Cloud computing has grown up. In the past two decades, the technology industry has fought against an occasionally brittle approach to security provisioning, with an increasingly sophisticated set of cloud functions beginning to emerge.
We now have access to a far more evolved notion of software-as-a-service (SaaS)-based technology. This is the era of ‘cloud native’. Natively built cloud services – which never existed in any kind of terrestrial version – are now coming to the fore. What factors, then, should you consider when working with cloud computing?
Adam Selipsky is CEO of Amazon Web Services (AWS). He’s been candid about cloud computing’s somewhat difficult adolescence: at the start, he says, it was expensive, slow and inflexible.
From the beginning, cloud vendors sought to achieve customer lock-in where they could, with interoperability far from an engineering imperative. That time has thankfully passed, and a new, more democratic and open approach to the cloud has prevailed.
“People used to question me about Amazon Web Services and ask what the connection with books was, but I think things have moved on,” said Selipsky, speaking at his firm’s December 2021 AWS re:Invent conference.
Vendors large and small are today putting a huge effort into making cloud services more consistent and easier to consume and use. Cloud specialists like to call this a simplified unified platform experience, but we could just call it cloud that works.
One of the companies talking in these terms is enterprise cloud specialist Nutanix. The firm has described its latest platform refresh as a simplified portfolio that brings together ‘rich product capabilities’ across private on-premises and public clouds. This aims to deliver consistent infrastructure, data services, management and operations for applications in virtual machines and containers.
Senior VP product management at Nutanix Thomas Cornely has said the firm has focused on delivering a simplified portfolio that delivers a consistent infrastructure with data services and management tools.
This theme resonates throughout the cloud vendor glitterati. It further echoes in the wider enterprise software platform space, such as Enterprise Resource Planning (ERM) or Customer Relationship Management (CRM).
It’s easy to see why it matters. Enterprise software is often too complex to purchase, let alone run; sometimes that’s down to acquisitions, or it could be due to product launch sprawl or relabelling existing services.
Understanding the cloud
The resounding message coming from the cloud industry is a promise. The vendors tell us they understand that customers need to work across private clouds and public clouds with a simplified application and data services product portfolio that is charged with more straightforward billing.
In practice, while cloud billing might be a dry topic, it’s a hot issue. Let’s remember, there is no actual cloud, as such. It’s just a global collection of server units housed in international data centres, running management software with various different optimisation parameters to make different virtualised cloud services behave and perform in different ways.
But how that’s all packaged and sold matters a great deal. The cloud industry has worked hard to simplify its packaging, metering and pricing structures. Take the use of so-called ‘reserved instances’, where customers agree to a specified amount of cloud based on knowledge of their own predictable IT workload requirement. That’s helped in some cases, but not in all.
Facebook’s Mark Zuckerberg famously complained that cloud is too expensive back in 2019. Still, the reality of cloud that can be turned on and off, providing the ultimate OpEx-only flexibility, is still something of a pipe dream.
Not everything is quite as perfect as the branding claims would have us believe. Even the most ardent evangelists typically agree that cloud perfection – and indeed cloud-native – is still a work in progress.
We should also remember that enterprise cloud is not just for Christmas. Customers need to think about ongoing supporting services, from load balancers to specialised accelerators to advanced security monitoring – and that’s just a start. Then there’s upgrades, testing and system maintenance: just like Windows, clouds need patch updates, too.
Making a plan
If cloud computing has been guilty of anything, it has been too shiny and new. The idea that we might all be able to tap into hyperconnected computing services built on virtual ‘machines’ running in cloud datacentres was always quite radical. Giving everyone massive computing power in the palm of our hand was a dramatic amplification of the internet’s initially promised freedoms. In short, it was a lot to take on.
The drive to cloud migration is of course one of the central moves repeatedly cited in the oft-discussed journey to digital transformation. But to make this journey, enterprises need a plan, a toolset and a capability arsenal.
Brent Schroeder is chief technology officer for enterprise-grade open-source solutions specialist SUSE. He says his firm has seen companies dramatically shrink their cloud migration windows. With the effects of the Covid-19 pandemic obviously still playing out, Schroeder reminds us that ‘traditional’ IT deployment windows of perhaps 90, 60 or even 30 days are now too long.
The way forward, says Schroeder, is process, planning and precision-engineered software production. SUSE insists customers work with private and public cloud infrastructure that is consistently deployed within robust security and governance guidelines.
In other words, organisations shouldn’t make the change to cloud without change management. As we make these changes, we can now shoulder a good deal of our total cloud software application lifecycle by using infrastructure as code (IaC), though that’s another subject for analysis in and of itself.
Cloud computing offers flexibility, scalability and changeable manageability, but only if we understand which direction the wind is blowing. The power of this weather system is still forming, so please wear a coat.