Care and support for the development of Navifleet local infrastructure

Navifleet. pl is a comprehensive platform for fleet management. It offers m.in: GPS locator and car monitoring, remote tachograph service, settlement of business trips or e-TOLL. It helps optimize fleet maintenance costs and facilitates route planning.

Beginning of cooperation

The infrastructure NaviFleet.pl from the very beginning is built on local servers, and the application itself is based on Open Street maps, which are used to monitor the fleet of NaviFleet Clients. Along with the company's intensive development and dynamic changes in the transport industry, it was necessary to implement further, new solutions.

Architecture

The Navifleet system from the very beginning works on the basis of dedicated servers and it is on them that the services required for the smooth functioning of the system work. As the company developed, new services were launched. The Client's infrastructure is now a dozen or so servers that carry out tasks related to the storage of telemetry data and analysis and sharing them with the end client.

The Navifleet system currently supports 3,700 users, which translates into almost 12,000 GPS devices. Support for several thousand devices requires an extensive system of receiving data, which are queued in the Rabbit MQ service, and then saved in SQL databases. Databases are run within pairs of servers in which one acts as a MASTER and the other slave. Writing always takes place in the MASTER database, and readings (depending on the requirements) are performed with MASTER or SLAVE. The services made available to Navifleet clients are built on the basis of the HA architecture. Thanks to this solution, Navifleet can provide the end user with access to data even despite the possible failure of the servers themselves.

The services that support the operation of the entire system are the local Open Street Nominatim server that speeds up the recognition of addresses for the given coordinates, on which we installed an application that automates the download of all map updates. The Spark Cluster is responsible for analyzing and providing the customer with the processed data, while the Cassandra Cluster serves as a place to archive and store the processed data, which accelerates the generation of reports.

Challenges of e-TOLL

The challenge for the system and architecture was to launch e-TOLL support. Due to the fact that in about 2 months about 4,000 new devices were connected, it was necessary to rebuild the architecture of the receiving system. Previously launched services on physical servers have been changed to virtualized systems in three groups: HProxy, RabbitMQ, and Tracker Servers. This solution now allows you to connect more machines at a very low cost and increase the efficiency of data reception.

Driving Style Analysis (ECO)

The developing fleet system and care for minimizing their maintenance costs encouraged Navifleet to create a new service: "Driving Style Analysis". Mechanisms have been developed that provide users with suggestions on how they can reduce fuel consumption and help select drivers who require training in this area. Support for these services required running additional data conversions that are performed in the Spark cluster.

Summary

Hostersi at every stage support Navifleet in the selection of hardware and installation of software indicated by the Client. An important aspect of our cooperation is round-the-clock monitoring of services and alerting if the parameters defined jointly do not fall within the set values, thanks to which we can jointly respond to critical problems with the application or architecture.

ANY QUESTIONS? CONTACT US

Read also:

• Nationwide call center, based on Amazon Web Services
• Implementation of cloud infrastructure for Magic Commerce, based on microservices, containers and IaC approach
• Implementation and maintenance of cloud infrastructure in AWS for Displate
• Migration of DANONE applications and websites to Amazon Web Services
• Design and implementation of cloud infrastructure for optAd360, based on Kubernetes clusters and the IaC approach