Abstract: A method comprising: capturing by software agents monitoring a plurality of network interfaces, telemetry data representing a plurality of data flow samples associated with an unknown Internet host connection which is assigned a unique identifier; processing the telemetry data to calculate a respective feature set for the unknown Internet host connection; applying, by each of the software agents, a respective instance of a trained machine learning classifier to the respective feature set calculated by the software agent, to obtain a respective field classification which associates the unknown Internet host connection with a particular application or Internet service category; and determining a final classification with respect to the at unknown Internet host connection, which associates the unknown Internet host connections with a particular application or Internet service category, based on a majority or plurality consensus among all of the field classifications.

Abstract: A method comprising: identifying, by a cloud server, a set of local area networks (LAN) associated with the cloud server, based on a similarity parameter with respect to an end device connected within each of the LANs; forming a communications network comprising all of the LANs in the set; detecting, by at least a subset of the LANs, a download file received by the respective end devices in each of the LANs in the subset; calculating an update event likelihood score with respect to the download file, based, at least in part, on a plurality of parameters associated with the download file; and issuing, by at least one of the LANs in the subset, a notification that the download file is associated with an update event affecting all of the end devices in each of the LANs, when the update event likelihood score exceeds a specified threshold.

Abstract: A method comprising: identifying, by a cloud server, a set of local area networks (LAN) associated with the cloud server, based on a similarity parameter with respect to an end device connected within each of the LANs; forming a communications network comprising all of the LANs in the set; detecting, by at least a subset of the LANs, a download file received by the respective end devices in each of the LANs in the subset; calculating an update event likelihood score with respect to the download file, based, at least in part, on a plurality of parameters associated with the download file; and issuing, by at least one of the LANs in the subset, a notification that the download file is associated with an update event affecting all of the end devices in each of the LANs, when the update event likelihood score exceeds a specified threshold.

Abstract: A system comprising: at least one hardware processor; and a non-transitory computer-readable storage medium having stored thereon program commands, the program commands executable by the at least one hardware processor to: receive a sequence of computer program instructions, analyze the received sequence of program instructions to indicate execution points associated with non-deterministic program instructions, execute the sequence of computer program instructions while capturing program state data, wherein said program state data is associated with at least some of said indicated execution points, and, generate a replay of said execution, wherein said generating comprises inserting said captured program state data at said associated execution points.