Announcing: Big Data Appliance X7-2 – More Power, More Capacity

Big Data Appliance X7-2 is the 6th hardware generation of Oracle’s leading Big Data platform continuing the platform evolution from Hadoop workloads to Big Data, SQL, Analytics and Machine Learning workloads. Big Data Appliance combines dense IO with dense Compute in a single server form factor. The single form factor enables our customers to build a single data lake, rather then replicating data across more specialized lakes. 

What is New?

The current X7-2 generation is based on the latest Oracle Sun X7-2L servers, and leverages that infrastructure to deliver enterprise class hardware for big data workloads. The latest generation sports more cores, more disk space and the same level of memory per server. Big Data Appliance retains its InfiniBand internal network, support by a multi-homed Cloudera CDH cluster set up. The details can be found in the updated data sheet.

Why a Single Form Factor?

Many customers are embarking on a data unification effort, and the main data management concept used in that effort is the data lake. Within this data lake, we see and recommend a set of workloads to be run as is shown in this logical architecture:

In essence what we are saying is that the data lake will host the Innovation or Discovery Lab workloads as well as the Execution or production workloads on the same systems. This means that we need an infrastructure to both deal with large data volumes in a cost effective manner and deal with high compute volumes on a regular basis.

Leveraging the hardware footprint in BDA, enables us to run both these workloads. The servers come with 2 * 24 cores AND 12 * 10TB drives enabling very large volumes of data and CPUs spread across a number of workloads. So rather then dealing with various form factors, and copying data from the main data lake to a side show Discovery Lab, BDA X7-2 consolidates these workloads.

The other increasingly important data set in the data lake is streaming into the organization, typically via Apache Kafka. Both the CPU counts and the memory footprints can provide a great Kafka cluster, connecting it over InfiniBand to the main HDFS data stores. Again, while these nodes are very IO dense for Kafka, the simplicity of using the same nodes for any of the workloads makes Big Data Appliance a great Big Data platform choice.

What is in the Box?

Apart from the hardware specs, the software that is included in Big Data Appliance enables the data lake creation in a single software & hardware combination. Big Data Appliance comes with the full Cloudera stack, enabling the data lake as drawn above, with Kafka, HDFS, Spark all included in the cost of the system. The specific licensing for Big Data Appliance makes the implementation cost effective, and added to the simplicity of a single form factor makes Big Data Appliance an ideal platform to implement and grow the data lake into a successful venture.

Let’s block ads! (Why?)

Oracle Blogs | Oracle The Data Warehouse Insider Blog