This is the next in our extended series of 'controversial questions.' The eleventh, we think.
Here's the question we put to a significant number of industry executives:
Everything happens in a combination of waves and cycles. For instance what is now cloud computing used to be time-share. Similarly there has been a constant yin and yang between centralised and decentralised computing - the early mainframes were sidelined by small departments buying minicomputers and then PCs in order to regain some control… and then they were aggregated into networks. But resistance followed and groups isolated again. Now we're looking at a major aggregation with cloud computing. What does the assembled throng think might cause the eventual waning of this waxing cloud hegemony?
Clearly, as technologies evolve, some systems will be able to advance to the new paradigm while others are left behind. Think of how few silent movie stars were able to make it in the 'talkies.' Mostly, we understand, because they had somewhat 'odd' voices that weren't easy on the ears. There's a more modern version of the same thing enshrined in the 1979 song "Video Killed the Radio Star."
However, lets pay a little more attention to the thoughts of the industry executives.
Setting the scene, Daniel Markuson Digital Privacy Expert at NordVPN tells us that "Everything happens in a combination of waves and cycles. To be more precise - in a spiral. Nothing dies in the computing world. Only new abstraction layers stack up.
"Computers were built into networks. Then networks were connected, content delivery systems were introduced, household devices that rely on different operating systems became interconnected and synced with each other over the cloud to provide a seamless and effortless experience. People won't stop to value comfort more than anything, not to mention that there's still a lot of space for improvement left. Industrial digitalization and the growth of developing countries are also vital factors for a cloud to get stronger and evolve."
Michael Warrilow, VP analyst at Gartner agrees, "Cloud computing is just the latest to-and-fro in the continual merry-go-round of centralised vs decentralised computing models."
Thor Essman, CEO & Co-Founder of Stax, adds "Ask to define cloud and you get more answers than asking what's the meaning of life? However, one thing all definitions will likely agree on is that moving to a technological capability of buying and consuming vs building and creating is a trend that will probably outlive us all."
Tim Hope, Chief Technology Officer at Versent, continues the same theme, noting that "As with the shift from Mainframe to PC, to networked PC and now cloud, it's less of a wave and more of a paradigm shift to a new technology platform - with new capability unlocked through the shift in approach. Through these shifts we haven't moved backwards to mainframes - but we have iterated on the new platform to introduce efficiency and the ability to match compute resources directly to the business outcomes and customer demand, in an ever increasing abstracted and distributed fashion."
"I've personally witnessed the cyclical evolution between centralised and decentralised computing; says Robert Merlicek, CTO - APJ at Tibco. "There's a constant push and pull between computing aggregation and segregation, which come and go in waves. Right now we're in a phase where companies are actively using cloud applications to accelerate digital transformation - especially by integrating systems, processes, APIs and people. As such, it's true that the cloud as it currently exists represents a broad move towards more centralised computing."
Offering a slightly alternative suggestion, Lee Thompson, Managing Director A/NZ, Nutanix thinks the swing is really only on one direction. "I believe cloud's mass aggregation is in full swing - this is evident in the breakaway from 'public cloud-only' and even 'public cloud-led' environments, in favour of the multi-cloud model. Enterprises don't want to be led by an IT environment; they want to lead with business imperatives through applications with access to any cloud, anywhere, anytime depending on their needs.
"Both sides of cloud show no sign of slowing down or being replaced. Indeed, cloud is vital to lead the digital economy-focused recovery the world needs. Even during the steepest economic contraction in modern history, we saw a swell in enterprise cloud investment driven by new ways of working and living. But the ratio will continue to shift in favour of private-led multicloud, as public cloud repatriation continues to be seen in A/NZ."
Michael Ewald, Director of Engineering, APAC at Contino is pressing forward in support of all things cloud: "Hybrid cloud, multi-cloud and all that cloud computing entails, will continue to be at the forefront of digital transformation projects and initiatives."
Similarly, "All eyes are on the cloud as the next de facto standard for IT systems," Brent Doncaster, Senior Director, NetFoundry tells us. "For the next decade or so we can expect cloud to continue to dominate IT strategy as the hyper-scalers build out their services to provide complete geographical coverage and continue to invest billions in the networking and compute infrastructure needed to support growing demand."
"When the cloud conversation started," Matt Cudworth, Cloud Practice Lead / Cloud CTO APAC, DXC Technology reminds us, "it was all around capabilities gained from the public cloud - someone else's server - but has really shifted to a full Enterprise Technology Stack approach when it comes to how companies orchestrate their technology. We're now seeing cloud technology emerge on premises and in computing, with cloud infrastructure being deployed at telecommunications exchanges to get the cloud closer to end customers. Cloud technologies are going to be everywhere."
"There are a number of forces driving this, Thompson reminds us, "sticker shock from large public cloud is a huge factor with budgets under pressure and cloud workloads increasing. Data sovereignty and security concerns have also hit an all-time high and this means it's never been more pressing for organisations to get control of their data. Ultimately, the right multi-cloud strategy gives the power back to the customer and ensures they can easily move data to the appropriately secure, scalable and efficient environment at any time."
Markuson wanted to take a couple of steps back to set some context. "Since the early beginning of computing itself, machines were large, simply because no one could make computers as compact as they are nowadays. Becoming more diminutive and more affordable, you could hardly find a place or a household without a PC, handheld, or IoT device. But massive computers are still present—only much more powerful.
"The term cloud by itself is very abstract and may differ regarding the context of where it's mentioned. At first, the network guys started drawing clouds on network diagrams where the cloud means something abstract and irrelevant in the context, where the details on how components are connected don't matter.
"Nowadays, cloud computing is even more abstract. There are many different use cases for them, such as laaS, PaaS, or SaaS. The latter provides the subscription-based service for the end-user, which doesn't care how the service functions under the hood only; it seeks the result of the service for which one pays.
"Nonetheless, big machines and computer clusters didn't go extinct and are still used for various purposes as science, movie making, etc.., you may find a personal computer in almost every home. Various government institutions, hospitals, and banks still own their own data centres and servers. Such servers need to meet regulations regarding how the specific data is being stored in them."
Nothing much will change
Many of the executives responding to the question were firmly of the view that 'cloud' was here to stay (in whatever guise it took - there were many!).
For instance, Markuson opined, "Cloud computing is still relatively new, so the end of it is pure speculation. However, I don't think it is likely to change radically in the foreseeable future. The amount of the existing data is growing with each second, and so far, we can only observe its expansion."
"While it's impossible to say what might happen in 50 or 100 years, over the next 5-10 years cloud won't be going away in a hurry, and it will be fascinating to see how the industry will evolve," Daniel Bradby, Emerging Technology / Innovation Principal at Mantel Group suggests. "The big three cloud providers are still in a rapid growth phase and are being challenged to differentiate themselves. They need to focus on their unique value proposition, or they risk becoming commoditised. We'd like to see each of the providers playing to their strengths a bit more. For example, Google needs to leverage its expertise in data and AI. Microsoft has a rich history of identity services and active directory it could be focusing on."
Essman agrees, " 'Cloud' will never end, and while the cycles of transformation may shift, we believe they will shift UP the technology stack."
Rob Bryant, Executive Vice President, APAC at InEight agrees, noting that "Hardware continues to evolve at pace, Moore's law reigns true, yet the volume of data we touch and draw from is also growing exponentially. As I see it, that means we will continue to find cloud a convenient way to store and process data, that we access, refer to and use anywhere at any-time. Providing we are connected, and improved mobile connectivity sees that remains the case, cloud computing has a solid place in our digital life.
"In a 'subscriber society,' " he continues, "I believe we have already moved on from the notion of having data in a place we can point to it or carry it under our arm. At the point where individuals are happy to trust cloud for the family photo album and music collection, I think data for the purpose of business decision making and corporate knowledge is very comfortable in the cloud."
Further, Markuson remains positive. "It's better to talk about the development of the cloud rather than the waning. It's no secret that people can get spoiled sometimes; one can hardly imagine a website that takes a second to load. This is where CDNs came into play, which caches static information as close to the end-user as possible, improving the experience for the consumer. That's not the case with dynamic information and tends to change for various reasons, e.g., video games, traffic management, multiple sensors, and many more. This is where edge-side computing comes into play and makes so CDNs cache the already known information and enable them to receive, process, store, and send the latest fresh, dynamic information." We discuss edge computing in a later section.
Offering some local context, Ewald reminds us that, "The digital economy remains high on the Australian government's agenda. With the recent announcement of the $1.2billion investment on the Digital Economy Strategy, an even greater number of critical systems will go into the cloud. This means that the concept of centralised computing will continue as the cloud evolves into software as a service, functions as a service and data as a service whilst infrastructure as a service and platforms as a service will remain the domain of the hyper-scalers - something we are starting to now see in across the industry."
Further, he notes that "The cloud can enable many businesses to adapt and find new ways of operating and is not going anywhere. In fact, there are four core benefits of adopting the cloud to achieve innovative and tangible outcomes continue to be part of the main reason for businesses shifting to the cloud:
"A funnel for innovation: Businesses armed with a well-developed cloud strategy will not only be able to leverage the cloud effectively and at scale, it also acts as a catalyst of meaningful innovation for end-customers.
"Security: The cloud has become a large and highly attractive attack surface for malicious actors who exploit poorly secured ports to gain unlawful access to valuable data. With enterprise-grade cloud protection from targeted attacks and data leaks, this ensures that there is enhanced data protection with encryption at all layers so that the business is from threats.
"Optimised operations: Cloud technology offers a flexible operating model while allowing organisations to save time and money, improve operations, and encourage innovation. A physical server typically takes days or weeks to become fully operational, but the cloud only takes minutes. The faster a business goes to market, the faster it starts seeing revenue, and profit.
"Data Gravity: Data is the true enabler for any digital transformation so where the bulk of your data will reside will drive a lot of subsequent and dependant architecture. As more data is collected, the cost model for on prem vs public cloud suits leans favourably toward cloud. On prem, compute is cheap, but storage is expensive, and it is the inverse model for public cloud. Plus, with new serverless models, cost is more on a transactional basis so good data architecture lends itself to cost optimisations in TCO towards cloud. When more and more organisations move to cloud and amass large volumes of it as they unlock the benefits of analytics and AI/ML, the appetite to move back to some sort of on prem model in the future could be cost and risk prohibitive."
Of course, not everyone thought that the cloud environment would grow forever - and there were even a few with an each-way bed on proceedings.
Markuson, for instance had previously told us of his support for the cloud, however, in a somewhat contrary comment, he offered, "Usually, change comes when the current infrastructure starts to become insufficient to satisfy the future's needs. For example, a significant part of digital innovation relies on AI development, and AI training relies on private or public datasets, where both input and output must be stored somewhere. Quantum computing, among other things, will reform the AI learning process by changing the fundamentals of data processing. Once we adopt it, we will enter a new realm of computing, in which the current cloud, as we know it, might become insufficient; however, this will way more likely lead to evolution rather than waning."
"In addition, we expect to see a focus higher up the cloud stack," Bradby added. "We can already see AWS trying to move away from the lower-level stack but it's in their DNA. Customers are going to be less interested in the nuts and bolts and see cloud as a more general-purpose platform for application deployment (as Salesforce is demonstrating). We also expect to see more of a push back to open standards and interoperability over the long term and more services around the edges of the cloud stack being provided by other vendors, such as Snowflake's data warehouse offering.
"Lastly, we're seeing a trend further down the stack for custom chips. The big 3 cloud providers are providing custom chips to deploy in their own environments and with Apple and Nvidia playing this space it will be interesting to see how this develops."
Offering a modicum of context, Essman suggests that "Whereas the last two cycles of cloud have dealt primarily with foundations and infrastructure, as we move above that layer there are decades of technological cloud enablement to come in the form of SaaS, PaaS, IaaS, and dozens that we have yet to invent. As you can see with all the major cloud providers, they are focusing their time providing ecosystems of solutions and products, and have well and truly sped past the "cloud infrastructure only" cycle."
Forget the hardware
Vercent's Hope took a slightly 'parallel' view by advocating for the 'no hardware' view of things.
"When customers start to move to the cloud, they start with a mindset around how to get their servers integrated and configured. This moves very quickly to conversations around Containers, Serverless and SaaS solutions - the server is being forgotten. Cloud isn't the next step in having a better server, it's the next step in not having a service - not having to think about the underlining servers and infrastructure. It's enabling people to work with technology directly to link it to the business problem - removing the underlining infrastructure, effectively commoditising it out of the value chain. The next step in this path is further towards SaaS offerings that are highly configurable to meet different customer requirements and market innovation. I would expect us to see capability growing around how we see highly configurable SaaS capability that has the ability to reach out of the cloud data centres and control IoT, wearables and services located close to the end users.
"The challenge with the ease of procurement enables customers to procure complex services on a credit card - this is far more powerful and simple that buying a minicomputer. Small departments can build highly functional and complex systems without any engagement [with] the technology teams - this introduces risk when security, supply risk, compliance, data and other key non-functionals are not managed or reviewed appropriately. The sprawl of business units creating services in different cloud ecosystems makes it difficult to integrate and understand where the business data and systems are actually being run. No longer can you point at a data centre and be sure all your data is in that building."
The cloud, but…
Beyond the 'no hardware' view, there were also comments that postulated a 'cloud plus more' view. For instance, Gartner's Warrilow opined that "What replaces public cloud is distributed cloud. It allows the customer to choose where to locate the cloud service, but still using the innovation provided by cloud computing."
Similarly, Deepak Giridharagopal, CTO at Puppet adds, "The cloud is only superficially a force for aggregation, but at its core it's the opposite. The cloud doesn't behave like one big computer. Rather, the cloud disaggregates the building blocks of software and serves them up à la carte. Users have a ton of choice (many CIOs would say too much so) in what cloud services they want to use, from which vendors/parties, and in how they assemble those pieces together to architect their applications. This wasn't really practical, not to this extent, before cloud computing.
"With the "cloud hegemony" rolling out a dozen new services every year, it's no surprise that the experience can feel sprawled out and incohesive. That my bill comes from a single company is of little comfort.
"I understand the desire to standardise and consolidate - the struggle is real, and reducing choice is _a_ way to bring things under control. But in my opinion, a better approach lies in smarter tools that abstract away the complexity of the cloud. Good infrastructure abstractions let you treat diverse, underlying technologies in a simpler, unified way. We're already seeing interesting, new abstractions and architectural patterns emerge from the cloud-native open-source community, with more yet to come.
"The industry must develop smarter tooling and better abstractions. The cloud is chess, not checkers. The more complex the cloud becomes, the more its fundamental promises of being cheap, easy, and flexible come into question."
Many of our respondents were quick to suggest 'edge computing' would be the evolution from current (relatively) monolithic cloud environments. Such implementations have a lot of dvantages.
Merlicek offers the thought that "Cloud is still evolving - it hasn't yet reached its peak. For instance, we're at the stage where multi-cloud has arrived and will continue to grow as standards evolve. But innovation is unrelenting, and forces change, and the next major change we'll see is the rise of edge computing.
"Edge computing will be a shift to decentralisation once again. With so many edge devices now being connected and providing information - whether it's the Internet of Things, smart autonomous vehicles, smart cities, or any of the multitude of smart consumer devices out there - we will see a massive rise of machine-to-machine interaction.
"The concepts of IoT, machine learning and AI leverage a great amount of real time data, much of which is not generated in the cloud but at the edge. To make real time decisions on this data, the traditional model of moving it all into a central cloud for processing is just not practical. Instead, we will see the computing nodes for transactions decisions move ever closer to the edge. The rise of Industrial IoT will drive this very quickly.
"As an example, think of self-driving cars. As these new vehicles proliferate, they will feed off each other's data. The problem is that this sharing will need to happen in real time to make useful decisions. For example, computing will need to happen almost instantaneously to determine how two driverless cars should avoid a collision. Autonomous drones are another useful example - imagine hundreds or even thousands of these drones occupying the same airspace. They would all need to update each other on positioning, location or flight paths, or share information on collision avoidance with birds. All this computing will need to take place at the edge, in innumerable and fleeting 'edge-clouds.'
"I see ultra low latency (such as low single digit milliseconds or less) and scale being vitally important in the future and that is where real time edge computing will flourish. These Interconnected 'edge-clouds' will provide the backbone for a new generation of machine-to-machine mesh interactions enabling a whole new suite of capabilities."
Mark Jobbins, vice president and CTO - APJ at Pure Storage adds that, "The growth of the Internet of Things (IoT) and self-managed edge technology will push compute and data out to satellite locations. The practicality of transmitting all data to the central location will force the decentralisation of technology again to be able to provide the independence of this technology. 5G and other advances in networking will support and accelerate this."
Pointing to another favourite 'hobby horse,' Markuson notes that "We believe that with 5G, there will be an increase in the need for such edge computing, as the consumer does care about the result of the service. And all of these computing actions are being done in the cloud, wherever the cloud is. Whatever the cloud is."
"Cloud infrastructure is evolving quickly to position cloud capabilities at the "edge" to support running applications close to where data originates in the physical world," Doncaster concludes.
He adds, "The odds are that cloud capabilities will even be delivered from outside the Earth's atmosphere. But after they've conquered space, the cloud providers will eventually run out of room - not physical or virtual space, but the laws of physics that determine how many bytes can be squeezed onto a silicon wafer and the speed at which electrons can be propelled over wires or through the atmosphere."
Extrapolating upon his previous comments, Doncaster wonders "So, what replaces today's cloud? The answer is most likely to be quantum computing, which is still in the realm of theory and early-stage experiments for now. Like supersonic flight, the barrier to quantum computing will be broken, unleashing previously undreamt-of power, but raising some interesting questions. Among the biggest is who dominates the provider landscape.
"Social media, internet and cloud have seen power concentrated in the hands of a small number of very large for-profit organisations - Facebook, Google, Amazon, Alibaba, and Microsoft. In a world struggling to come to terms with the economic, political and social influence of these technology giants and with geopolitical forces pulling back from globalism, what might this mean for the development of quantum computing?
"One possible scenario, given the huge investment needed and strategic value of controlling the world's IT - which is to the global economy of the 21st century what oil was to the 20th - is that governments may seek to take control of quantum computing infrastructure. Could the quantum cloud of the future be run not by AWS but by Washington or Beijing? Might quantum computing power fuel a new wave of nationalism and/or geopolitical alignments? Is this the next space race whose ostensible purpose is to expand human endeavour beyond the Earth's atmosphere but whose real purpose is to become the "winner" with technological and economic competitive advantage delivered on the planet's surface?
Finally, we give everyone a chance to offer some concluding remarks.
Essman starts, "Our prediction is we will see cloud evolve as the primary form of technical enablement, like the industrial revolutionary shift from producing power individually to consuming it. The future is up the stack, in cloud."
Hope continues by observing that there is already a massive barrier to entry for any new players, "The pace, capability and functionality required to compete with the major cloud providers will make it difficult for new cloud infrastructure providers to enter the market. It's important that the environment continues to have a critical mass of cloud providers operating in the wider ecosystem to ensure that different business and consumer requirements and values are serviced."
InEight's Bryant adds that "Personally, I see cloud providing enormous opportunity for vast volumes of data to assist in augmented decision making and augmented reality, drawing on machine learning and pooling of databases that cloud provides. In the business world, security and integrity of data is critical. Being able to outsource those responsibilities and operations to trusted hands allows business in any sector, construction, law, retail or medical, to focus on their core value and IP.
"If that is working, I don't expect to see an appetite for business to go back to buying and maintaining server hardware and staff working in 'IT black ops'. With an increasingly remote workforce that decision is almost removed. Cloud has become the great enabler. "
James Bergl, Regional Vice President, ANZ at Datto summarises the many advantages of cloud silutions. "It's difficult to debate the many benefits that cloud brings, such as scalability and agility, which is why is has proven to be so popular within the enterprise over the last decade. But cloud has shown to be particularly useful from a business continuity and a backup perspective. In the event that a primary server is compromised in an attack, a clean backup image will be immediately available in the cloud. This enables normal business operations to continue (on a virtual machine) while the primary server is being restored. Having this kind of failover has become critical as businesses face increasing volumes of sophisticated cyber-attacks, especially ransomware where hackers will steal and encrypt your data making it impossible to restore unless you have reliable back-up solutions in place."
Looking into the future, I believe the process of cyclical evolution will continue to play out and we'll again begin moving towards more decentralised forms of computing," adds Merlicek. "However, I don't think this decentralisation will see us move away from cloud. Instead, and maybe a little counter-intuitively, it will see us moving even deeper into cloud technology."
Ever the optimist, Markuson opines that "The only way for clouds to go extinct is to lose networks or for them to become very unreliable. As we talked about before - the cloud is something abstract and irrelevant in context. Without getting to the cloud and between components in the cloud, it becomes very relevant and not abstract. Then this is not a cloud anymore. But in such a case, humanity will face much bigger problems than to talk about the waning clouds."
Cudworth wants us to look further forward, "As we head towards what we call the 'pervasive cloud' stage of the cycle, I predict we're going to see a real plethora of emerging technologies suddenly become accessible. As cloud matures, its capabilities will become almost self-service, enabling new technologies to be adopted much more quickly throughout the business. At the same time, companies are rewiring for innovation - no longer willing to spend a million dollars on hardware for five years - and so new technologies will be much more fluid, coming and going as they meet immediate and future needs."
However, it seems fitting to give Doncaster the final word. "Finally, a caveat about our powers of prediction. The first supersonic passenger airliner went into service in 1976 amid claims of a revolution in high-speed air travel. Forty-five years later, there are no commercial supersonic services operating."
So, to summarise, the majority of our executives are quite sure that this cycle is different; that the cloud trajectory will be in place for a very long time and that it will give rise to and encompass many other technologies - edge and quantum computing, various shared AI and so on. A few also pointed out the commercial benefits of moving to the cloud as opposed to maintaining major devices in-house. Oddly, though, no-one touched on the impact security might have - either as a positive or a negative.
We'd love to hear readers' thoughts on this.