Organizations moving Big Data environments from lab environments, into production, have discovered that:
- in-database analytics, statistical analysis engines, full text retrieval, predictive scoring, computationally-intensive algorithms, and automated decision management require far more from their storage environment, in terms of IOPS and throughput, than conventional distributed commodity compute-and-storage architectures can provide, economically
- the staggering second-order costs of operating large-scale distributed computing environments, in terms of energy, rack space and personnel, are and largely invisible when small-scale distributed computing environments are running in kick-the-tires testbeds and lab configurations, but immediately and painfully obvious when those testbed environments are moved into production, and scaled to production levels
- the ability to integrate Big Data environments into an IT organization’s methods and practices for operating production technologies is at least as important, in terms of cost-effectiveness, as the performance characteristics of those environments.
Additionally, conventional RDBMS engines have demonstrated that, practically speaking, conventional RDBMS technology coupled with scale-up SMP server architectures are as effective, and less expensive, than exotic Big Data environments like Hadoop, for most workloads operating on data sets of 500 terabytes or less.
What our customers say
“The X-IO ISE storage solution is competitive on a cost per TB basis and also provides us with high performance, 100% utilization and superior reliability. These benefits are then amplified by the five-year warranty provided by X-IO. We can expect at least a five year duration on all ISE systems and in some cases we have seen in excess of five years and uncompromised performance.”
“We have been able to draw a direct line between the level of performance we’re getting out of the ISE units and the performance we see at the application layer. It’s important to us that we were able to monitor and track that. In a lot of instances, with a lot of other vendors it is so abstracted in terms of what the performance actually means to the application.”
– Managing Director and CIO/Argus Information and Advisory Services
X-IO’s Intelligent Storage Elements allow organizations to operate conventional RDBMS-based applications and exotic Big Data environments on top of a common, high-performance storage pool. ISEs provide extreme levels of performance acceleration for all data-intensive applications, whether those applications are deployed on top of merchant DBMS technologies, or exotic Big Data environments. Providing absolutely linear scale up to 2048 ISEs, and using iSCSI, direct FibreChannel attachment and FibreChannel fabric attachment, we provide hundreds of gigabytes of throughput and millions of IOPS to data-intensive, analytically-complex applications, with zero-touch, never-fail robustness, and energy, space and personnel costs that are typically less than a third of both traditional enterprise storage arrays, and commodity “scale-out” distributed storage environments.
Whether your Big Data workloads are characterized by complex, high-volume SQL, proprietary statistical analysis, large MapReduce streams, or intricate extraction, transformation and loading (ETL) logic executed against petabytes of Hadoop-stored data, X-IO’s Intelligent Storage Elements provide the industry’s most cost-effective method of underpinning your Big Data environments with ultra-high performance, high-reliability storage infrastructure that scales out, or scales up, as your applications require.