Modern Cloud
Cloud computing is a key technology for us at Trice. About two decades old it continues to grow in importance for product and service delivery. At the same time a lot of organizations struggle with their cloud journey and, to be honest, there are still a lot of things we miss in terms of simplicity, developer experience, and guidance when working with cloud today.
Our main points are the following:
Software and hardware as a service are becoming the defaults. Your ability to innovate and move fast are limited if you can’t use a cloud environment. That is where top tier suppliers and partners will focus their effort. A solid cloud and data strategy is the foundation for being able to pick up technologies like AI.
Developers and organizations struggle with the complexity of the cloud of today; resulting in poor data quality, high cost and large projects without enough technical direction. We recommend moving in smaller chunks with clear guardrails to solves some of these problems.
For the issues that remain, such as developer experience, data quality, or use case specific compliance, we see a lot of interesting things on the horizon. Industry dedicated clouds, European pure cloud challengers such as Evroc and new data platform products suitable for the AI age.
Using Cloud is Key to Success in the Next Decade
The title of this post is modern cloud and with that we aim to go beyond the virtual machines, containers and services of the 2010s. Instead we want to highlight some key technical shifts that further strengthens the case for cloud.
The cloud of tomorrow solves more of the complexity and increases innovation speed
The last five years of development have moved cloud-based applications in a direction where you can build even more with a small team. While large organizations may still need strong infrastructure teams and developer experience organizations to guide in the plethora of cloud services, there are now good, well-structured defaults available from some cloud providers. Use these wisely and your organization will be able to place five or ten innovation bets with the same resources that were required for one in the past. Or operate large systems with small crews using a high level of out of the box automation.
The supplier ecosystem is moving into the cloud
A lot of key services and suppliers are now cloud first. Data tooling, AI services, monitoring and security systems are built cloud first. Your suppliers may expect you to provide an authenticated blob storage bucket or event queue to connect to in their service. While most still offer on-prem support it is seldom an out of the box, supported integration as with cloud solutions. If you want to be able to use state of the art services, they will expect you to connect to them in cloud environments.
To build AI-enabled products access to hardware is key
Behind the meteoric rise of Nvidia is a huge demand for hardware to train everything from language models like GPT to personalized recommendation models by innovative product companies. The big buyers of this hardware are cloud providers and they provide a shortcut to access. When building AI products your needs are also hard to forecast. Training a model may require enormous amounts of GPUs and storage while the resulting product uses a much smaller footprint for predictable inference and day to day delivery. Sharing resources through a cloud provider cuts cost and development time significantly.
A wider range of options allows more organizations to fully use cloud services
Beyond the big cloud providers, the ecosystem is growing. Oracle is making a play with a different cost structure and focus on areas like defence. In Europe we are excited about Evroc making an entry with a focus on becoming a sovereign and sustainable cloud provider for Europe. Evroc will also be a pure cloud provider without a lot of other business stealing focus and as they are going live in 2024 we have high hopes for a developer friendly and very modern platform. Combined with specialized offerings from the big players a lot of new domains can now consider cloud in earnest.
Security is more important than ever
We have experience securing high profile systems and applications at Trice. Securing cloud applications has challenges, especially when it comes to navigating configuration options, but it is vastly easier than running everything on your own. Having access to always up-to-date, ephemeral serverless systems on top of authenticated, encrypted, storage and transport layers from a cloud provider limits the number of details you have to pay attention to. Not to mention the value of your infrastructure being supported by some of the best security personnel in the business.
Become a Role Model in Cloud Computing
A core belief for us at Trice is that we want to help organizations become the role models of the future. That is to say that you should aim for what will make you successful five years from now, not copy what made others great five years ago. We have seen too many poorly executed attempts at copying some aspect of Spotify, Google, Apple or Tesla. Learning from great tech companies is incredibly useful but you should aim for where they are going, not where they were. And know what applies to your reality and what does not.
So where would we start a cloud journey today to become a role model in 2030? While the topic is too big to cover in a few hundred words here are some high-level ideas we have seen work in practice.
Keep it simple, start small with a focus
You do not have to make the cloud transformation journey into an enormous project from day one. We have created amazing products on all three major cloud providers and a couple smaller ones as well. They all work well for building most applications and platforms. Spending years evaluating edge cases to make a choice is not money well spent. Cloud is billed by the second and that allows trying it out in however large or small chunks you need.
Focus on your business value and understand what parts would benefit most from running in a cloud environment. Set a clear goal for a small piece and try it out. With a clear goal we mean a business outcome, not that you are “now fully serverless” or “doing something with machine learning”. Cloud services enables you to deliver better customer experiences or cut costs and development time. Define such measures of success and do not get stuck on buying tools.
By starting small you can put together a talented team of internal and external expertise to lead the way. Bring in a couple of people who have done this before and pair them with growth focused people from your organization. Your first small piece should either create a good example for the rest of the organization to follow or provide a lot of insights quickly on what to change when you try again. A few quick failures to learn can often be expected and it is better to make them small than to fail on multi-million-dollar transformation project.
Set clear guardrails as you grow your cloud presence
Most cloud providers have many overlapping services and options. Twenty years of history is an eon in tech and not all the services available are suitable for your situation. As you continue to move more small pieces of your products following the good example, there is a need for guardrails. This means limiting technical choices, encoding security and compliance requirements into your solution. With clear guidance on technical trade-offs your teams can focus on the customer value instead of arguing about which database service to use.
Let your business goals set the pace for large scale adoption
One reason that we argue strongly for moving piece by piece is that it is hard to predict exactly what will happen along the journey. Innovative technologies emerge, crises occur, and focus will shift. If your organization works in a modular fashion a system migration can be paused, switched out or removed entirely and kept as a legacy solution until end of life.
Invest in understanding your data
On the topic of learning the good parts from famous tech companies we would like to emphasize investing in structuring and understanding your data. Google, Amazon, Spotify and others have spent years investing in indexing data and improving metadata quality. This enables them to use cloud tools and services to build great data driven products. Buying the tools without structuring the data is an all-too-common waste of effort and money.
Invest in always evolving
The next big piece about becoming a future role model is to set your organization up to continuously evolve. Moving to the cloud is not one big five-year project and then you are done. It is aligning, sequencing, and running many small, impact focused projects to get you closer to your vision by using cloud technologies. Once you are using cloud services to drive your business forward continue to work in that way. That will allow you to use the latest hardware and services for innovation in parallel with keeping your stable revenue sources available, secure, and compliant on well-tested managed infrastructure.
Move forward with confidence
Any organization that can manage their data with high quality in a cloud environment is set up well for the next decade. If you also run modular applications and systems that can be independently improved and updated, you will be able to make the most of that data. This good start is aspirational in some domains but blockers to getting there are being removed by cloud platforms having evolved to serve more regulated and critical use cases like healthcare. New players are also adding options that did not exist before.
We believe in a world where collaboration around data from many sources with many consumers will unlock new waves of innovation and partnerships. A lot of these collaborators will get help from or be AI users. Working together in this way requires a platform that can encode business goals, legal aspects, data models, access rights and security in a seamless manner. That is the modern cloud and we are already involved in projects that use it to drive innovative products in healthcare, cybersecurity and defence.
