Plenty of attention has been given to the ways in which cloud computing and analytics are changing the way organizations do business, but the changes being wrought by these technologies on underlying infrastructure are no less significant. Alain Andreoli, the head of HP’s server business, recently appeared on theCUBE to shed light on what’s going on behind the scenes of the enterprise.

For the last few decades, CIOs concentrated most of their time and budget on a handful of back-office applications corresponding to the most critical business processes within the organization. Now, they have entirely new data sources in others parts of the stack as well as outside the firewall to worry about as well. That’s growing the operational burden beyond the capacity of the traditional server farm, Andreoli said.

The search for ways to handle the exponential growth of unstructured information is sending organizations into unfamiliar territory such as high-performance computing, a market historically confined to government and academia that HP has begun actively pursuing recently. But at the same time, the proliferation of cloud services is creating pressure from the opposite direction to cut costs.

“High-performance computing, which used to be a bit of a niche market, is becoming mainstream as more people want to crunch their data. So this is blowing new life into HPC while the cloud is driving an all new model for compute,” Andreoli told theCUBE host Dave Vellante. As organizations adopt different approaches to implementing that model, HP is adding variety to its server portfolio.

Most traditional enterprises are taking the converged infrastructure route, replacing their dis-aggregated architectures with integrated modules that combine hardware building blocks into a single chassis with built-in management software to reduce administrative overhead. HP has its own spin on the paradigm that revolves around allowing customers to tune what goes into the box based on their needs.

“You can have compute and object storage in the same rack; you just decide what the compute nodes look like – whether it’s disk or SSDs – and you have it integrated for the whole rack,” Andreoli said.

Some organizations are taking the opposite approach, refining their requirements down to the component level and decoupling management functions from the hardware. HP has its finger in that pie, too, Andreoli said, through a joint venture with Foxconn Technology Group that focuses specifically on producing the kind of workload-optimized bare-metal servers that the top providers demand. He said that his company acts as a broker between the Chinese manufacturing giant and customers, handling logistical aspects that a direct relationship would place on the shoulders of the buyer.

“I don’t know what the future will be, but I want to be ready for the future and I want my customers to be,” Andreoli explained. “If the open designs become dominant, we are ready, we are working with the key players there, we have an offering– we are ready to go.” But HP’s vision doesn’t stop there. The company is preparing not only for the day when the hyper-specialization model reaches the enterprise but also the future beyond silicon, which it’s trying to address with a project called The Machine.

The bold initiative is an evolution of converged infrastructure that aims to fuse compute and storage one level below the chassis on the motherboard. The goal is to eliminate the need for organizations to mix and match different kinds of hardware in order to optimally run multiple workloads in the same environment, Andreoli said.

“There is a wave of reunification over the horizon,” he predicted. “Over time, you’re going to have a self-governing body that we will load into applications so that they self-optimize and you don’t have to buy different servers. That’s what we’re working on with The Machine.”

Watch the full interview (19:30)