Wednesday, June 12, 2024

Pivot3 Introduces Serverless Computing

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

At first glance, the concept of serverless computing might seem hostile to the beleaguered server administrator. Pivot3 Serverless Computing by Pivot3 (Palo Alto, Calif.) is a clustered storage solution that simultaneously runs server applications on IP SAN hardware. Is this the end of the server and the end of the need for a server admin? Actually, it’s the reverse.

“This concept allows server managers to take on the management of large-scale storage arrays that are based on the technologies they already know — server hardware, Ethernet and virtualization” said Lee Caswell, cofounder and chief marketing officer of Pivot3. “It’s a great way to leverage server skills into a new area.”

Here’s how it works:

Most SAN hardware is based on proprietary hardware, including solid state memory, batteries and disk arrays. Pivot3 employs its RAIGE software to create a reliable, high-performance SAN made up of many off-the-shelf x86 server hardware appliances running in parallel. Server virtualization software enables Pivot3 to concurrently run its storage virtualization software across all appliances and one additional server application per appliance. Like with conventional server virtualization, you are eliminating server hardware. However, unlike conventional virtualization, there is only one application running per appliance, and it has access to the entire clustered storage array’s bandwidth and capacity. This works because it has plenty of CPU overhead to tap into.

Obviously, this setup isn’t suited to environments where loads of applications are running. But for storage-intensive environments that focus on relatively few applications, it might make sense. Video surveillance, for example, is one area where this approach is getting a lot of play. These days, companies want to store vast quantities of video images and need only a few apps to monitor them.

“In a typical surveillance application, the elimination of stand-alone servers generates savings of 52 percent in rack space, 44 percent in power and 22 percent in cost,” said Caswell. “These servers would not ordinarily be considered for consolidation because they are I/O bound and carry a heavy CPU load.”

Let’s stick with the surveillance example. Normally, companies deploy servers that manage between 16 and 64 cameras. Typically, these servers are something like a Dell 2950 running Windows with a Fibre Channel (FC) Host Bus Adapter (HBA) attached to an external RAID box. In a large application, 40 of these application servers might be tied to a Petabyte of external storage. With Pivot3, all of the 40 external servers can be eliminated, and the applications could be run in virtual machines (VMs) on 40 of the 100 appliances that provide the clustered storage.

As far as “serverless” goes, the server functionality is re-packaged in the IP SAN. However, you don’t have to spend more than $100,000 buying and managing 40 servers. In addition, these machines would require over 40kW of power and gobble up over 2 racks of data center space.

Pivot3 uses a Xen hypervisor running on each appliance. Because only one application runs per node, there is no need to use sophisticated tools to manage VMs. Also, because the underlying clustered storage can tolerate appliance failures as simple RAID failures, it supplies application fault tolerance since applications running in VMs on a specific node can be restarted on a new node in case of a hardware failure.

“The technology is ideal for environments with large amounts of storage and relatively few applications, such as surveillance, medical imaging, video streaming and data de-duplication,” said Caswell.

This article was first published on

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles