Geeglee Deployment: is it SaaStisfying ?


Why do our customers prefer "on-site" installation when SaaS is the general trend?


First, our Geeglee Pattern solution uses a 3-tier architecture: a web-based graphical user interface; a database; and the central module that manages both.

This future-proof architecture makes it possible to deploy the whole solution either on a standalone PC, or on multiple servers in a datacenter.


Technically speaking, moving it to a full SaaS solution would not take too long a time, but our customers do not ask for it:

As Geeglee Pattern is a software in which users put their knowledge but also their know-how, which are sensitive data and sources of competitive advantage, the harm would be tremendous if this information were to be disclosed, altered, or made unavailable. Indeed, in addition to the risk of loss of commercial secrecy, the data used by users is the results of calculations that can take several hours, and the unavailability of the results or their possible corruption could lead respectively to a decision delay or a bad decision.

Therefore, our customers, when considering the sensitivity of the data, the use cases, and who the users are, prefer to deploy directly on the PCs of these users, the Total Cost of Ownership (TCO) being only slightly more expensive than if the service had been offered in SaaS; and the benefits of SaaS would not overcome the risks. Last, but not least, many of our sensitive customers simply do not have the right to go for SaaS.



What about the Geeglee Pattern Light solution?

Geeglee Pattern allows users to model the topics that interest them, by describing the building blocks and their relationships (in the form of mathematical formulas), and then running the calculations, to, eventually, browse through all the alternatives.


Some calculations can take several hours, depending on their complexity but also on the machine on which they are run. So, we have developed a lighter version, Geeglee Pattern Light, which allows to model the topics but without being able to run the calculations. This version is particularly interesting, for example when it comes to sharing knowledge between different experts, or to preparing a model before loading it on a more powerful server to run the calculations.

Arnaud