Retour

Cloudifying the uncloudifiable!

“My company’s apps are cloudifiable, super-cloudifiable in fact”. Unfortunately, repeating this phrase over and over is never going to result in a miracle. Launching an IT system into stratospheric orbit remains a complex topic. Of course, if you start with a blank canvas or if you only move forward gradually, the cloud ceiling will be lower, with the propulsion required to breach it consequently less too. Having effective solutions to the key issues of control, management, governance and change management for users allows for the complexity inherent in today’s world to be increasingly better controlled.

The two usual objections are the quality of the access networks and security. With the former, what is certain today is that the capacity to provide increasingly more effective solutions is steadily progressing, even without there being total control yet. When it comes to security, it is rather the case that risks tend to increase in line with the growth of the cloud. Whatever the quality of the products and techniques for protecting data and apps, the nature and virulence of the threats evolve in a globalized environment subjected to increasingly brutal stresses, and one where our IT infrastructures also become targets. “The man and his safety must be the first concern of any technological adventure”, and in keeping with the spirit of what Albert Einstein said on the subject, it is essential for security concerns to remain central to the cloud project, from its design through to its production management. In this regard, costs and efforts will only increase over the coming years.

Looking under the hood of our rocket reveals the infrastructure of the cloud environment that underpins our apps, with industrialization currently masking its technicality and complexity. As such, deploying complex virtualized environments that reproduce any type of architecture is easily organized using orchestration products, which renders what we were implementing manually just a few years ago prehistorically obsolete.

On the upper level, finally, are the apps. They are the very heart of the system. In 2010, Steve Ballmer announced that there were 4 million apps available for the Windows platforms, of which, at that time and even still today, we leave the vast majority in the hands of our users . A relatively small portion has been rewritten or has been natively designed for web environments. Mobile apps have also emerged and, for the professional world, are extensions or complements rather than the foundation of business apps.

Windows apps, which are not fundamentally different from the first Windows, are therefore still a key factor. Obviously, they have evolved in terms of ergonomics, their capacities and power, and they offer a considerable advantage in that—brace yourselves—they work perfectly and meet users’ productivity needs. At a time when the cloud and any type of device have flourished, while remaining viscerally installed on Windows systems, these apps have been reproduced in little and large, heavy and light, Linux and Mac, mobile and still the classic desktop. This has been made possible by various technologies, be it the publishing of remote apps or desktop virtualization, with the formidable and considerable advantage for the publishers and producers of this mass of almost 4 million apps, of not having to redevelop them.

The issue is thus not users’ access to these apps. It is, however, clearly their life cycle with regard to IT services. Their bond with the Windows base is absolute, complicating any installation and updating with procedures that have never really changed. This is barbaric. While solutions have emerged to reduce this bond between the distribution solution and app isolation, often, the problem is merely bypassed. Microsoft has tried several times to revisit this internal issue with development tools that facilitate the life cycle, but this has come coupled with the terrible difficulty of having to convince the creators of this protean mass to break everything. Just knowing the direct costs, including human and financial, and the indirect in terms of risk-taking, is enough to understand that there is not much at risk here.

The cloud will take a decisive step forward the day that this complexity in the life cycle of the apps, which is the result of lack of industrialization, is overcome. This is changing. The emergence of container technology, championed by Docker, applied to the world of Windows for apps with interfaces, is opening up new avenues. Together with publishing and virtualization technologies for restitution, this approach could enable strong industrialization of this life cycle. Let’s dream a moment: it could even go beyond this with a supply chain for the life of the app, which would run from the publisher to the user, smoothly and without complexity, with an industrialized operation that would no longer require hours upon hours of manual interventions. A truly disruptive technology.