BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Capgemini: The 'Cloud-Native' Age Is Now

Following
This article is more than 6 years old.

Capgemini

Technology consultancy and outsourcing specialist Capgemini would this month have us believe that its latest research project has identified a significant rise in cloud computing adoption. More specifically, Capgemini points to a what it claims is a tangible rise in so-called 'cloud native' software applications and services i.e. those apps and online services built specifically for an online cloud-connected existence.

"[As many as] 15 percent of new enterprise applications are cloud native today with adoption set to increase rapidly in the next three years, jumping to 32 percent by 2020," asserts Capgemini.

What is a cloud-native application anyway?

But what really is a cloud-native software application in the first place and what do we mean by this term?

SVP for enterprise cloud services at Capgemini Vikrant Karnik explains that a good example would be an application that is updated often, or one in an area where a firm is trying out a new business model. Karnik cites a specific Capgemini example of a new UK 'disrupter bank' trying to create a fully digital banking experience. The bank in this case needs to perform a lot of A/B testing (just to find out what works and what doesn't). Because of this it needs to have the ability to add new features or roll them back quickly. This is the kind of app that is perfect to be build cloud-native.

Cloud-native allows risk taking 

"Equally any apps that need to be able to make 'live' decisions and calculations quickly are good for cloud-native. For example, apps that are focused on payment processes need to calculate interest rates in real-time. Organizations using these kinds of apps are more nimble and as a result, they are able to take more risks, drive business model innovation and generate more revenue," said Karnik.

Other prime examples might be apps that require deep data analytics. Capgemini works with a North American bank that is ingesting a huge amount of data at up to a terabyte a second (combining everything from social media moods, onward to the weather etc.) then calculating ‘risk’ in real-time for the prospect or current customer.

However, its not all plain sailing. Capgemini's research suggests that many CIOs are facing challenges in building business cases to invest in cloud-native apps from business leaders that see cost reduction as the priority for IT teams. These challenges range from the organizational, including battling an ingrained culture that is opposed to the nature of cloud-native working (65%) and a skills shortage when developing cloud-native apps (70%), to the technical, such as difficulties integrating with legacy infrastructure (62%) and being locked in to vendor contracts (58%).

How do we build cloud-native software?

If we accept the suggestion that the cloud model might catch on then, how should we build software applications to run in cloud? Do we just build normal software and 'shove' it up there? Actually it's not 'up' it's 'to' as in 'into' a cloud datacenter, but you get the point. Does any traditional terrestrial non-cloud software 'just work' in cloud?

In terms of shape and form cloud-native apps are different says Karnik, "The kind of memory, CPU and storage drastically changes. For example, the bank we work with that processes a terabyte of a data a second, they will need a huge amount of computing power to get through that and then to ingest it and put it in the right place. Once it moves to processing, the computer needs to figure out what is actionable and what isn’t worthwhile – at that point you need a lot of Input/Output (I/O) muscle and computing power. Once you’ve weeded out the data that isn’t useful, you’ll feed it into an Artificial Intelligence model and then the computing need changes again. If you tried to do that on-premises [i.e. not in cloud], it would be a nightmare for your IT staff. This is exactly when cloud-native apps help. You can match the need to the task and in turn, save money."

For all the technology associated with cloud computing, Karnik gets touchy-feely and says that cloud-native is not just about technology i.e. it’s about people, process and technology together. This is because, successfully implemented, cloud technology enables people and processes. If a customer is doing a true cloud-native cloud deployment, Capgemini says that automation deployment functions should be at least 80% or more of the total project.

Cloud + open source = a thing of beauty

Karnik is upbeat on the nature of cloud computing and its intersection with open source. He says that the 'beautiful thing' about cloud-native is that the whole technical trend is tied to the rise of open source.

"What people don’t realise is a lot of the start-up community, like that in Silicon Valley, is based on a sharing of knowledge and it contradicts how the big tech companies were initially built [through the sixties and onward], where they were focused on protecting IP. With the tie between cloud-native and open source, we’re also seeing a related burst in Artificial Intelligence and machine learning. Now there is a mechanism for people to open source their learning framework and give feedback. This has completely changed the game and allows companies to really innovate," said Karnik.

The bad news about cloud computing

Pushed for a negative to balance the PR-fuelled 'cloud is just dandy' feel-good factor, Karnik hedges his bets and says that the only complaints he sees surrounding cloud native adoption is that it all just runs so fast and so well.

This, if anything, means that firms can't always get their entire operation running cloud native overnight.

This, in turn, means that firms end up with a portion of the business running cloud native and a portion of the business sitting on in the 19th century on an old framework.

"The speed disconnect [here] causes a lot of clash internally," said Capgemini's Karnik. "The other problem is focused on information, particularly around how companies ingest and process all of the new data and how they integrate this into their IT system. We are seeing new companies try to bridge the gap between legacy and cloud native systems, but it is still difficult for many large organizations and few are driving meaningful change here."

The cloudy road ahead

Looking at the wider picture then, Capgemini says its study (involving around 900 senior professionals working in both IT and business from 11 countries across Europe, the Americas and Australia) has found that where adopted and implemented successfully, cloud-native applications almost always help increase revenue and reduce operating costs.

The company also notes that the cloud skills gaps and the whole 'technology cultural change thing' are still the two biggest obstacles to cloud adoption.

Perhaps we can continue to get firms like Capgemini talking about the specifics of how cloud applications work and what shape they really take. In this way we may possibly help demystify some of the real guts that go into the cloud model.

It's not a cloud. It's a server running a highly-connected software application with specific storage and networking demands in a datacenter connected via a pipe that we usually call the Internet.

There's we've said it. Do you feel better?

Capgemini

Follow me on Twitter or LinkedIn