SFR (Société française du radiotéléphone) is the second largest telecommunications company in France. Headquartered in Paris, it provides a range of voice, video, data, and internet telecommunications and professional services to consumers and businesses. Its services include fixed-line and mobile telephony, high-speed internet, IP television and TV boxes. SFR has more than 21 million customers, ranging from individuals to large corporations to governmental agencies.
SFR is responsible for managing a national telecommunications network consisting of millions of interconnected components. If just one element of this network fails, it can have an enormous knock-on effect on its customers. As SFR is bound by service level agreements, with financial penalties for unplanned downtime, they were keen to predict any points of failure, and minimise the impact of downtime on its customers. The network was monitored by 30 different systems which were inflexible and expensive to maintain. Any attempts to enrich data modelling within these systems would prove costly and painful. Consequently SFR consequently it was necessary to deploy an entirely new solution.
Through our initial Discovery (link: to Discovery content) work with SFR’s internal project team, we were able to confirm their core needs and reach a consensus on a proposed solution – deployment of Neo4j, a NoSQL graph database. This technology offered SFR the best potential method of modelling their network, as Neo4j’s 1:1 mapping would provide a more natural representation of their network’s physical components.
Following acceptance of the proposed approach, we progressed to delivery. Neo4j has the added advantage of being an open source product which allows for rapid development of prototypes. This enabled us to quickly demonstrate the viability of the technology, and ensure it met SFR’s needs. The result was a solution that empowered SFR: with a database of their own design, and the ability to query their network (analysing millions of components at country level), they could detect and assess any points of failure. With this information they could better plan maintenance activities, and reduce their costs, which in turn would provide their customers with an improved and even more stable service.
“Neo4j is an obvious solution to explore for a graph network problem. It manages risk on the client side, and it also makes processes more efficient” – Project Manager, SFR
“Getting the model right allows everything to flow naturally, which is a process that is simple and powerful and works well using Neo4j” – Project Manager, SFR
Collaborating with SFR at every stage, from concept through to production, ensured the solution stayed aligned with their needs, whilst affording SFR an opportunity to gain knowledge of graph database implementation and contemporary software practices.
As a result of our partnership, we were able to leave the SFR team with:
SFR is now self-reliant, and no longer dependent on legacy third-party systems. By working in partnership with us, SFR now possess the technology, knowledge and expertise to proactively manage their network, and offer their customers the best possible service.