Each server operating at the edge of a network runs in a lights-out environment where there is no direct access to the server and where it may be deployed outside the convenience and security of a data centre.
These edge servers run autonomously without supervision while collecting data in real-time and responding to any arbitrary queries submitted from the network centre without any prior knowledge of those queries. These queries can be very diverse. For example, a query reaching an edge point may require the edge server to aggregate a handful of columns across billions of rows or drill down and retrieve the majority of columns from a handful of rows. Whatever the query, the server has to respond in reasonable time without any prior design to achieve that response time.
Moreover, messages being collected at edge point may be arriving at network speeds reaching many hundreds of thousands of messages per second from multiple gateways, devices or probes. These messages are materialised as rows in real-time and made available for queries.
More than one server may be operating at an edge point to provide resilience and high-availability for data collection and queries. Multiple servers at an edge point will collect data from the same data sources; and the servers will automatically synchronise their message data to provide a comprehensive set of messages on every server at the edge point.