Recent Posts by Peter Jeffcock

How can Oracle NoSQL Database be cheaper than "free"? There's got to be a catch. And of course there is, but it's not where you are expecting. The problem in that statement isn't with "cheaper" it's with "free". At Strata Hadoop World last month one of our product managers, Robert Greene, did a talk summarizing what one of our customers found. You can check out some of the slides from Oracle OpenWorld or even a short video (skip to 2 minutes 15 seconds if you are pressed for...

The original question was: why is Oracle Big Data SQL different from those other "SQL on Hadoop" solutions out there? There are different ways to answer that question. One way is to listen to the questions we actually get asked (or end up answering) from the customers we speak to. They tell us that there's actually a better and more important question to ask. It's something like "how do all those SQL solutions help me transparently work with the all data I have in new data...

Minimizing data movement is important. So for any solution linking two or more different data management technologies (like Hadoop, NoSQL and Oracle Database) you need to ask "where is the data processed"? If you're looking at a pure "SQL on Hadoop" solution with no RDBMS link, then the answer here is easy. Hadoop data is processed on Hadoop. Of course, RDBMS data is processed in the database with a different dialect of SQL. Which means you need "at least two different...

Hadoop has not had a reputation for security but the good news is that things are improving. The Apache sentry project provides basic table level access control on Hadoop data, there are solutions using Kerberos for authentication and you can encrypt data on Hadoop. But you need much more than that. Data on Hadoop will often contain personally identifying information and many companies don't want their analysts to see that kind of data in the clear. In fact, regulatory or...

It's great that SQL is well supported on Hadoop.In fact, there are so many different versions of "SQL on Hadoop" it's a bit like going into one of those ice cream stores with a huge variety of flavors and trying to pick one. So which flavor should you pick? Many of them are based on or similar to SQL-92. That was state of the art 23 years ago, but things have moved on since then. More modern versions of SQL are much more powerful, enabling both more and better analytics...

Databases have metadata to tell the database where to find the data. Hadoop is the same. For the most part, its metadata is stored in HCatalog. And all the "SQL on Hadoop" solutions use it to locate data, determine how it's laid out so it can be accessed in parallel, and what the schema is. But if you want to join data in Hadoop with data in your RDBMS, then you need a unified view of that metadata. If you're just "SQL on Hadoop" then you don't have that view, and you need...

We announced an update to Oracle Big Data SQL 1.1 as part of our news for Hadoop World this month. The product has been out for a while now and here I want to focus on a common question: why is Oracle Big Data SQL different from those other "SQL on Hadoop" solutions out there. Or, if you prefer, what's the big deal about Big Data SQL? The thing that gets missed is that Big Data SQL is not just a "SQL on Hadoop" solution. It's a "SQL on Hadoop and NoSQL and Oracle Database"...

SQL on NoSQL is almost a palindrome (that would be "SQL on NoLQS") and at first glance for many people it's also something of an oxymoron. What's the point of using a NoSQL database if you're going to query it with SQL? It's important to note that the applications developers who are writing based on NoSQL databases, are using APIs to access the database. That's to be expected. SQL on NoSQL is not for them. It's for the rest of your organization who don't write Java...

That's the title from #2 in our series of predictions for 2015. You can find an in-depth article here that covers all 7, but I'm going to focus on that second one. Here's what we think: Hadoop and NoSQL will graduate from experimental pilots to standard components of enterprise data management, taking their place alongside relational databases. Over the course of the year, early majority firms will settle on the best roles for each of these foundational components. The demand...