This breeds creativity to try new concepts. Personally if you adhere to the rules which are set out(no persistence beyond 3 months, not for production, no backups), I think we should not do any audits or worry about data readiness/quality.

    You could envision adding couple of more rules around not using data/insights from playground for decision making, those have to be replicated and gleaned from production/high quality/governed data.

    Reply to this comment

Leave a Reply