The cloud: present and future
The Chief Architect of Amazon Web Services is convinced that cloud technologies are in the early stages, that they are indispensable for companies and an essential prerequisite for IoT and artificial intelligence.
Text: Urs Binder, Images: Raisa Durandi, 19 February 2018
Glenn Gore is the Chief Solutions Architect for Amazon Web Services and is a genuine “shooting star” of the cloud. Australian-born Gore has worked on internet technologies since the 1990s. His stellar career began with the first retail ISP in Australia; he later had various positions with a backbone provider, a hosting provider and, from 2013, AWS. He now supports customers on their journey towards the cloud and helps them to find the perfect solution. At the World Web Forum 2018, we were able to talk to him about the status and future development of cloud computing.
Glenn, is it still possible for a company to ignore the cloud these days?
I don’t think that any organisation can give the cloud the cold shoulder any more. Companies have to keep abreast of this topic and develop a strategy for cloud transformation. It is certainly a journey – a journey that begins with moving individual applications to the cloud. At the same time, the skills and processes which make these transitions possible also have to be developed. Depending on the complexity, it can be a short trip or a long journey that takes several years.
What does “cloud” mean in this context?
It encompasses everything from IaaS to SaaS via PaaS. Ironically, some companies say they have no interest in the cloud; but then they talk about software like Salesforce... there’s a great deal of misunderstanding about what the cloud is and whether or not an organisation uses it.
Initially, “Private Clouds” were very popular. Now, the talk is mainly on “Hybrid Clouds” – do you also observe this development?
“Private Cloud” was always a completely inaccurate term, because the actual advantages of the cloud – its elasticity and the “pay-as-you-go” tariff model – are only provided by Public Cloud Services, supplemented where necessary by the company’s own IT system in a hybrid model. With a private cloud, you always pay for the entire infrastructure, whether you use it or not . Of course, old-guard providers did sell complete infrastructure packages as “clouds”; but in 2018, that has virtually disappeared.
Glenn Gore, Chief Solutions Architect of Amazon Web Services
It’s easy for a start-up to move directly to the cloud from scratch. But how does an established company with an existing IT infrastructure do this?
The best way is just to start it. Many companies get stuck in feasibility studies, and nothing happens. Take an application, a small team, and do an experiment. You don’t have to move everything to the cloud all at once, and you can learn a lot just by moving one thing.
Initially, the idea was to move processing power and data storage into the cloud. Nowadays, IT systems are built for the cloud from the start. Are the available tools adequate for this task?
Yes, there is a whole ecosystem of development tools, services, partners and software providers that can be combined. These include open-source tools such as Puppet, Chef and Jenkins, but also partnerships with commercial providers such as Atlassian and services by Cisco and F5, which companies use to manage security in the cloud.
On average, are developers ready for the cloud?
The role of the software developer is currently undergoing a transition. Developers and development teams have to become “cloud natives”, which means they have to understand what faster release cycles, microservice-based architectures and DevOps mean – for the organisation and corporate culture, as well.
Major changes to the job profile, constant flux – is software development an attractive occupation?
I think there's never been a better time to be a developer. Cloud Services give developers superpowers. They can build solutions, which were simply not possible in the past.
What does that mean for the significance of IT as a whole?
Not so long ago, it was normal for the IT department in an organisation to have a low status – the CIO reported to the CFO, and IT was outsourced and seen only as a cost factor. Now, companies realise that technology can make a fundamental contribution to their business success and bring competitive advantages. IT is now rising to the surface more and more: the CIO is directly subordinated to the CEO, and development teams are part of the company’s strategic future.
Are we currently already experiencing the cloud’s technological peak, or are we still at the beginning?
It's pretty much Day 1 of what customers can do with the cloud. New technologies and services are constantly being added, as you can see from AWS’s product range. We launched Amazon Web Services in 2006. Since then, the rate of innovation has been increasing each year: our platform had 1,017 significant new features and services in 2016, but over 1,400 in 2017.
Do customers really use these innovations?
When I talk to customers, they want to tackle more, and more diverse, IT problems using Cloud Services. They want us to simplify problems, which would otherwise be difficult to solve. Cloud Computing now encompasses lots of things which were unthinkable just a few years ago – for example, artificial intelligence, machine learning, language, image and video processing.
What significance do research and innovation have for a cloud provider?
We have a lot of research and innovation underway, across the entire product range. One example is elastic GPUs: previously, a GPU had to be connected 1:1 to a particular server; today, GPU power can be virtually assigned and removed according to need. Another is field programmable gate arrays (FPGA) in the computing cloud, which enable “hardware microservices” – that’s extremely innovative.
How do you make technological decisions?
On the basis of data. We passionately measure everything that can be measured – we are strongly driven by operations metrics. With this data in the background, we discuss and debate and are able to make a sound decision about what to do. We also use customer feedback, which we obtain from customer advisory boards, for example.
What will the cloud look like in 5 years’ time?
This is difficult to say. In general, the service range will continue to grow. We will increasingly be supporting customers in transforming their legacy technology into the cloud. The scope of services will also increase, and there will be additional regions across the globe. For example, we have just announced the Stockholm region and a region in Bahrain for the Middle East.
AWS offers over 100 services. Isn’t it difficult to keep so many Cloud Services under control?
Of course, it is a much bigger challenge today than in 2013, when I started working for AWS. On the other hand, the platform is also far more powerful. The choice of technologies is larger, and the services go deeper. The expectations of employees are changing: developers are transforming into machine learning engineers, and entirely new disciplines such as chaos engineering are being born – creating new skills and career paths. And that’s a good thing. Imagine if instead there was less choice ...
What about the customers – aren’t they confused by the enormous range on offer?
I don’t think so. The platform is this varied because that’s what customers want. An SME in a particular sector will be unlikely to use all the services – it will choose the precise combination that it needs. There is also more than one way to solve a problem. Customers are free to decide what they need in their current situation and to adapt their choice of technology to new requirements later.
Machine learning, AI and IoT are the hot topics of IT. What is the connection between them and the cloud?
With IoT, we connect up billions of devices. Each one provides measurement data, and they can interact, which produces even more data. The cloud is the natural habitat in which to store and process this data. This immediately gives us another challenge: the data volume is becoming extremely large – we’re talking petabytes. It’s getting more and more difficult to see what is happening. That’s where machine learning comes into play, to identify patterns and processes. This, in turn, generates even more data – a self-fulfilling cycle. Cloud, machine learning and IoT are closely linked together; you can’t have one without the other.
This is surely great news for cloud providers...
Of course. But also for companies which want to reinvent themselves and exploit new revenues. You only have to look at all the start-ups which are doing things which, 10 years ago, were literally science fiction.
AWS is the world’s largest provider of Cloud Services and runs a lot of mission-critical applications. How do you handle this systemic relevance?
Firstly, the security of the AWS platform is our highest priority. Secondly, we have a very good understanding of scalability and availability. The underlying architecture is built to be resilient. We have the concept of regions – there are currently 18 worldwide – and of availability zones which are independent of each other. If one of them fails, the others are still available. Customers are free to choose the regions and zones in which their applications can run. We generally recommend an active to active configuration across two or more availability zones.
What does working with local providers such as Swisscom mean for you?
Our customers like to see local faces, and they already have a relationship with local partners. The advantage of partners like Swisscom is that they can cover everything: from support for existing IT up to cloud transformation. They also have the experience of numerous transition projects. This benefits customers for whom the world of the cloud is still new.
What influence do you have, personally, on the cloud business?
Amazon’s uppermost management principle is “customer obsession”. Everything starts with the customer. Our innovation process works on the principle of “working backwards” from the customer: the initial basis is always the customer requirement. I myself and my teams go to the customers, collect feedback and suggestions for innovations and pass these findings onto the engineering department.
Does that work to your satisfaction?
The engineering teams love customer input. It’s a fantastic working relationship. At other providers, engineering is often pretty isolated and far removed from the customer for whom it is actually working.
What do you find fascinating about your job?
I am happiest when working with customers who have a very difficult problem to solve. When they arrive, you see their frustration. They want to build something up, to transform their business and inspire their customers, but aren’t getting anywhere. Accompanying the customer on the way to a solution, making the impossible possible – that’s what drives me and that’s why I love to get up each morning.
From morning to night: is there anything about AWS that makes you wake up in a cold sweat?
No, nothing about AWS. I wake up happy and full of energy each morning. To be more accurate, I don’t sleep that much at all because my brain is always active and thinking about all the things I want to do the next day.
More on the topic