HP Unveils First Purpose-Built Server for Big Data

 
 
By Chris Preimesberger  |  Posted 2012-11-15 Email Print this article Print
 
 
 
 
 
 
 

Most enterprise servers already have the capability to handle processing, storage and analysis of sizable workloads in the terabytes-to-hundreds-of-terabytes neighborhood. But HP is talking about multiples of that: petabytes and multiple petabytes.

Hewlett-Packard, making a large investment in the latest trend in IT, on Nov. 15 introduced two new servers, including one designed specifically for big data-type workloads.

Most enterprise servers already have the capability to handle processing, storage and analysis of sizable workloads in the terabytes-to-hundreds-of-terabytes neighborhood. But HP is talking about multiples of that: petabytes and multiple petabytes.

"Conventional siloed servers and storage just doesn't work anymore," Jim Ganthier, HP vice president of Marketing and Operations for Servers and Software, told eWEEK. "With increasing storage requirements and business needs changing, we figure that this is an unsustainable approach. These siloed architectures are no longer viable. They cost you too much, require too many admins, use too many different tools and frankly, from a physicality perspective, it's no longer applicable.

"What we decided to do is to use converged infrastructure optics to come up with the ProLiant SL 4500 server. The thing that separates this one from the others is that it is built new and from the ground up for big data workloads."

Workload Requirements Now Vastly Different

There's no question that we're talking about vastly different types of processing and analytics use cases when big data is the subject.
 
Big data application environments such as Apache Hadoop, MPP data warehouses, big data analytics and object stores have very different processing requirements, said Dan Vesset, vice president of Business Analytics Research at IDC.

"Given the large and varied amounts of fast-moving data that needs to be stored and accessed quickly and the different requirements of end users, these workloads can be highly varied, complex and inefficient to manage if run on traditional hardware infrastructure," Vesset said. "In order to fully embrace the promise of big data, it is critical that the underlying infrastructure be optimized for the workload."
 
HP's ProLiant SL4500 Generation 8 (Gen8) provides high-end performance in the type of dense package required by these workloads, Ganthier told eWEEK. It is built on HP's Converged Infrastructure, which is more powerful, yet smaller in size and consumes roughly half the power of older servers, thanks to new, more power-efficient Intel Xeon processors.

The ProLiant SL4500 can take a whopping 240TB of storage in a single 4.3U chassis, or 2.16PB with nine servers in an industry-standard 42U rack. It supports multiple Apache Hadoop distributions, including Cloudera, Hortonworks, OpenStack Cloud Software and MongoDB, Ganthier said.

The SL4500 server in a single-node configuration is available immediately worldwide for a starting price of $7,643, Ganthier said.

Updates to HPC Lineup

HP also announced updates to its high-performance computing lineup. The ProLiant SL270s (also Gen8) features "maximized processor density," Ganthier said, with the ability to support up to 8 Intel Xeon Phi co-processors or 8 Nvidia Kepler graphic processing units per server.

The ProLiant SL270s servers will become available in December, with prices starting at $6,166.

 
 
 
 
Chris Preimesberger

Chris Preimesberger is Editor of Features and Analysis at eWEEK.

 
 
 
 
 
 
 
 
 

Submit a Comment

Loading Comments...
 
Manage your Newsletters: Login   Register My Newsletters























 
 
 
 
 
 
 
 
 
 
 
Thanks for your registration, follow us on our social networks to keep up-to-date
Rocket Fuel