What happens when you combine two field-proven, widely-adopted technologies from tech power houses? You get a super flexible, scalable Hadoop platform to address your big data analytic needs. And the best part is that it really doesn’t matter where you currently fall on the Hadoop adoption spectrum. Deploying a traditional Hadoop distribution? Expanding one that’s already deployed? Starting a brand new Hadoop implementation to merge with your existing IT infrastructure? VCE and Isilon have come together to make your life easier.
So how does it work? Let’s assume you understand some of the core concepts of Hadoop, like NameNodes, DataNodes, MapReduce, and Resource Management. For simplicity sake, let’s classify these components into two categories: storage and compute. In short (extremely short), Isilon takes care of the back-end storage, while VCE takes care of the compute. Check out our latest ESG Lab Report, which goes into far more detail about the architecture and functionality. We also look at things like performance, flexibility, and resiliency.
There are two important things I want to mention here. First, because the storage and compute components are separate, you can scale your components independently. Need more compute components? Just add them. Need more storage? Add another node to the Isilon cluster, rebalance, and off you go. This leads to better resource utilization by not being forced to over-provision storage just because you need more compute, or vice versa. Second, all unstructured data stored in the Isilon cluster can be analyzed in place. That means you don’t have to worry about ingest or staging.
On to the last question: Where does VMware tie into all of this? After all, when you’re talking VCE, you gotta talk VMware. The VCE and Isilon joint-solution for Hadoop is already powerful, but once you factor in VMware Big Data Extensions (BDE), it’s even more impressive. BDE automates the deployment and management of the Hadoop compute clients. It lets you simply create the compute-only parts of the Hadoop cluster in new virtual machines on the Vblock, and all you have to do is provide an address to the HDFS interface on the Isilon cluster. Within minutes you have a virtual Hadoop deployment connected to all your unstructured data stored on your Isilon cluster. Time to analyze!