How to deliver the benefits of the public cloud experience on-premise

In the hybrid cloud era, CIOs are increasingly targeting cloud-like infrastructure in the enterprise, or on-premise.

The DCIaaS (Dedicated (Local) Cloud Infrastructure-as-a-Service) market will be worth $14 billion by 2025, according to the IDC.

ESG Research estimates that 60% of those who have moved to the cloud are now looking at bringing certain workloads back to the enterprise, and the report adds that 46% will invest in bringing the cloud experience on-premise. This is because there are datasets that don’t belong in the public cloud, such as servers in manufacturing or 5G towers, or in hospitals and banks. Regulation requirements and security challenges mean that this data must be stored on-premise or on the edge of a network.

There are also huge cost savings that can be achieved by moving workloads back to the enterprise.

Speaking during The IT Press Tour to San Francisco, Siamak Nazari – CEO and co-founder at Nebulon – compared the experience to renting a car. Deploying all workloads to the cloud is viable and cost effective in the short term – it helps companies scale – but in the long-term, this is an expensive route. He did note that there were workloads that absolutely have to be stored on the public cloud.

Research from Andreessen Horowitz says that enterprises can save 50% in costs, as a result of cloud repatriation. As an example, significant costs can be achieved by removing potentially thousands of virtual machines in the enterprise to on-premise, which consume energy even when idol. When these machines are all turned on at once, this intensifies the cost problem.

>Read here: To find out more about the other companies in the latest edition of The IT Press Tour in San Francisco and Silicon Valley

Adopting a cloud model in the enterprise

The challenge is producing the public cloud experience on-premise – something agile, elastic and flexible – that delivers hyperscale infrastructure.

“Enterprises have to adopt a cloud model in the data centre to stay competitive, but they can’t – the technology isn’t mature,” said Nazari.

The standard AWS cloud operation platform for hyperscale infrastructure consists of three areas, which need to be added to for the cloud-experience to be viable in the enterprise.

1. AWS management console: automation, non-disruptive ops and instant updates.

Enterprises need to add Zero Trust to the SaaS delivery model, which is beyond what cloud providers can offer.

2. AWS nitro system: offloaded enterprise data services, secure platforms and isolated fault domains.

On-premise these services need to be available with any supplier of choice for the flexibility that cloud providers can’t offer.

3. Amazon machine images: consistent deployments, consistent maintenance, and application variety.

Immutable instance or updated servers need to be available from day 1, which gives the enterprise the agility and variety necessary to work on the fly.

This equips customers to operate their on-premises infrastructure like a hyperscale public cloud.

A key inhibitor to delivering these IaaS cloud efficiencies on-premises is overcoming the manual, server-by-server deep infrastructure operations in public and private environments.

Nebulon’s server-embedded, infrastructure software delivered as-a-service – implemented on a dedicated card that functions as an IoT endpoint – allows the enterprise to manage their deep infrastructure operations, such as BIOS updates, SSD firmware updates and component health monitoring.

Customer story – global cloud storage data centre

As an example of the service in action, the COO and co-founder, Craig Nunes, provided a customer case study.

He said: “In its current estate the customer has 10 data centres, 3-tier infrastructure with 250 VM servers and 60 arrays, equivalent of approximately 2 petabytes of storage.

“The project involved decommissioning the entire current environment, replace it with Nebulon-enabled servers and Nebulon’s infrastructure operations SaaS.

“We deployed nearly 100 enabled servers and over a petabyte of storage across four centres.

“The benefits they saw was 40% savings on infrastructure, higher margins on existing services and new customer acquisition benefits.”

See also: Have you traded on-premise lock-in for in-cloud lock-in? – Martin Gaffney, VP EMEA at Yugabyte, discusses why CIOs need to head off long-term lock-in by adopting as open an approach to their cloud strategy, as to their software choices

Avatar photo

Nick Ismail

Nick Ismail is a former editor for Information Age (from 2018 to 2022) before moving on to become Global Head of Brand Journalism at HCLTech. He has a particular interest in smart technologies, AI and...