This week Oracle announced the availability (yes you can right away buy and use these systems) to Big Data Appliance X3-2 Starter Rack and Big Data Appliance X3-2 In-Rack Expansion. You can read the press release here. For those who are interested in the operating specs, best to look at the data sheet on OTN.
So what does this mean? In effect this means that you can now start any big data project with an appliance. Whether you are looking to try your hand on your first project with Hadoop, or whether you are building your enterprise Hadoop solution with a large number of nodes, you can now get the benefits of Oracle Big Data Appliance. By leveraging Big Data Appliance for all your big data needs (being this Hadoop or Oracle NoSQL Database) you always get:
Assume you want to start your first implementation on Hadoop, you can now start with the BDA Starter Rack, 6 servers which you can fully deploy for HDFS and MapReduce capabilities (of course we also support for example HBase). All the services are pre-configured, so you have Highly Available NameNodes, automatic failover and a balanced approach to leveraging the 6 servers as Hadoop nodes.As your project grows (and you need more compute power and space to store data) you simply add nodes in chunks of 6 using the In-Rack Expansion, filling up the rack.
Once full you can either add another Starter Rack or add Full Racks to the system. As you do that, Mammoth - the install, configure and patch utility for BDA - ensures that your service nodes are in the appropriate place. For example, once you have 2 cabinets assigned to a single cluster, Mammoth will move the second NameNode to the second Rack for higher fault tolerance.
This new release of Big Data Appliance (the software parts of it) now also include Cloudera CDH 4.2 and Cloudera Manager 4.5. On top of that, you now create multiple clusters on a single BDA Full Rack using just Mammoth, which means you can now patch and update individual clusters on that Full Rack. As you add nodes to a cluster, Mammoth will allow you to choose where to add nodes, how to grow a set of clusters, etc.
Lastly, but not least, there is more flexibility in how to acquire Big Data Appliance Full Rack, as it is now a part of Oracle's Infrastructure as a Service offering, allowing for a smooth capital outflow for your Big Data Appliance.