This blog article is for developers, engineers and everyone else who is driving innovation in their company. If you are one of these technically skilled people who is having trouble convincing management to add cloud services to your toolbox, you can stop reading right now and just send them this article.
To Cloud or not to Cloud? In the last couple of years, a lot of companies moved their services (at least partly) to the cloud. For many people, the cloud question is a topic they are very passionate about; in either direction. They either love “the cloud”, or they hate it and think “it’s just another hype that will pass”. Others burned by a bad experience with a cloud project in the past do not even consider using it for any future projects. This article introduces you to the advantages and disadvantages of adding “the cloud” to your toolbox.
At first, we have a look at “the cloud”. Most managers do not even know what’s inside the cloud. They see “the cloud” as just another tool for moving their servers to a public data center. So, they refer to just the computing services e.g. EC2 in the cloud which is just one single service in “the cloud”. As shown in the picture below, there is a wide range of available cloud services. The depicted services are from AWS, but Azure and Google Cloud have similar offerings, starting with basic computing services up to sophisticated services such as IoT-related or AI-related ones.
Basic computing services are really easy to implement in the data center of your choice but moving to higher-level services (so moving from IaS, PaS to more SaS) becomes quite a bit more complex. Smaller data center providers currently move in that direction. They expand their service portfolio and provide more services either as managed or even truly as “as a Service” meaning managed completely by APIs. In current modern infrastructures that are services-based (like microservices, etc.), new products can be generated by coupling existing services (self-made or from e.g. cloud providers). Still, some codes need to be written and integrated, but a lot of existing services can be leveraged to provide high-quality products, easier and cheaper than ever before. Adding cloud services to your service toolbox enables teams to create better products in less time because they simply have access to more services and they can use them to improve their services.
How is this related to innovation?
Innovation is often experiment-driven. Let’s say one of your teams wants to try a new speech-based interface.
The classic way
You have to order new servers, find a vendor to provide speech recognition software, integrate this software, maybe get some new dedicated hardware with GPUs for AI processing. Most of the contracts for the servers have a duration of at least one month, maybe even a year, while the servers are used only for a couple of hours a day. Until now, it was a very expensive process and took maybe a couple of weeks. You evaluate the new interface and learn that it does not meet your requirements. The project gets discontinued. You are still sitting on your servers and have to pay the monthly fee or even own them with no use case in sight. All in all, this experiment cost you a 4-5 digit amount of money – not considering the bill for the hours worked.
The cloud way
You still have to integrate the API in your software, but you can rent the needed servers and only pay the hours you need them. This is also true for really specialized hardware e.g. GPU instances. You have them available right when you need them and if the project fails, you just turn them off – no further costs.
So far so good, the real benefit is in the mindset of the people. You enable your team to experiment with new approaches, really cheaply and fast. Imagine you have e.g. 10 000 € available for innovation. Your team can do one project with this sum using the classical way. For the same amount, your team can do five projects in the cloud – you increased your chances to have a product breakthrough by 500%. This also enables the company to organize specific innovation events like Hackathons. People working 24-48 hours on their ideas can’t wait weeks for their servers to be ready in your data center – they need them right now. If you force your people to wait to try out their ideas, it will kill these ideas and reduce the likelihood of innovative tasks going on. For innovation, you really need to reduce the friction to bring ideas to life and verify that they are working in the real world. Cloud services enable you to do exactly that. Are all these ideas going to work? No of course not, but that’s why you have to try out a lot of them. Innovation is about trying out a lot of things and learning on each integration. If you can choose for the same money, making 10 or 20 iterations – what would you go for?
Most managers do not see costs that are not shown on any paper. Direct costs such as salary, expenses for hardware, etc. are quite easy to measure – just use the most sophisticated accounting tool on the world (Excel) and sum it up. But wasted innovation potential is really expensive. These numbers are hard to guess and are not shown in any cost report. At first, you are not getting the new products and features that you could, which is a competitive disadvantage and second, you are not providing a work environment where creative and innovative people want to work – and never forget talent attracts talent.
We know that with power comes responsibility, at least since we watched Spiderman. Depending on your current procurement process of buying cloud resources is quite different. More teams than ever before are responsible for managing their own servers. Compliance and data privacy are also quite differently managed in cloud environments than in on-premise settings. You have to get your IT department, legal, etc. on-board, partner them with cloud experts to get the best out of it. Managing Cloud infrastructure needs quite different knowledge than managing bare metal servers in a data center. Get partners and experts to support you. Sounds like a lot of work? It is, but totally worth it.
Rene Schakmann, Head of Technology