Hyperscaling enterprise datacenters through logical aggregation

Enterprise data centers can't compete with hyperscale clouds, mostly due to labor costs, which are as much as 70 percent of EDC expense. But what if enterprises could logically scale far beyond Amazon and Google? That's what Maxta - among others - is working to do.
Written by Robin Harris, Contributor

Yes, the cloud providers have grabbed IT's mindshare for the last decade with their mind-boggling scale, PhD-heavy staffs, big data analytics, and labor costs as low as 3 percent of total expense. Despite that though, they still account for a small fraction of total IT purchases. The vast majority of IT sales go to governments and businesses, not cloud providers.

Which means that if enterprises logically aggregated their infrastructure metrics using machine learning and big data analytics, they could achieve an effective scale dwarfing Amazon's AWS. That's what Maxta, a Silicon Valley firm, is taking the first steps to enable.

Cloud vs on-prem

Cloud vendors deploy billions of dollars worth of new hardware every year. Their systems are optimized for their application needs and custom built. They claim that because they buy in volume direct from manufacturers, their hardware costs are some 30 percent below what enterprises pay.

But that doesn't include the healthy profit margins that cloud vendors enjoy. Nor does it factor in the non-financial costs of cloud services, such as higher latency, and the considerable switching costs.

Hyperconverged infrastructure (HCI) is the enterprise's answer to cloud vendors, using many of the same ideas, but in bite size chunks. And that is where Maxta is focusing. This week they announced Maxta MxIQ, which they say

. . . combines configuration metrics with visibility into capacity, performance, and system health trends across data centers and clouds to provide a granular overview of customers' IT environments. MxIQ leverages metrics and trends across the entire Maxta customer base to recognize potential impacts on system health and availability to enable proactive technical support.

(emphasis added)

This isn't a new idea. HPE's Nimble Storage unit built something similar starting almost 10 years ago, which HPE is extending to their other storage and server products.

But Maxta's focus isn't just storage, but on the entire health and efficiency of HCI infrastructure. They're working on incorporating machine learning into future versions, so they can automate even more management tasks. Today though they are helping major server vendors, such as Lenovo, provide diagnostic and preventative service to tens of thousands of customers, using cloud-based analytics.

The Storage Bits take

Scale changes everything in IT. Warehouse size computers blew the doors off the traditional enterprise vendors and their fiddly high-margin systems. Now the smart money is moving to incorporate the architectural concepts of cloud systems into products sized and priced for the enterprise.

The bad news for IT workers is that outsourcing the intelligence required to manage critical infrastructure means that the remaining work will be much more focused on integrating new IT services with enterprise needs. IT techs will be primary care doctors, not highly specialized - and paid - tech gurus.

The key takeaway is that the IT arms race is no longer only between AWS, Azure, and Google, but now includes companies that are building the lessons of the cloud into products for enterprises. Choice and competition are a good thing, and I'm pleased to see companies like Maxta and HPE driving to bring the advantages of scale to the enterprise.

Courteous comments welcome, of course.

Editorial standards