we write about the things we build and the things we consume
Romain Deville

metadata, from one cloud to the other

A few weeks back, my first big metadata project at MetaBroadcast came to an end with its official go-live.

Born from the need to migrate from a legacy API to its more advanced successor, we had to play the role of metadata-interpreter between an historic British broadcaster and a modern television platform.

By switching to a more recent API, our aim was to translate a more advanced data-set to the TV-Anytime format and improve things for the receiving party in the process.

Being involved in a project of that scale was a first for me on many levels:

  • Having to act in the middle of two large teams, one on the client side—the biggest European broadcaster—the other on the side of their beneficiary
  • Having to deal with both other teams expanding each time new knowledge, skills or decision making were needed
  • Having to work on a project that would impact about 2 million households around the country
  • Having to deal with metadata on such a large scale
  • Having to work on a really tight deadline

how it all started

The project started in October with an original go-live planned for early 2015. Straight away, we started talking about the original scope and the improvements we could bring to it, after which the fun really started: planning, “spec-ings”, mapping and coding… Quickly, we discovered that the schema of our new data source was extensively different from its predecessor, therefore making our task quite challenging, yet very enriching.

Following a very tight schedule, we were able to, iteration after iteration, implement new features and tackle bugs very quickly. Once our core code appeared to be fairly stable, we had to go through the exhausting task of comparing and analysing screens and screens of metadata in order to understand and find a solution to our final issues.

This particular task felt like a long journey during which emails were bouncing like ping-pong balls; fortunately, thanks to the invaluable help and flexibility of our two main partners (and some good teamwork around the ticketing system), we reached a point where we could finally pretend to have our work pushing metadata on production system.

During that phase, communication and flexibility proved to be key, allowing us to overcome all the issues that came across our path. It’s also during that period that colleagues completely new to the project joined our team, which introduced potential complexity to the MetaBroadcast family but we had to keep the rhythm steady for the last lap of this race. And despite having to work harder myself, I’m happy to report we pulled through.

metadata

how it came to an end

That final sprint ended up being the most challenging for us: technical hiccups, hidden bugs and a tight schedule put us through demanding times but we powered through.

As we got closer and closer to the finish line, the number of people involved in the discussion started to grow, making this project ever more important in our eyes. Fortunately, after all the hard-work from everyone involved, seeing it being released was a big satisfaction for us all.

Thanks to the flexibility of our partners, the official launch was granted a complete week which allowed us a comfortable schedule to set everything in motion. Everything went almost flawlessly and before we could actually sit back and realise what happened, everything was already up and running while by that time, our minds were occupied with a new set of challenges.

Now, being weeks away from launch, the pressure went down drastically and it’s became easier to look back on what we did well and what we might have overlooked all along this adventure. In retrospect, this final sprint did brought us some mixed-feelings, somewhere between nostalgia and deliverance.

If I had to take one thing out of this, it would be, as I mentioned on multiple occasion in this post: flexibility.

Throughout the project, things did evolve in quite an unexpected fashion and flexibility is what allowed us to push through and get everything ready and working in a timely manner.

This project was for me a great introduction to metadata in a large scale and even if it wasn’t easy every day, I enjoyed very much the learning experience that it gave me.

What about you people? Any interesting experience that you want to share with us? Please let us know by leaving a comment below or tweeting us.

blog comments powered by Disqus
sign up to #metabeers
slideshow