Even though at present we have a clear vision on the Cloud services that we are deploying, it is clear that the industrial view that we want to apply suffers from the presence of major players who are taking over the market. Even for a company like Orange the situation is not simple.The extension of our development activity (build) to production (run) is an obvious step, but how can we approach it without the support of major communication campaigns?If among the fundamental values of Cloud Computing there is a total abstraction of the geographic location of data, this particular point brings forward the notion of trust in relation to the service provider which handles a considerable part of the company’s digital assets. If tomorrow production, payment or CRM data are lost, how is the company supposed to function?What if there were a solution that offered to keep data hosting within the company? The Cloud service provider only provides the required power of execution in its datacenters. This appealing idea regarding information security obviously implies technical constraints that can easily be overcome:- The client’s site has to be close to the datacenter in order to maintain a synchronous accessibility (or replication).- Identical storing equipment has to be used in order to limit conversion platforms.Now, if major players of the field are flooding companies with business offers, the community approach is interesting for offering services that are as industrial as the others, but taking into account more concrete problems:- Geographical community: a CIO’s interest is to use Cloud Computing with “local” actors.- Business community: the interest is in proposing packages (from IaaS to SaaS) to business managers, with already proven migration plans.As regards the geographical community aspect, it is true that the resource mutualization capacity is greatly reduced. Moreover, except for major French towns, it is difficult to find reliable, available and modular hosting resources. I found the COLT initiative in this field to be in line with this constraint. Their offer consists in building a 500m2 datacenter in 4 months based on containers. The other problem is related to the interconnection of the IP and electrical networks. On the initiative of local authorities, local fiber networks are extending. France is not covered, but this is improving when you see plans that are in place. As for electricity, although there is little interest in placing these data centers close to a production site (nuclear or other), it is certainly easier to find a place which is redundantly covered.On the other hand, we have already been asked several times to offer infrastructure outsourcing solutions outside EU frontiers (each had their own reasons that don’t concern me). We cannot provide a solution at the same level of performance, but we already have answers to these problems that are close to Cloud Computing.The business community is more difficult to target because, in order to succeed, we need credibility. The main service category is SaaS and the others stem from the first. Work on software editor reconversion towards these Cloud Computing activities is going in this direction; even Syntec (the French IT association) is active in this respect. It is therefore in these editors’ best interest to group together and organize their client community in order to make them resource mutualization propositions and therefore offer more services. But not all these editors have the capacity, competence or desire to include the “run”.In short, if we have explored the paths of generic, general or industrialized Cloud, we are also paving new ways that will allow us, together with our partners, to bring correct and precise answers to community expectations, be they geographical or business communities.