Categories
Analysis

Cloud Computing Integration: A pain in the ass!

Cloud computing providers should face reality and tell their potential clients the truth! The integration within the entire cloud stack leads to the biggest complexity in the cloud. This also includes the other costs which no public cloud provider, in addition to the on-demand and pay-per-use promises, transparent has on its marketing slides. An overview.

What is the expectation of the user?

Besides the on demand use and pay per use cloud computing provider promise easy access to IT resources. That sounds fantastic on paper and normally works very well. However, the fun stops when it comes to combine existing systems or even new services in the cloud. The integration caused the greatest pain.

Discussions with users and inquiries from cloud interested companies reflect a harsh reality. Expectations, for example, are to replace existing on-premise solutions with 1:1 equivalents in the cloud. The hope is that thus the integration effort – which formerly existed on premise – can be realized much easier, since the systems of the providers “are already within a centralized infrastructure”. It is thus assumed that the familiar interface problems that bother every IT department for decades can be implemented easier. From the user perspective this is a comprehensible expectation. Finally, the vendors keep the threads within their cloud infrastructure (including SaaS offerings) in their hand and could offer a simpler integration between multiple systems within the own services.

Similar feedback is coming from the cloud system integrators. APIs and integration of multi-vendor services but also provider proprietary services in the cloud lead to the same effort as on premise.

What is the reality?

In principle, the cloud should be the Mecca for every IT manager. He can pick out the best services for his use case, test it in advance without much effort, and then integrate it. However, the full extent of the disaster is reflected by current cloud marketplaces. Cloud marketplaces are trying to catalog best-of-breed solutions by merging individual cloud solutions in different categories and thus offer companies a comprehensive range of different solutions, with which a cloud productivity suite can be assembled. Nevertheless, these marketplaces, although they are very heavily controlled and supposedly integrated, at a crucial point massive problems with the integration. There are many individual SaaS applications that do not interact with each other. This means that no common data base exists and for example the email service can not access the data in the CRM service and vice versa. This creates individual data silos and isolated applications within the marketplace.

The number of new SaaS applications is growing everyday and with this the need to integrate these different solutions with each other to exchange data. Today, the cloud market moves uncontrolled in the direction of many small isolated application, each offering an added value for themselves but, in the end, lead to many data silos. With such developments, companies have fought back in the times before the cloud – for nothing.

What to consider?

One thing (prospective) cloud users should be aware of and especially understand. The cloud is unequal to the known on-premise infrastructure. This means that even the migration of an e-mail system, one of the supposedly highly standardised applications, becomes an adventure. The reason is that many companies do not use a standard e-mail system and instead have connected proprietary add-ons or self-developed extensions, that do not exist in the cloud.

There are also other applications that belong to the IT enterprise architecture to meet very individual needs. I spoke with the strategy manager of a German DAX company with a global footprint, which is currently evaluating a cloud strategy. The company has about 10,000(!) applications globally (desktop, server, production systems, etc.). This example is extreme and does not reflect the average challenges of a company. But it shows the dimension in which one can move and which can also be broken down to much smaller application infrastructures.

Before their path into the cloud, IT decision makers should take time to think about it and use the chance to eliminate the over the years evolved special cases. A 1:1 migration without great effort and additional costs would not work and the potential savings, which the cloud provides, put into perspective. Anyone who decides to move a individually developed application to a cloud infrastructure (IaaS) and Platform (PaaS) should plan with the development of the application on the Greenfield. A 1:1 migration (maybe depending on the cloud infrastructure/ platform) will not work and is a myth.

Who needs help for the operation of its virtual infrastructure (IaaS) and development platform (PaaS), should not rely on a public cloud provider, but on a managed cloud provider or hosted private PaaS provider, who both offer additional professional services.

Cutting out the middleman doesn’t work!

Due to the self-service the public cloud has a flavor of “cutting out the middleman” since the enduser purchased directly from the vendor. For distributors who have specialized exclusively on the pure resale of hardware and software, this is true. Software-as-a-service will lead to their death if they do not change. Cloud natives (start-ups, innovative enterprises) will also use self-service. However, this does not apply to the mass in the corporate environment.

System integrators are more important than ever before. “Cutting out the middleman” in the public cloud does not work. Considering that interface problems and the lack of cloud knowledge are present. With his assumption, that Accenture already makes a $3.6 billion revenue with cloud consulting and integration, Carlo is tend to be right. The evidence and feedback from the market are clear.

In any case, system integrators and integration-as-a-service providers can look forward to a rosy future.