Taking full advantage of the Cloud in Switzerland
Every digital product requires a very concrete part: infrastructure. Thousands of physical servers are needed to enable this digital revolution. Cloud computing is the name chosen to abstract this computing power.
In this field, we mostly hear about Amazon, Google and Microsoft. These American giants are disrupting every market, starting with IT infrastructure.
To create a digital product in Switzerland, do we need to do without these giants? Is this possible today?
Why the Cloud?
Cloud services, from IaaS to SaaS, promise above all to focus on value creation. Automation facilities make it possible to turn what is no longer a competitive advantage into a commodity. With a reduced financial and human investment, companies can create digital services focused on their core business.
For example, Swiss private banks, like retail banks before them, automate the opening of customer accounts, while the bankers themselves intervene in more specialized areas (advice, analysis, etc.).
Why develop a service in Switzerland?
The regulation of the cloud ecosystem is forcing companies to be aware of the issues at stake, particularly in terms of data confidentiality and privacy.
The U.S. Cloud Act, for example, gives U.S. government agencies access to service providers' offshore infrastructures. Closer to home, the Swiss financial center's umbrella organization requires that technical, organizational and contractual means be put in place to guarantee data confidentiality in the cloud. Respect for privacy, a strong element of Swiss culture, can also be a powerful market differentiator.
These are all reasons to keep your data, and therefore part of your IT resources, on Swiss soil.
Is it possible to develop a 100% Swiss service? No.
Clearly, the Swiss Cloud offering in terms of richness of services is behind what Amazon, Google, Microsoft or Stripe and other specialized services can offer. Does this mean we have to forego the promises of the cloud? Not necessarily. Confidentiality and privacy issues don't force us to make an all-or-nothing choice. There's still a middle way: hybridization.
Hybridization means finding the right compromise between the need for services only available from the big American players and the need to keep data on Swiss soil. But this is no simple matter. We illustrate these difficulties with three hybridization models we have encountered.
Hybridization models
Green Field model
Let's imagine a young start-up founded around a new digital service, which, for legal or reputational reasons, must guarantee a certain level of confidentiality and therefore cannot benefit from American suppliers. It must agree to set up its infrastructure in Switzerland, with all the associated costs.
But the cloud isn't just about Infrastructure-as-a-Service, it's also about Software-as-a-Service. Operating your own virtual machines with a Swiss provider doesn't mean you have to re-develop the project's ancillary functionalities (e.g. a mapping service) yourself.
The challenge is to define the functional scope, identify the buy vs. build trade-off, and compare it with the nature and sensitivity of the data exchanged with any external services.
Sidecar model
Some SMEs have their own infrastructure, in their own data centers in Switzerland. This information system generally comprises a range of applications to support business processes. E.g. core banking system, customer repository, ERP.
In addition to technical difficulties, migrating these applications to a cloud provider is sometimes impossible: the data handled is subject to excessive confidentiality constraints (typically customer identities) and compliance requirements (FinMa and other regulatory bodies).
However, not all new digital product opportunities require access to this data. A company can set up a dedicated digital division without necessarily needing personal data, and can choose cloud services without restriction.
Interactions between the two systems must be minimized: each communication channel has a technical and organizational cost. Once again, a compromise must be found between the autonomy of the division and the scope of the data it requires.
Shared Responsibility model
The strong autonomy of the division in the previous model is not always possible: its need for integration with sensitive data and services may be too great for it to be able to decide alone which cloud services it can use, as the impact of such decisions goes too far beyond its scope of responsibility.
Historically, this responsibility was carried by a "platform" team. Having to instantiate any cross-functional service at the request of customer divisions, it quickly became a bottleneck.
Nevertheless, American cloud providers all offer organizational and access control concepts. The platform team can provide teams with controlled access. Each division has full autonomy over its resource space. The platform team is responsible for the robustness and watertightness of allocated cloud accounts.
The difficulty with this model lies in clearly defining the scope and sharing of responsibilities.
Conclusion
The digitalization of our economy is forcing companies to accelerate their adaptation to change. It's hard to imagine a digital service that's only open from 8am to 6pm, 5 days a week.
Companies therefore need to focus increasingly on the value of their core business - their DNA - and turn what is no longer competitive into a commodity. This is what the Cloud promises.
While the local Cloud offering still lags far behind the American giants, there are still good reasons to keep part of your resources on Swiss soil. Indeed, respect for privacy and data confidentiality, strong elements of Swiss culture, often represent a legal, ethical or marketing issue for companies.
We are convinced that the country possesses clusters of high added-value know-how, and that these clusters have the capacity to build X-as-a-service - hybridized - offerings on Swiss soil for the local market or for export.
