Announcing a Hortonworks Hosted Sandbox

Bit Refinery is now offering a hosted version of Hortonworks popular Sandbox. The HDP Sandbox is a full Hadoop environment with numerous tutorials for people that want to explore this new technology.

We are offering a limited amount of free 14 day hosted Sandboxes for anyone that wants to get to know the Hortonworks flavor of Hadoop. If you and your company are exploring the need for a big data solution, head on over to http://sandbox.bitrefinery.com to apply.

Amazon EC2 Costs vs. Bit Refinery

We’re often asked how we compete with Amazon. After all, they are a huge company with billions of dollars in revenue. When it comes to big data and Hadoop, they have a good plan which is to allow you to transfer and store your data for a low cost which sounds great at first. The problem is if you actually want to analyze and retrieve your data which can add up to a small fortune if you don’t watch out.

We here at Bit Refinery have a very simple business model where we provide your infrastructure just as if you were going to do it yourself. We believe you have better things to do than operate your own clusters and servers AND you should expect a consistent bill from month to month based on your usage with no surprises at the end of the month.

We’ve put together a document that illustrates the cost differences between a 20 node Hadoop cluster run on Amazon EC2 vs. Bit Refinery. The document can be found here.

amazonec2_pricing

Don’t use Hadoop – your data isn’t that big

We ran across a interesting blog post by a Chris Stucchio from last year. It can be found here:

http://www.chrisstucchio.com/blog/2013/hadoop_hatred.html

It is a very entertaining article and one of the best quotes is “Too big for Excel is not “Big Data””. You would think since we offer Hadoop hosting that we would be against this article but in truth, we couldn’t agree more. Having been in the hosting and IT world for many years, the co-founders of Bit Refinery have seen many fads come and go. Big data is definitely here but how many people really have “big data” and even if they did, would the cost of analyzing it be worth the additional sales it would bring?

There are many RLDBs (relational databases) out there that can easily crunch hundreds of millions of rows with no issues. With the performance of most of the hardware out there, we’ve only run into a few situations where companies truly have a big data problem.

We had a blog post yesterday that points our readers to the Hortonworks site where they have some good examples of true big data. We even have a customer that is using SQL Server to house a large web retail database with a few tables over 1 billion. You could even take one of our Hadoop nodes with 64 or 128GB of Ram and use one of the many in-memory databases out there to crunch your data. Lots of different options out there…

So, in conclusion, we feel your pain Chris. We enjoy reading the hype of all the startups and new products coming out on a weekly basis but sometimes the obvious answers/tools have been around for 10 years. (ie. Perl, Python and of course good old Korn shell scripting. 🙂 )

– The Bit Refinery Team