Cadulis uses many advanced technologies. You will find below some information on some of them, to give you a more precise idea.
Do you think Cadulis works very well? so do we :) but thanks to this top organization, your business is growing and you wonder if Cadulis will always be as responsive?
Rest assured, our architecture is scalable, that is to say that the server resources adapt according to demand. We did tests by injecting up to 10,000 procedures at the same time.
A company that believes, on the other hand, requires a specific organization to partition the activities and the teams. The still Cadulis has the tools to help you.
Responsiveness is one of the essential demands of our customers. No one wants to wait, even for about ten seconds for a page to load. Here are the tips we have implemented to make our application completely fluid, “frictionless”:
- an automatically scalable system. With Kubernetes hosting, the Cadulis infrastructure adapts its own resources to user needs. In slower periods (holidays, specific periods of the day / night), the cost of accommodation drops. We can therefore offer you the best rate on the market.
- Most of the processing has been made asynchronous, tolerant of “breakdown”. For example, sending an automatic email following the end of an intervention can be carried out a few minutes after the end, if the system is loaded. In the event of an error, Cadulis will retry the shipment a little later. You can thus continue your activity without suffering from delays which would be unpleasant to you. With this in mind, we have been able, with the advice of our customers, to define certain treatments as lower priorities.
- Efficient calculation funnels. Imagine the number of calculations necessary to determine the best niche for your intervention: You must consult all the niches of all your technicians, calculate the distances and travel times between all the interventions and calculate all the profitability! And yet the niche proposal identifies the most profitable players and niches in seconds. For this two tips: the extrapolated calculation of travel time (see the article on machine learning) and a funnel that quickly eliminates field workers too far away, who do not have the right skills, those whose planning is already full.
- Caching. Our users (business conduct or stakeholders) have to visit the same page several times. In these cases, the page is cached, which prevents the system from recalculating each time the journey time, distance, profitability data is used. Part of this cache system is moreover what allows use of Cadulis without having a network. (See the article on the offline).
Machine learning helped us to solve a fairly common problem: how to make a calculation simple but keep the precision. The calculation in question is to determine the journey time from the GPS coordinates of two points.
The GPS coordinates can give us the distance as the crow flies, by taking into account the curvature of the Earth, the calculation is precise and extremely fast. But from there, how to determine the real or even approximate journey time, without using an external mapping system, very greedy in precious seconds of calculation? These 0.2s duration calls constitute the most important part of the calculation time.
What will the journey time depend on? The density of road infrastructures such as expressways and motorways near the point of departure of the intervener mainly. This is where our machine learning comes in: To go from the distance as the crow flies to the journey time, we use a correction coefficient, which initially takes a low value for each participant.
With this value, a maximum of participants pass the filter steps of the slot proposal and therefore, for the first calls (often made in tests), the response times are not completely optimized.
Each time a call to a third-party mapping system is actually made for this participant, the coefficient is adjusted according to actual traffic data.
After a few calls, each stakeholder has a realistic coefficient and the niche proposal filters only keep the most relevant stakeholders to make the calls
Machine learning describes a system that becomes more and more efficient as it is used. This is the case here.
With cryptocurrencies, Bitcoin, blockchain is one of the buzz words. And yet beyond the circle of finance, this technology can really provide solutions to problems that we encounter on a daily basis.
For example in our intervention planning and optimization application, Cadulis, the blockchain is a way to secure the dematerialization of documents.
Here is the concrete case:
Imagine a technician coming to install the fiber in your apartment. At the end of the intervention, you sign a document (paper) to certify that the intervention has taken place. This signature is not binding on the proper functioning or the successful completion of the intervention, nor does it certify your identity, that you are the owner of the apartment or the indirect sponsor of the intervention. There is also no guarantee that the contributor will not modify the content of the document after signing.
And yet this signature has legal recognition that the digital signature does not yet have.
Our digital system, however, carries in itself more security: at the time of signing, the date is recorded. And each modification is traced with recording of the identifier, the date and the content of the modifications.
To go even further, the intervention report (or more precisely its hash) can be stored in a blockchain. Thus it is easy to demonstrate legally that the document is inviolable, falsifiable, with a time stamp and even a geolocation.
This is enough to stifle many disputes in the bud.
Another fairly concrete case concerns the transport of hazardous waste. The responsibility for these products resting with the producer until their destruction, the latter can with the blockchain safely trace the course of the various documents attached to the products it transports.