Gartner Symposium: Time to get rid of the dead wood

0
36
An encouraging new conversation around sustainable IT, says Nordic CIO

Source is ComputerWeekly.com

The idea of lifting and shifting workloads into the public cloud is never going to deliver a good return on investment. Instead, according to analyst firm Gartner, IT leaders need to focus on consolidating applications. In particular, CIOs need to “get rid of the old wood”, Philip Dawson, a research vice-president at Gartner said in a podcast interview with Computer Weekly

Dawson, who presented a session called The Future of Infrastructure Is Distributed and Hybrid at the Gartner Symposium in Barcelona, said cloud management of IT infrastructure is key. He said this becomes increasingly important as IT leaders choose to keep or rehost elements of their enterprise IT platforms on-premise.

As IT becomes more complex, Dawson urged IT leaders to rationalise or retire applications. “Get rid of your dead wood and only modernise the stuff you really need to modernise,” he said. “If it has high IT complexity and low business value, get rid of it. You don’t need that cost.”



This also applies in the cloud, where applications have been moved to public cloud infrastructure without first being optimised to take advantage of the elastic compute and platform capabilities available from public cloud providers.

A distributed hybrid infrastructure incorporates cloud-native infrastructure principles, such as programmability, elasticity, modularity and resiliency, and can be deployed and managed in any location the customer chooses, including on-premise, at the colocation, at the edge or in the public cloud. For Dawson, this means having a single control plane that can manage the system in any location.

Dawson said modernising and refactoring applications is far easier to do on-premise than in a cloud environment.

His presentation identified the main companies offering on-premise to cloud capabilities as IBM, with IBM Cloud; Nutanix, with its Cloud Platforms and Cloud Clusters offerings; and VMware, which includes a number of cloud products such as Azure VMware Solution, Google Cloud VMware Engine and VMware Cloud on Amazon Web Services (AWS).

A company like Red Hat, which IBM acquired in 2018, is pitching its OpenShift platform as an environment for virtual machines and containers. Dawson regards the tactic Red Hat has adopted with OpenShift as being focused on customers looking for application modernisation. But such platforms only deliver tangible benefits when applications are refactored. As Dawson points out, moving a workload from a virtual machine into a “fat container” is never going to deliver the efficiencies IT wants to achieve. Moving to containerisation requires “a big refactoring of applications”, he said. 

Even if there is an ambition to containerise 30% to 40% of new applications, Dawson said: “You’ve also got to think that there’s two-thirds of stuff that need to be modernised such as legacy Java.net and other legacy environments.”

This means IT departments will end up with a heterogeneous IT environment comprising some private cloud, some public cloud and some cloud-native environments, along with business application clouds. On top of this, datacentre equipment providers are selling equipment on a pay-per-use basis, with the ability to extend this into their own cloud or to public cloud providers. In addition, the likes of Alibaba, AWS, Microsoft Azure, Google, Oracle and TencentCloud offer public cloud to on-premise capabilities.

As Dawson points out: “You not only have multicloud heterogeneity, but multiple platforms plugging into common infrastructure. Comparisons not only of the infrastructure, not just the platform itself, but the software sitting on top as well.”

This is the reason he recommends that IT leaders look closely at which applications to refactor, which ones should be moved to the cloud, and then identify applications that should eventually be discontinued. 

Source is ComputerWeekly.com

Vorig artikelThe future of cloud: A Computer Weekly Downtime Upload podcast
Volgend artikelStorage explained: Consumption models of storage procurement