Dell Intros Flexible Server, GPU, Storage Mix With PowerEdge C8000

Dell on Wednesday launched a new blade chassis that allows customers to mix-and-match server, GPU and storage blades to meet a variety of high-performance workloads.

That chassis, Dell's new PowerEdge C8000 series, provides a single server, GPU and storage platform that can be adopted for a variety of purposes, said Brian Payne, executive director of server sales for the vendor.

"It gives customers the ability to adopt a single platform but use it for several purposes including hosting managed services or high-performance computing or big data," Payne said.

[Related: Dell Targets HP, Cisco With New Blade Arrays, Converged Infrastructure ]

id
unit-1659132512259
type
Sponsored post

Unlike the converged infrastructure offering which Dell unveiled in June as a way to unify blade servers, networking and the new EqualLogic storage blade into a single unified management system, the PowerEdge C8000 is more of a traditional server and storage offering, but it's optimized for performance and capacity.

The PowerEdge C8000 consists of a new 4U chassis with 10 vertical slots that can be filled with three types of sleds.

The first is the PowerEdge C8220 compute sled, which combines a two-socket Intel E5-2600-series-based server with up to 16 DDR3 RDIMM memory modules and two 2.5-inch SSDs or SATA drives. Each C8220 takes up one C8000 slot, Payne said.

The second blade is the C8220X GPU sled, which takes up two slots. In addition to the same compute capabilities of the C8220 sled, the C8220X also has one or two Nvidia Tesla GPUs for use in high-performance computing tasks.

The third, the PowerEdge C8000XD storage sled, is a two-slot blade device for adding high-density storage to the C8000. Up to four of the C8000XD sleds can be daisy-chained together for a maximum capacity of 144 TB of SAS, SATA or SSD storage in a 4U form factor, or 1.4 petabytes in a full rack, making it suitable for big data. Payne said.

The PowerEdge C8000 is slated to ship this month with two power sleds and eight open slots in which customers can choose from the C8220, C8220X and C8000XD sleds, said Armando Acosta, product manager for the C8000.

By December, however, Dell plans to start shipping a separate 3U power distribution unit (PDU) that will replace the two power sleds in up to three PowerEdge C8000 chasses, freeing up the two power slots for use with more compute or storage sleds, Acosta said

NEXT: Dell PowerEdge C8000 And The Channel

Dell's new PowerEdge C8000 platform comes to market at a time when customers are re-evaluating how they purchase IT infrastructure equipment, said David Cantu, co-founder and COO of Redapt, a Redmond, Wash.-based solution provider and Dell partner.

Redapt, which has grown its business from $60 million in 2010 to an estimated $180 million to $200 million this year in large part due to a fast-growing social media and social gaming business, recognized over two years ago how cloud adoption is impacting the system integrator and solution provider business, Cantu said.

"We saw that customers would start buying equipment by the tens of racks or hundreds or thousands of racks," he said. "So for use, we wanted to be flexible and make it easy for customers to take their pick of storage and compute resources."

Redapt has already installed a couple evaluation units of the PowerEdge C8000 system, Cantu said.

"It's a natural evolution of Dell's C6000 line, except that customers are getting into higher CPU and memory density per rack, which is what they are asking for," he said. "We're already seeing fairly good demand for the C8000 based on the customer evaluations."

The compute and storage density customers require depends on their applications, which is where the flexibility of the C8000 comes into play, Cantu said.

"We have the C8000s at public cloud providers where they are running Web-scale apps or public-facing websites," he said. "Apps today are demanding more memory. For public cloud providers and hosted cloud companies, for business reasons, the more memory they can get into a rack, the more efficient they are."

PUBLISHED SEPT. 20, 2012