fdmon (Fast Deployment Monitoring) : a new step forward in terms of performance !

Our fully REST API based distributed file system, API-FS, is undergoing a major change to make an fdmon cluster permanently immune to network latencies and temporary bandwidth limitations.

Indeed, nodes of an API-FS cluster now communicate asynchronously according to a publish/subscribe mechanism that nevertheless guarantees data integrity. During subscription, processing can be applied to the provided data, specific to the type of data and the type of the subscribing node (example : updating an index), which adds an additional dimension to the notion of object storage.

Before this change, API-FS objects underlying information provided by the fdmon user interface were systematically checked in terms of versioning, on “partner” nodes, which could degrade the user experience beyond 400 ms of network latency (including TLS handshakes).

This mechanism disappears with the new version of API-FS, which allows nodes of the same fdmon cluster to be distributed anywhere in the world.

In case of network latency peak, information is provided to the user with a delay of a few hundred milliseconds or a few seconds in the worst case (but respecting the order in which it has been built), which remains acceptable for monitoring or orchestration purposes. However, the user interface achieves extreme fluidity, unmatched on the market, regardless of the number of configuration items.

Furthermore, architecture of fdmon natively prevents any split-brain situation and automatically manages data reconciliation in the event of temporary loss of one or more nodes.

Until now, we guaranteed 100% availability of our fdmon cloud platform. This guarantee is now combined with a guaranteed user interface response time regardless of the status of the internet network and resources provided by cloud providers.

Thanks to its resilience and infinite scalability, fdmon has become a true enterprise IT analytics solution, able to be deployed on a worldwide scale and capable of managing several hundred thousand, even several million configuration items of the same organization.